Newsletter · 2026-08-19 · Ashutosh Agarwal

It All Comes Down to Jackson Hole as the Fed Splits Nine to Three - The Dollar Brief - Week of August 19, 2026

The Dollar Brief for the week of August 19, 2026. Kevin Warsh's August 28 Jackson Hole speech has become the only date that matters for the dollar, the July minutes revealed the widest Fed split in two decades, and central banks now hold more gold than Treasuries as the dollar's reserve share slides toward 54 percent.

The Dollar Brief

Week of August 19, 2026: It All Comes Down to Jackson Hole as the Fed Splits Nine to Three

For the rest of this year, almost everything the dollar does now runs through a single ten o'clock speech in Wyoming.

On Friday, August 28, new Fed chair Kevin Warsh takes the podium at the Jackson Hole symposium, the annual gathering where the world's central bankers meet in the mountains and talk big ideas. He has spent his first weeks in the job saying as little as possible on purpose, and the entire market is now leaning in to hear what his "reaction function" actually is: in plain English, what will make him raise rates, cut them, or sit still.

The strange part is where the debate has landed. Six months ago the argument was how many times the Fed would hike. This week you could hear a top Wall Street economist stick to a call for three hikes, another shop warn the bond market is asleep on a September hike, and, on the other side, respected voices putting real odds on cuts. Nobody agrees. And while they argue, a quieter story keeps building underneath: the world's central banks are voting with their reserves, and they are voting for gold.

Here's the week, in plain English.

TL;DR

The one date that matters now is August 28. Precious-metals analyst Craig Hemke pinned it down: Warsh gets the traditional Fed-chair slot "about 10 o'clock Eastern… on the Friday," and "that's going to be a volatile day" (The KE Report, Aug 17).
The Fed is more divided than it's been in a generation. The July meeting minutes showed "a 9-3 vote… on whether to keep rates the way they were," which Hemke called "the widest it had been in a couple of decades" (The KE Report, Aug 17).
What the market actually wants from Warsh is a rulebook. JPMorgan's cross-asset chief Kay Herr said the bond market is "worried about what's the Fed's reaction function," and "adding some clarity… can help calm the market down" (Bloomberg Surveillance, Aug 18).
The hawkish extreme: three hikes, and the long end could "get unanchored." Bank of America's Aditya Bhave held his call for three rate hikes because "the unemployment rate is down. Inflation is up… and yet the policy rate is 75 basis points lower than it was" (CNBC's Fast Money, Aug 12).
The bond market is "underappreciating the risk of a hike in September." RenMac's Neil Dutta warned that if the Fed doesn't move by September "they probably don't get it done at all for the year", and thanked Warsh for creating "a bull market… for Fed watching by design" (RenMac, Aug 14).
The dovish read: Warsh is trapped and "praying for the immaculate economy." Vincent Daniel argued the Fed can't really tighten because "interest expense is now bigger than the defense budget," so Warsh needs "a strong economy, a slightly weaker dollar, and lower rates" all at once (RiskReversal Pod, Aug 14).
The wildcard nobody is pricing: Warsh could change how inflation is measured. Texas A&M's Jill Satina flagged that one of Warsh's new task forces is looking at "CPI methodology," and the last time that happened, in 1995, it "downwardly revised the U.S. CPI computation by 1.1 percentage points" (WSJ's Take On the Week, Aug 16).
The dollar is drifting under 100, but maybe not much further. FX veteran Marc Chandler sees "scope for one more push lower," targeting roughly 99.20, while cautioning "this might be the end of the boxer's punch rather than the beginning of it" (The KE Report, Aug 14).
"How much lower can the dollar fall?" David Woo made the contrarian case against the crowded dollar-bear trade: "if the euro is very weak and the Japanese yen is very weak, how much lower can the dollar fall?" (Wealthion, Aug 18).
A September Bank of Japan hike is now the base case, but the yen keeps sliding anyway. MUFG's Derek Halpenny, once "one of the most hawkish forecasters," noted the market now prices about 20 basis points of hikes for September, yet the yen has "quickly given back around half of its intervention-driven gains" (The MUFG Global Markets Podcast, Aug 14).
The contrarian yen warning: even 2% Japanese rates might not be enough. JPMorgan's team held a bearish 164 yen target; economist Steve Hanke called the rescue a "dead-cat bounce" and expects another intervention, because the real problem is Japan's money supply "growing very slowly," not the interest-rate gap (At Any Rate, Aug 14; The David Lin Report, Aug 18).
The quiet story: the dollar's share of world reserves has fallen from 71% to about 54%. One markets podcast put the number bluntly and noted "gold holdings on central bank balance sheets has now exceeded U.S. treasuries" (InvestTalk, Aug 19).
China is buying gold like never before. Wealth manager Clive Thompson walked through the ramp: from "30 to 40,000 ounces per month" a year ago to "640,000 ounces… in July", with China's gold still just 8% of its reserves versus "60, 70%" for the West (ITM Trading, Aug 14).

What's new

The whole market is now pointing at one speech

The single most repeated idea this week was simple: stop trading the data, and start waiting for Warsh.

The clearest calendar came from Craig Hemke on The KE Report (Aug 17). Jackson Hole runs "Thursday, the 27th through Saturday the 29th," and "the slot typically reserved for the Chairman of the Federal Reserve is about 10 o'clock Eastern… on the Friday. So… next Friday the 28th, when Warsh speaks, that's going to be a volatile day." He also flagged the appetizer before the main course: the July meeting minutes (due Wednesday afternoon), which he said showed "a 9-3 vote of those FOMC governors on whether to keep rates the way they were. And that was like the widest division, because usually it's unanimous or 11-1… 9-3 was the widest it had been in a couple of decades." A Fed that splits nine-to-three is a Fed nobody can confidently read.

Why does one speech carry so much weight? Because Warsh has deliberately stopped telling markets what he's thinking. On Bloomberg Surveillance (Aug 18), JPMorgan cross-asset CIO Kay Herr explained that the recent rise in long-term borrowing costs "is not being… inflation-driven. It's more of a term premium story", that is, investors demanding extra compensation just to hold long bonds when they can't tell what the Fed will do. Her point: "the market's worried about what's the Fed's reaction function. When… what is the Fed looking at for inflation?" She's watching two things: the minutes, and "Chair Warsh speaking at Jackson Hole." Her verdict: "his adding some clarity to the reaction function… can help calm the market down." (A "reaction function" is just the Fed's if-this-then-that rule, what data makes it move.)

Her colleague, BMO's chief FX strategist Mark McCormick, added a subtler read on the same show: the "vibe" from Warsh is that "he's thinking more about what's driving inflation", "industrialization, electrification, fiscal dominance, energy policy, even geopolitics… these are all inflationary", and "trying to move away from precisely measuring a particular level of inflation." His warning to the doves: despite all the hand-wringing about higher yields, "the Bloomberg Financial Conditions Index has eased. So the Fed is not tightening. They're actually… adding more fuel to the fire."

The tone across the desks was mild exasperation. On Saxo Market Call (Aug 12), John Hardy passed along strategist James Aitken's take that Warsh "really flubbed the communication" at the last meeting, making Jackson Hole "a key… chance for Warsh to get a more coherent and cogent communications policy." And on RenMac (Aug 14), Neil Dutta put the silence-strategy best: "it's definitely a bull market… for Fed watching by design. So, thank you, Kevin Warsh, for that. By being so circumspect around his own views and letting so many disparaging voices fill the void, you know, he's made my voice a little bit more important."

The rate debate blew wide open

A month ago the range of opinion was narrow. This week it stretched from "three hikes" to "get ready for cuts", and that uncertainty is itself the story.

Start at the hawkish end. On CNBC's Fast Money (Aug 12), Bank of America's head of U.S. economics, Aditya Bhave, refused to blink on his out-of-consensus call for three rate hikes. His logic is almost arithmetic: "the unemployment rate is down. Inflation is up and further away from target. And yet the policy rate is 75 basis points lower than it was. So that doesn't make sense to us. We think they need to take back those 75 basis points of cuts." His scary scenario is about the long end of the bond market, the 10- and 30-year yields that set mortgage and corporate borrowing costs. "If they don't hike now and the data turn hawkish again early next year, they're facing a very serious risk that the long end could get unanchored", meaning yields spike on their own because investors lose faith. He even did the math on where inflation really sits: strip the temporary stuff and "if everything goes right for the Fed and nothing goes wrong… you're still at 2.5. That's a pretty big overshoot when the labor market is looking basically in equilibrium." And on the politics of it, he was blunt: they won't start in October, "that's too close to the midterms," so "they either start in September or start in December", and if they go, "75 is totally reasonable because you need to do more than just a symbolic hike."

RenMac's Dutta lands close to Bhave on the near term. On RenMac (Aug 14), he said flatly, "I think the bond market is underappreciating the risk of a hike in September." His reasoning: core inflation has run above 2% "for really six of the seven months of this year," and a plain rules-based model "is saying hike." But he laid out a fascinating twist, the doves might win the battle and lose the war: "if they don't get it done by September, they probably don't get it done at all for the year," but of the "nine people that want to hike," six see multiple hikes, and "maybe they shave down to one for the year and they start pushing more hikes into 2027." In other words, a dovish 2026 could set up a more aggressive 2027.

Now the dovish, and more colorful, end. On the RiskReversal Pod (Aug 14), investor Vincent Daniel described Warsh as having been "dealt 2-7 offsuit", the worst hand in poker. The economy is "hotter than people think, mainly because… government spending" (with deficits running 4-6% of the economy) and a war adding "annoying inflationary pressures", all of which argues for tightening. The catch: "you really can't tighten that much because more and more of our interest expense is moving towards bills. Interest expense is now bigger than the defense budget." So what does Warsh do? "He is in prayer mode… praying for the immaculate economy", "a strong economy, a slightly weaker dollar, and lower rates… and somehow, some way lower inflation so he doesn't have to raise rates." Daniel's kicker cuts through the whole "is Warsh a hawk" debate: "if you actually just go across the street and look at what the Treasury is doing, it's wide effing open and they are printing money like nobody's business. Dollar swap lines here, budget deficits there, yen intervention here." His bottom line: "I really don't think the Fed's going to raise in September."

That dovish camp has real macro logic behind it, not just wishful thinking. On Real Vision's Macro Mondays (Aug 17), Andreas Steno Larsen laid out four reasons inflation should keep fading into year-end: energy prices have peaked after the spring's Middle East spike and "will turn outright negative into early next year"; companies "are currently getting paid by the U.S. Treasury to stand pat on prices" through tariff rebates, so they're not passing costs through (consumer-goods wholesale prices actually "fell… roughly 75 basis points on the month"); the World Cup's hotel-price surge is "mean-reverting" back down; and tighter immigration is quietly cooling rents. Put together, he argued, that's "a Goldilocks setup" that will leave the Fed "wrong-footed" and force a dovish repricing. Even the ratings-agency economists are edging that way: on Moody's Talks – Inside Economics (Aug 14), the panel put a 25-35% probability on a rate cut by early 2027, against a market pricing essentially zero, on the view that recession risk is being underrated.

And there's a genuinely different framing worth keeping in mind. On The Higher Standard (Aug 18), Chris Naghibi argued the whole hike-or-cut fight misses the point: today's inflation is "cost-push" (driven by energy and geopolitics), not "demand-pull," so "the Fed cannot effectively combat it with rate hikes", raising the specter of a "1973-style stagflation," where prices rise even as growth stalls, and rate policy is the wrong tool either way.

The wildcard: Warsh could change how inflation is measured

Here is the most under-the-radar idea of the week, and it comes from an actual inflation-market specialist rather than a pundit.

On WSJ's Take On the Week (Aug 16), Texas A&M finance professor Jill Satina explained why real yields on inflation-protected bonds jumped around the July Fed meeting, and offered a reason most people missed. Warsh has launched five task forces, and one is looking at economic data, including "CPI methodology." That matters because "TIPS", Treasury Inflation-Protected Securities, bonds whose payout rises with the official inflation index, are only as good as that index. If the government were to redefine how it measures inflation downward, TIPS holders would get paid less.

The precedent is real: "there is a historical precedent dating back to 1995 when Chair Greenspan spoke about" CPI being overstated, which led to the Boskin Commission and "methodological changes to U.S. CPI that downwardly revised the U.S. CPI computation by 1.1 percentage points on a… ongoing basis." Satina's telling detail: "I was unable to find in my lifetime any examples of CPI methodology changes that had resulted in upward revisions of CPI. I could only find evidence of methodological changes resulting in downward revisions."

If that sounds like a technicality, follow it through. A lower official inflation number would make the Fed look closer to its target, give it cover to keep policy easier, and hand the Treasury a cheaper inflation bill on its protected bonds. Satina also underlined why the Fed feels boxed in regardless: financial conditions are "the easiest that they've been in the last decade," and the Treasury's habit of funding itself with short-term bills, now "over 20 percent" of the debt stock, is effectively "a Treasury form of quantitative easing" that quietly pressures the Fed to keep rates low. It's the same plumbing that keeps showing up in this newsletter, seen from the bond desk.

The dollar itself: real slide, or a mirage?

The dollar index has slipped under 100, and this week produced the cleanest two-sided read on whether that continues.

The "it keeps sliding" case came from FX veteran Marc Chandler on The KE Report (Aug 14). His starting point is that the dollar tracks short-term U.S. rate expectations, and "it looks like the U.S. economy is hitting a soft spot", jobs actually fell in July, inflation was soft, and retail sales came in negative. On the charts he sees a "bear flag", a pause before another leg down, with the next support around 99.20 (a level "the dollar index has not traded below… since the middle of May"), then closer to 99 and 98.75. But he was refreshingly honest about conviction: "the momentum indicators are stretched, oversold," and after the July jobs report "the dollar got sold, but there was no follow-through." His caution: "this might be the end of the boxer's punch rather than the beginning of it." He put the priced odds in perspective: at the end of July the market saw 18 basis points of September tightening; after the soft data, just 8, "about a one in three chance."

The pushback was the smartest contrarian point of the week. On Wealthion (Aug 18), David Woo described how crowded the dollar-bear trade has become, "a lot of people, especially non-Americans… think that this is a moment to sell the crap out of the U.S. dollar" on the view that America is losing leverage abroad, and then asked the question that deflates it: "gold ultimately is a dollar trade… if the euro is very weak and the Japanese yen is very weak, how much lower can the dollar fall?" His point is mechanical: you can't sell the dollar without buying something, and its two biggest rivals are also weak. He added that the recent gold rally "has been driven by the news of Chinese buying" more than by dollar weakness itself. (For what it's worth, Woo is short the yen, long oil, and thinks bonds are "very, very cheap.")

The yen: a September hike is now the base case, but will it matter?

The near-term story has flipped fast. Two weeks ago a Bank of Japan rate hike was barely on the table; now it's close to consensus. The fresh debate is whether it changes anything.

On The MUFG Global Markets Podcast (Aug 14), Derek Halpenny, who admits MUFG was "one of the most hawkish forecasters" on Japan, noted the market now prices "around sort of 20 basis points of hikes" for September, "which even a couple of months ago would have been seen as… very unlikely," helped by "much more hawkish rhetoric from BOJ officials, including Governor Ueda." Yet the yen "quickly [gave] back around half of its intervention-driven gains." The official CFTC positioning data (which tracks speculators' bets) showed "a sharp unwind of short yen speculative positions" right after the late-July rescue, but "leverage funds" have been "cautiously rebuilding those short yen positions" since. His conclusion: intervention "on its own really is just… buying time," and a lasting yen rally needs the fundamentals to change.

Marc Chandler added the most revealing detail on why the rescue faded. On The KE Report (Aug 14), he explained that even though Japan sold "roughly $80 billion" and the U.S. maybe "$10 to $15 billion" to prop up the yen, in the very next week "Japanese investors bought the most amount of foreign bonds and stocks… in two years." In plain terms: Japan's own big institutions used the stronger yen as a chance to buy cheaper foreign assets, "the deeper pockets, the real money, seem to be taking advantage of that bounce in the yen to sell into it." So the intervention squeezed the speculators but handed the real money a better entry to keep sending savings abroad. He also noted the yen's weakness is "exaggerated", down only 3-4% on the year despite 40-year lows.

JPMorgan's FX team, on At Any Rate (Aug 14), held a bearish 164 target, noting the market already prices "just over three hikes for the BOJ over the next year." Their sobering point: even that may not be enough, "should 2% BOJ [rates] be enough… I'm not sure that that's the case," because a 2% Japanese rate is already "in the price." What could actually turn it, they argued, is a "sea change in psychology", an unambiguous signal that Tokyo will let the BOJ tighten, backed by real domestic money flows, like reallocating Japan's giant pension fund (GPIF) or new tax incentives to get Japanese savers and insurers to buy home-market bonds. Their honest coda: "we will see it to believe it."

And the sharpest skeptic was economist Steve Hanke on The David Lin Report (Aug 18). He'd called the rescue a "dead-cat bounce," and it played out: the yen has "fallen about halfway back" toward its pre-intervention level. His diagnosis is that the yen problem isn't the interest-rate gap at all, it's that "the money supply is growing very slowly in Japan," which means slow nominal growth, which means a weak currency. "The only way you can change the course of a currency is to change the fundamentals. And they're not changing the fundamentals of anything." His prediction: "they're going to end up having to intervene again."

One fresh plumbing detail tied it together this week. Several podcasts, including Many Happy Returns (Aug 19) and Goldman's own Exchanges (Aug 13), explained the emergency tool letting Japan raise dollars without dumping its Treasuries on the open market: the Fed's FIMA repo facility, where a foreign central bank pledges its U.S. bonds as collateral for short-term dollar loans. And on Money Metals (Aug 12), the host highlighted an even more telling wrinkle: the U.S. Treasury contributed to the rescue by selling euros to buy yen, "thus limiting the volume of dollar sales needed by the Japanese authorities", because, quoting Berkeley economist Barry Eichengreen, "the U.S. is reluctant to see foreign central banks use dollar reserves due to the potential ramifications for its own financial markets." Which leads straight to the week's quiet blockbuster.

The quiet story: the race into gold

If the dollar's day-to-day is a Warsh story, its long-run is a reserves story, and this week that theme, thin in recent weeks, came roaring back with actual numbers. A note on who's talking: some of the loudest voices here are gold dealers and newsletter writers talking their own book, so treat the framing with a grain of salt. But the core data points showed up on mainstream markets podcasts too, and the underlying logic came from a Berkeley economist.

The number to remember came from InvestTalk (Aug 19): "the dollar share of global reserves has fallen from 71% in 1999 to around 54% today," and "gold holdings on central bank balance sheets has now exceeded U.S. treasuries." The mechanism they described is a debt-and-sanctions story, not an apocalypse: with U.S. deficits running 6-7% of the economy while it grows just ~1.5%, "the market is starting to price in that debt spiral. And so are central banks around the world," who are letting Treasuries mature and "buying gold to protect them from U.S. sanctions as well as fiscal instability. Because gold is a liability-free asset… It can't be printed or defaulted." Their tart line on the much-hyped "BRICS currency": "You know what that is? It's gold."

The single most striking data point came from wealth manager Clive Thompson on ITM Trading (Aug 14). China's gold buying has gone vertical: "they were buying 30 to 40,000 ounces per month" a year ago, then it "stepped up… to 160,000 ounces" in February, "260,000 ounces" in April, "320,000 ounces" in May, "480,000" in June, and then "the astonishing figure… 640,000 ounces of gold that they bought" in July. His read isn't Armageddon; it's catch-up plus insurance: China's gold is still only "8% of their reserves… whereas most of the major Western countries have gold reserves of 60, 70%," and after Russia's dollar assets were frozen in 2022, "if they fear that there might be… sanctions down the line… it would make sense to hold other things." He also passed along Senator Rand Paul's Fort Knox line that the dollar "has lost 85% of its purchasing power" since 1971.

The intellectual anchor for all of it was Eichengreen's argument, relayed on Money Metals (Aug 12): the yen rescue quietly revealed that "the dollar is not the attractive reserve currency it once was," because Washington doesn't want big holders freely selling their Treasuries. "Why in the world would you hold dollars as your reserve currency if you can't use them when you need them?" The answer, increasingly, is gold, a point echoed on Money Tree Investing (Aug 14), which noted central-bank gold buying has "doubled from 500 tons annually" a decade ago to "1,000 tons annually" now.

Two more colorful threads, clearly labeled as such. On the gold-industry side, Andrew McGuire on Kinesis Money (Aug 13), a gold platform, so a promotional voice, argued the "reserve race has started," with the People's Bank of China "shifting reserves from London to Hong Kong" and building a gold-referenced yuan settlement corridor to let trading partners move money "outside the dollar-controlled rails." And on the political side, Chris Markowski's Watchdog on Wall Street (Aug 18) seized on a genuinely notable comment: the Vice President "being interviewed, talking about, it's not a good thing that the United States has the world reserve currency status, and maybe that should go away." Markowski's worry is practical, not gold-bug: reserve-currency status is what gives Americans cheap, plentiful access to capital ("Why do you think all of these AI companies set up shop in the United States? … it's where the money is"), and losing it "will change the complete dynamic of our country."

The debate

Does the Fed hike, hold, or cut? The range of respected views is unusually wide this week.

Hike, maybe a lot: BofA's Bhave holds three hikes because policy is "75 basis points" too easy with unemployment low and inflation above target, warning the long end could "get unanchored" if the Fed waits (CNBC's Fast Money, Aug 12). RenMac's Dutta thinks "the bond market is underappreciating the risk of a hike in September" (RenMac, Aug 14).
Hold, drifting dovish: MUFG argues the soft data give the Fed "more leeway to leave rates on hold," which "leaves the potential… for the dollar to re-weaken as we head into year end" (MUFG, Aug 14). RiskReversal's Daniel: "I really don't think the Fed's going to raise in September" (RiskReversal Pod, Aug 14).
Cut, on the tail: Moody's puts 25-35% odds on a cut by early 2027 against a market pricing zero, citing recession risk (Moody's Talks, Aug 14); Steno's four-factor disinflation case points the same way (Real Vision, Aug 17).
The referee: Jackson Hole on Aug 28, the July minutes, and the run of CPI/PPI/PCE prints before the September 16-17 meeting.

Is the dollar's slide real or a mirage?

Real, one more leg lower: Chandler sees a "bear flag" and a path to ~99.20, on a softening U.S. economy (The KE Report, Aug 14).
Mirage, you can't sell it against what? Woo: "if the euro is very weak and the Japanese yen is very weak, how much lower can the dollar fall?" (Wealthion, Aug 18). Even the dollar bear Chandler concedes momentum is "stretched" and "this might be the end of the boxer's punch."

Will a Bank of Japan hike fix the yen?

Yes, eventually, if fundamentals shift: MUFG and JPMorgan see a September hike as the base case and a genuine (if slow) support, provided Tokyo unshackles the BOJ and domestic money flows help (MUFG, Aug 14; At Any Rate, Aug 14).
No, wrong diagnosis: Hanke says the driver is Japan's slow money supply, not the rate gap, so expect "another… intervention" (The David Lin Report, Aug 18); Chandler shows Japan's own real money is selling yen into every bounce (The KE Report, Aug 14).

Is de-dollarization real, or gold-bug noise?

Structural: Eichengreen's point, a reserve asset you can't freely use is a less attractive reserve asset, plus the hard number that the dollar's reserve share has fallen to ~54% and gold now tops Treasuries on central-bank books (Money Metals, Aug 12; InvestTalk, Aug 19).
Overstated: it's a slow bleed, not a break, the dollar is still 54% of reserves, China's gold is only 8% of its own, and, as Woo notes, there's no scalable rival currency to rotate into. Much of the loudest commentary comes from gold sellers with an obvious incentive.

The trades in play

Only where a podcast named an actual position:

Short the yen / long oil / long bonds, and looking to short AI. David Woo said he's "been short the Japanese yen over the last two weeks, ever since they started intervening," thinks "the yen is doomed," is "long oil," finds bonds "very, very cheap," and wants "better levels" to short AI (Wealthion, Aug 18). (An individual macro investor's book.)
Short the dollar into Q4. Steno Research reiterated a weaker dollar as its "highest conviction" call and the "final piece" of a Goldilocks setup (Real Vision, Aug 17). (Independent macro shop's book position.)
Bearish yen (target ~164), and prefer the Swiss franc as a funding currency over the yen. JPMorgan's FX desk sees dollar-yen pressure persisting unless the Fed cuts, and favors the Swiss franc to fund carry trades because "you're not… countering the domestic policy biases there" (At Any Rate, Aug 14). (Bank FX desk view.)
Own gold as the reserve-shift trade. The gold-industry voices are unabashedly long, and even mainstream desks note central-bank buying as a structural pillar under the price (ITM Trading, Aug 14; InvestTalk, Aug 19). (Treat the gold-dealer framing as talking-its-book; the central-bank-flow data is the durable part.)

Read-throughs

Aug 28 is the whole ballgame for the dollar's next month. With the Fed split 9-3 and Warsh deliberately quiet, his Jackson Hole speech is the market's first real chance to learn the rulebook. If he clarifies a credible reaction function, the term-premium worry (and some of the pressure on long yields) eases; if he stays cryptic, the "bull market for Fed watching", and the volatility, continues. Watch the July minutes first as a warm-up.
The debate has widened, which is itself a signal. When serious desks range from "three hikes" to "cuts on the tail," conviction is low and positioning is fragile, small surprises will move the dollar more than usual. The honest base case from the bank desks (MUFG, JPMorgan) is hold-in-September, which points to a softer dollar into year-end unless the data re-accelerate.
On the yen, watch 160 and the fundamentals, not just the headlines. A September BOJ hike is now largely priced, so the hike itself may be a "sell the news" event. The durable tells are whether Tokyo signals it will truly let the BOJ tighten (and back it with domestic money flows), and whether the U.S. has to lift the FIMA facility or intervene again. If Hanke and Chandler are right that the driver is structural, expect another round of intervention rather than a clean fix.
The reserve-share number is the slow story that matters most. A dollar at ~54% of reserves (from 71%) and central banks holding more gold than Treasuries won't move a currency this week, but it is the backdrop that makes every U.S. fiscal headline heavier and gives foreign capital a reason to ask for a little more to hold American assets. A sitting Vice President musing that reserve-currency status "should go away" is worth filing, even if markets shrug for now.
A midterm risk premium is being named, but not yet priced into the dollar. The clearest link came from commentator Gerald Celente, who argued the administration will push for lower rates into November's midterms, which "will weaken the dollar" (ITM Trading, Aug 12); Bhave's point that the Fed "won't insert themselves… six days before the midterms" says the same thing from the other direction. No one is pricing a clean dollar consequence, but the political calendar is now shaping the rate path, which shapes the dollar.
Stablecoins stayed in the background this week, with two small forward markers: a broad bank-and-network consortium (Visa, Mastercard, Coinbase, Stripe, Bank of America, ~140 firms) building an open stablecoin project (The Insurance Guys Podcast, Aug 12), and mounting pressure to pass the crypto market-structure "Clarity Act" before the midterms, with betting-market odds "plummeted… to 10%" (CRYPTO 101, Aug 18).

What changed this week

The calendar took over. The conversation shifted from parsing each data print to waiting on Warsh's Jackson Hole speech (Friday, Aug 28) and the July minutes, the first hard chances to learn his reaction function.
We learned just how split the Fed is: a 9-3 vote to hold in July, the widest division in roughly two decades, which is exactly why one speech now carries so much weight.
The rate debate stretched to both extremes, a live "three hikes" call from BofA on one side, 25-35% cut odds from Moody's on the other, replacing the narrower hikes-only argument of early summer.
A genuinely new wildcard surfaced: Warsh's task force could revisit CPI methodology, with a 1995 precedent that cut measured inflation by 1.1 points, a lever that would flatter the target and pressure inflation-protected bonds.
The September Bank of Japan hike moved from long shot to base case (about 20 basis points priced), even as skeptics like Hanke argued it won't fix a yen weakened by slow money growth.
De-dollarization got its numbers back: reserve share down to ~54% from 71%, gold now exceeding Treasuries on central-bank books, China's monthly gold buying up more than tenfold in a year, and a Vice President openly questioning whether reserve-currency status is even worth keeping.

Levels referenced are approximate, from mid-August U.S. sessions: the dollar index just under 100 (down from ~101.60 before the late-July yen intervention), with chart support watched near 99.20; dollar-yen back around 159.5, roughly halfway back from a post-rescue low near 155-156, with 160 the widely-watched pivot; the July U.S.-Japan rescue involved roughly $80bn of Japanese bill-selling plus an estimated $10-15bn from the U.S. (partly via euro sales); the Bank of Japan's September hike about 20bp / two-thirds priced, with the market pricing just over three BOJ hikes over the next year; the U.S. September decision on Sept 16-17, with markets pricing roughly a one-in-three chance of a September hike (down from 18bp of tightening at end-July to about 8bp), and the July FOMC minutes (a 9-3 hold) due this week; Warsh speaks at Jackson Hole Friday, Aug 28, ~10am ET; headline CPI ~3.4% and core ~2.5% (July), with core PCE running near 2.8-3.3%; the 10-year Treasury yield in the mid-4% range (near multi-year highs) and the 30-year above 5%; the July federal deficit a record ~$432bn; the dollar's reserve share ~54% (from 71% in 1999); China's July gold purchase ~640,000 oz.