Newsletter · · Ashutosh Agarwal

Warsh Just Put the Rate Hike Back On - US Macro Recap - Week of September 4, 2026

US Macro Recap for the week of September 4, 2026: Kevin Warsh's Jackson Hole speech pushed September rate-hike odds from about 35 percent to as high as 67 percent, sitting Fed presidents joined the hawkish camp, and the biggest bond managers took the other side as payrolls fell and the consumer cracked.

US Macro Recap

Week of September 4, 2026: Warsh Just Put the Rate Hike Back On

Three days ago the loudest macro voices were telling you the September rate hike was basically dead. Then everyone actually digested what Fed Chair Kevin Warsh said at Jackson Hole on August 28, and the whole conversation flipped. This week the podcasts are full of money managers and even sitting Fed officials arguing the hike is not only alive, it may already be a done deal, all while the underlying economy quietly looks weaker than the confident talk suggests. Here's the week, sorted.

TL;DR

  • Warsh's Jackson Hole speech reset everything. In one speech the new Fed Chair told markets that inflation "is not slowing," that the central bank owns "65 months of sustained elevated inflation," and that unless he sees underlying inflation heading to 2% "clearly and at sufficient speed... we have work to do." The odds of a September rate hike jumped from around 35% before the speech to somewhere between 55% and 67% after it. Apollo's chief economist called it a 10-out-of-10, pivotal speech.
  • Real Fed officials are now openly in the hike camp. This isn't just pundits. Chicago Fed's Austan Goolsbee says he's "worried the economy is overheating" and that if data-center-driven demand keeps spilling over, "we're not restrictive enough." Richmond Fed's Tom Barkin frames a live "65-month vs. 47-plus-18-month" argument about whether the Fed has been too patient. Former policymaker Adam Posen thinks the Fed hikes 75 to 100 basis points over the next six months.
  • But the real economy keeps flashing weak, and immigration is why the jobs math is so confusing. Payrolls actually fell 23,000 in the last report, wage growth (3.15%) is now running below inflation for a fourth straight month, and the "break-even" number of jobs the economy needs each month has collapsed to roughly 30,000 because immigration has dried up. Underneath a resilient headline, the consumer is cracking, savings are gone, and one of the sharpest minds on the tape is now short the consumer outright.

What's new

1. Warsh's Jackson Hole speech: "we have work to do." (Operator / policymaker, the actual Fed Chair.) On Balance of Power (August 28), Bloomberg played the line that moved the whole market: "Responsibility for 65 months of sustained elevated inflation sits squarely with the central bank. And that's where it belongs. So here is my standard. We must be confident that underlying inflation is moving to our objective clearly and at sufficient speed. Otherwise, we have work to do." Bloomberg's Michael McKee, reporting from Jackson Hole, said the markets had "pressured him into" the hawkish tone after an underwhelming July press conference, "today he gave them what they wanted," and flagged that Warsh pointed to inflation's breadth, noting 54% of the components in the last report were still rising. Why it matters: this is the speech every other item this week is reacting to. Warsh didn't say "hike," but he said enough that the market now infers one is coming if the September 12 CPI doesn't cooperate.

2. Apollo's Torsten Slok: a "10 out of 10" pivot toward inflation. (Operator, chief economist at a roughly $1 trillion asset manager.) On The Economics Show (August 29), Torsten Slok of Apollo rated the speech's importance a 10 out of 10 and explained the mechanics in plain terms: with inflation "not 2%, it's 3.5%," Warsh made two things newly clear, that the Fed's number-one tool for an inflation problem "is interest rates, not tighter financial conditions and not the balance sheet," and that he is now putting far more weight on the inflation half of the Fed's mandate than on jobs. Slok's bottom line: "financial markets are now pricing that we will get a hike already at the next meeting in September. That was not the expectation going into the speech," the market had penciled the first hike in for December. Why it matters: when the economist for a trillion-dollar manager says a speech pulled a rate hike forward by three months, that's the buy-side repricing in real time.

3. A sitting Fed president says he's "worried the economy is overheating." (Operator / policymaker, an FOMC voice.) On Odd Lots (August 28), Chicago Fed President Austan Goolsbee, who dissented against last year's rate cuts, laid out why the hawks have a case. He decomposed past inflation as "loosely two-thirds supply and one-third demand," but warned that services inflation "is not really caused from tariffs... that's not really caused from oil prices. That's a deeper level of concern." And he made the AI-capex-to-inflation link concrete with color from his own district: "You go around the Midwest... what's the biggest problem? And they're like, the data centers are buying up all the land... You can't get an HVAC person." His warning: if that competition for workers and materials "gets out of its lane" and drives up wages and prices economy-wide, "then we're not restrictive enough." Why it matters: the data-center boom isn't just a growth story anymore, a Fed official is now treating it as a potential source of the very overheating that justifies a hike.

4. The confusing jobs data is really an immigration story. (Analyst plus operator desk views.) Two threads converged this week. On Today's Market Explained (August 28), FourStar Wealth's Chris Reardon walked through the ugly print: July payrolls "fell by 23,000... the first negative headline print of this cycle," against expectations of an 80,000 to 95,000 gain, with private payrolls up 30,000 but government down 53,000, and the prior two months revised down by a combined 100,000-plus. Wages decelerated to "0.1%... the lowest since May of 2021." But on Monetary Matters (September 3), trader Jared Dillian explained why that can coexist with a falling 4.1% unemployment rate: immigration. "Under a Biden administration, when we had like 300,000 non-farm payrolls a month... that wasn't necessarily a sign of a strong economy," he said, whereas "during Trump, when migration is neutral or even negative, maybe the break-even rate is like 40,000 or 50,000." He also revisited the recession alarm that never rang: unemployment "got up to 4.5, everybody was saying that it triggered the Sahm rule... and sure enough, the unemployment rate has come down to 4.1." Why it matters: the number the economy needs just to hold unemployment flat has quietly fallen close to zero, which is exactly why a negative payroll print no longer automatically means recession, and why Warsh feels free to focus on inflation.

5. The break-even math, spelled out by J.P. Morgan. (Operator, runs real money.) On Notes on the Week Ahead (August 31), J.P. Morgan Asset Management's chief global strategist Dr. David Kelly put hard numbers on the labor puzzle. Average monthly job growth has fallen from 88,000 (the 20-year pre-pandemic average) to "just 46,000" over the past two years, yet unemployment fell to a "4.09%... 18-month low." The reason: the labor force participation rate dropped 1.34 percentage points, which "amounts to 3.7 million missing workers," driven mostly by an aging population. His punchline for the inflation debate: with wage growth at 3.15%, "the smallest gain since May 2021 and lower than... CPI for a fourth consecutive month," "it's impossible for the economy to generate a price-wage spiral," which should "keep CPI on a downward track." His verdict on the crowd's sudden hawkishness: "markets may have been premature in now assigning a 60% probability to a September rate hike." Why it matters: the single most influential asset manager's house economist is publicly on the other side of the Warsh trade, a real fault line, not manufactured balance.

The debate

Both camps showed up this week with real evidence, so both get built out. The fight is simple: does Warsh's hawkish turn describe an economy that actually needs a hike, or is the Fed about to tighten into a slowdown? Unusually, this time the loudest hawks include actual Fed officials, and the loudest doubters include some of the biggest money managers.

Camp A, inflation is too sticky and the Fed should hike now (Fed officials and heavyweight economists). The steel-man: strip out the excuses and underlying inflation is stuck near 3%, the labor market is at full employment, and waiting only means bigger, uglier hikes later. Warsh anchors it on Balance of Power with the "65 months" line and the breadth point (54% of components rising). Chicago's Austan Goolsbee on Odd Lots adds the overheating worry and the services-inflation "deeper level of concern." Former Fed and Bank of England policymaker Adam Posen, also on Odd Lots (September 1), goes furthest: "my view is the Fed is going to hike. If they don't hike in September, they're certainly going to hike in December... six months from now, Fed funds will be 75 bps or 100 bps higher than it is now," with inflation stuck in a "three and a half to four and a half range" until then. Even the data crowd backs the hot read: on Moody's Talks, Inside Economics (August 28), Stripe's Matt Contiguglia (formerly of the Council of Economic Advisers) said that even after stripping out a wacky imputed financial-services component, which alone accounted for "40% of the increase" in July's core PCE, "core market-based PCE year over year... is still 3%," and "I would probably lean on the hawkish side." Richmond's Tom Barkin, on Odd Lots (August 31), framed the whole argument honestly: you can take the patient "47-month plus 18-month" view that the plane was landing before the tariff and oil shocks, or the impatient "it's been 65 months and it's been over" view, "and that's the argument I'm sure we're going to have." The soft spot: this camp needs services inflation to stay hot; if the labor market gives out, the case evaporates.

Camp B, the Fed is about to tighten into a slowdown (the money managers and forecasters). The steel-man: the hard data outside of inflation is deteriorating, wage growth is below inflation so a price-wage spiral is mechanically impossible, and the bond market itself says inflation isn't the problem. J.P. Morgan's David Kelly on Notes on the Week Ahead owns this side with the wage and break-even math and the "premature... 60% probability" line. Jared Dillian on Monetary Matters is blunt: "not only is inflation not a concern, it looks like we are entering a slowdown in growth," Chicago PMI "10 points lower than expected," JOLTS "terrible," ISM below expectations, yet "we have a 66% chance of a rate hike. It's a madness." And on Eurodollar University (September 2), Jeff Snider makes the market-signal case: TIPS break-even rates (what bonds imply for future inflation) and the 5-year, 5-year forward "have never been pricing 1970s-style inflation this entire time." His read is that "fragile labor market plus energy shock does not lead to inflation... it leads more likely to disinflation in the long run." The soft spot: this camp is now openly betting against both the Fed Chair and several regional presidents, a lonely place to be into a live meeting.

The referee both camps are watching: the September 12 CPI. As J.P. Morgan's Bob Michael put it on Closing Bell (August 31), Warsh "gave such a compelling argument" that the question is whether "he just talked the rest of the FOMC into hiking rates." But there's an out: if the three-month annualized core PCE comes in soft, "the three-month annualized rate drops from 3.1 to 2.4%... I would say that's a sufficient improvement." One inflation print now decides the meeting.

The trades in play

Positioning was unusually specific this week, and the split between operators and pundits is stark: the people actually running bond money are leaning toward duration and quality, betting the growth scare wins out over the inflation scare.

Operators (positioning real money):

  • "I'm insanely bullish on bonds" (Jared Dillian). On Monetary Matters (September 3), Dillian said he "moved a huge portion of my money into bonds in the last month" as a three-to-five-year trade, calling "5.2, 5.3 on 30 years... an incredible deal" and "4.7 on 10s an incredible deal." His thesis is contrarian sentiment: "people believe that deficits are out of control. People believe that inflation is out of control... The reality is that inflation is not that high." He notes the deficit is scary in dollars ("$2 trillion") but "only 6% of GDP" versus 12% in 2010 when auctions were three-times oversubscribed, the market obsesses over bond supply and ignores demand, which would flood in on any risk-off.
  • Own the front end, quality, and staples, play defense (UBS's Ulrike Hoffmann-Burchardi). On Alpha Exchange (September 1), UBS's CIO Americas (overseeing $7.3 trillion) said she likes "short duration... across all risk scenarios, whether it's rates, whether it's AI, geopolitics, that's the real ballast." She sees "risk in the back end" because the extra yield for extending to 10s or 30s doesn't compensate "in this type of environment where the pace of change is just so aggressive," and pairs it with "quality businesses... staples" whose earnings are "independent of an AI buildout."
  • Test 5%, then buy (J.P. Morgan's Bob Michael). On Closing Bell (August 31), the head of J.P. Morgan's Global Fixed Income group said the 10-year backup to 4.75% is "a continuation... the backup of real yields," with inflation expectations "anchored at about 2.3%," bond investors want compensation for "competition for capital" (sovereigns versus hyperscalers) and "policy confusion." He expects a test of "4.90, 4.95... I think we're going to test 5%," at which point "a lot of the cash on the sidelines is looking to come into the market," but only if the Fed commits to "three, four rate hikes." He sees real value now: a core bond fund yields close to 6%.
  • Short the consumer, long gold and energy (Ben Hunt). On Excess Returns (September 3), Ben Hunt described his model portfolio as "long gold... long energy... short the consumer," with small short positions in tech and financials. His logic: the one-time stimulus of tax rebates and tariff refunds is fading, savings are gone, and "I don't know how you do more stimulus for the consumer at this point." He expects the Fed and Treasury to lean into "financial repression," keeping rates artificially low and backstopping data-center financing, which is why "gold is a phenomenal investment" as long as they keep trying.

Read-throughs

  • AI capex is now the load-bearing wall under GDP, which cuts both ways for the Fed. On Excess Returns, Ben Hunt did the arithmetic bluntly: for a year of roughly "one and a half percent GDP growth... the only thing that keeps that from being a recession is keeping the AI hamster wheel going." On Macro Hive (August 28), Jon Turek argued the hyperscalers spend from a "we have to do this" position, "if they don't do it, the business could be gone," so a 50-basis-point rise in Google's borrowing cost "is not changing the calculus." He treats the "over a trillion dollars" of next-12-months capex as "a fiscal program," comparable to Germany's. The scale is corroborated on Motley Fool Hidden Gems Investing (August 28): NVIDIA's Jensen Huang projects the top five hyperscalers spend "$1.3 trillion" next year, up from about $800 billion this year and about $500 billion in 2025, a 60% jump. The Fed's dilemma: that spending is both the reason growth isn't a recession and, per Goolsbee, a potential source of overheating.
  • The financing of that boom has quietly flipped from cash to debt, and it's pushing up your mortgage rate. On The Meb Faber Show (August 31), Paul Kedrosky said data centers and AI capex have driven "over 50% of US GDP growth for the last six quarters," but that as of mid-2026 "more than half of the financing for data centers... no longer comes from internal cash flows," it's ABS, private credit and sovereigns, with "15 to 18% of the investment-grade marketplace now data center related." His warning is that this is "one of the reasons behind the bond freakout we just saw, because it's pushing up longer-term rates." Dillian made the same crowding-out point about Google's record bond issuance. In other words, the AI build is now competing with the U.S. Treasury for the world's savings.
  • Underneath "resilient," the consumer has run out of cushion. On Thoughtful Money (September 2), Stephanie Pomboy pointed to "the greatest wave of corporate bankruptcy since the global financial crisis" and private-credit assets "trading at huge haircuts to what they were marked," arguing the stress is real but "just hasn't reached that point where it devolved into a crisis yet." Ben Hunt was more visceral, citing the Dick's Sporting Goods earnings call, the stock down 30% because "people aren't buying... tennis shoes anymore," alongside "an all-time low in savings rate." Even Richmond Fed's Barkin, on Odd Lots, described consumers "de-saving" and "going into debt" to keep spending. And on Today's Market Explained, Reardon flagged that University of Michigan sentiment fell to 51, with "only 8% of consumers now expect their income to grow faster than inflation." The K-shaped split is now showing up in single-company checkout data, not just surveys.
  • The bond selloff is being widely misread, and that matters for what the Fed does next. Jeff Snider on Eurodollar University argued the spread between the 30-year and the 3-month has actually compressed, from 379 basis points in 2011 (when there was $14.3 trillion of debt) to 134 basis points now (with over $40 trillion), which is the opposite of what "Treasury rejection" would produce. His read: long rates are drifting up on yield-curve mechanics with the Fed "in the way," not on inflation fear. If he's right, a Warsh hike into a fragile labor market risks re-inverting the curve, a recession signal, rather than proving the Fed's resolve.

What changed

The last recap ("The Market Buried The Rate Hike. A Fed President Didn't.") argued the hike was mostly priced out, that core inflation was quietly drifting down, and that the money managers were on the dovish side with only a lone Fed dissenter pushing back. This week the ground shifted hard:

  • From "the hike is basically off" to "the hike is back on." Tuesday's recap leaned on mid-August audio (Rieder, Goldman's desk) saying the market priced a small September cut. This week, after digesting Warsh's August 28 speech, the same probability ran 55% to 67% for a hike, and Apollo's Slok said the market pulled the first hike forward from December to September.
  • The hawkish camp went from one dissenter to a chorus, including sitting officials. Last week's hawk was a lone regional president (Musalem). This week Chicago's Goolsbee is "worried about overheating," Richmond's Barkin is openly airing the too-patient argument, and heavyweight Adam Posen is forecasting 75 to 100 bps of hikes. The debate is no longer crowd-versus-one-guy.
  • The money managers switched sides of the argument. Two weeks ago BlackRock's Rieder was the calm-inflation anchor. This week it's J.P. Morgan's David Kelly, Bob Michael, and trader Jared Dillian carrying the dovish/growth-scare case against a hawkish Fed, a near-total role reversal.
  • Energy stayed a risk, but the framing moved to the September CPI as the single swing factor. Last week's worry was that a hot August print would reheat inflation; this week the entire hike-or-hold call collapsed onto one number, the September 12 CPI and the three-month annualized core PCE it feeds.
  • The consumer-crack evidence got sharper and someone put a trade on it. Tuesday's tell was Walmart's weak comps. This week it's Dick's Sporting Goods down 30% on an earnings call, an all-time-low savings rate, and Ben Hunt explicitly running a short-the-consumer book, the thesis moved from observation to positioning.