Newsletter · · Ashutosh Agarwal

Payrolls Went Negative and the Rate Hike Talk Faded - US Macro Recap - Week of August 11, 2026

A US Macro Recap for the week of August 11, 2026: a surprise negative July payrolls print collapsed the odds of a September rate hike, the break-even jobs reframe took center stage, and everything now hinges on tomorrow's CPI as money rotated into gold, silver, and copper.

US Macro Recap

Week of August 11, 2026: Payrolls Went Negative and the Rate Hike Talk Faded

On Friday the jobs report did something it hadn't done since February: it went backwards. Employers cut 23,000 jobs when forecasters expected 80,000 added, and the two prior months were revised down by roughly another 100,000. In a single print, the coin-flip we flagged last week landed, and the case for a September rate hike quietly deflated with it. Now everything rides on tomorrow's inflation number.

TL;DR

  • The jobs report was soft across the board. Payrolls fell 23,000 against about +80,000 expected, prior months were cut by roughly 100,000, and wages rose just 0.1% on the month (3.2% over the year, the weakest in years). The one catch: unemployment actually fell to 4.1%, but for the "wrong reason," because people are leaving the workforce, not finding jobs.
  • The debate now turns on one number almost nobody was discussing a month ago: the "break-even" payroll level, how many jobs the economy needs each month just to hold unemployment steady. With immigration slowing the inflow of new workers, the White House says that number has dropped from about 120–130,000 a few years ago to roughly 40,000 now, and some Fed staff estimates are near zero. That reframing turns a scary negative print into something closer to noise.
  • Odds of a September hike collapsed from near-certain to a coin flip or worse, and it now hinges on tomorrow's CPI. Forecasters see a soft reading (around 0.2% on core). A couple of hot prints could still put a hike back on the table; anything tame likely takes it off.

What's New

1. The July jobs report went negative, and Wall Street cheered it. (Operators and economists reacting live.) The headline: payrolls −23,000 versus +80,000 expected, with the prior two months revised down by roughly 100,000 and wages up a "scant tenth of a percent." On Bloomberg Surveillance "Instant Reaction" (August 7), economist Claudia Sahm (New Century Advisors) said the report reminded her "exactly" of a year ago and cautioned "there is signal here" but "probably a fair bit of noise and some seasonal issues" from the summer school calendar, while Constance Hunter (EIU) was blunter: "you saw the unemployment rate fall for bad reasons, not good reasons... This is definitely warrant to hold." The bear reading was starker still. On Eurodollar University (August 8), Jeff Snider argued the "resilient labor market" is a statistical mirage: the household survey shows employment "down by over 900,000 just since January" and the labor force down "1.4 million just since January," so once you hold participation steady the real count of unemployed is "closer to 11" million versus the official ~7 million. Why it matters: the report the whole rate-hike debate was waiting on came in weak, and, crucially, the stock market read "weak jobs" as "no hike," and rallied to a record.

2. The reframe of the week: a negative payroll print may no longer be a warning sign. (Official + bank research.) This is the piece that changes how to read Friday's number. On Bloomberg Businessweek (August 7), White House economic adviser Kevin Hassett spelled out the math: because participation is falling, "the break-even jobs number, that is, the jobs number you need so that unemployment rate doesn't go up, has gone from maybe 120, 130,000 a few years ago to maybe about 40,000 now." Deutsche Bank's Matt Luzzetti, on the same show, went further: "there's various Fed estimates out there suggesting that the break-even number could be as low as zero per month." Bank research echoed it. On Global Research Unlocked (August 7), Bank of America's Aditya Bhattacharya said his advice to the Fed would be that "breakeven job growth is very close to zero. We will get a few negative months per year if breakeven job growth is close to zero. It should not shock you," and noted the unemployment rate is "two tenths below where any of them thought it would be by the end of this year." Why it matters: strip out the shrinking labor force and the "shocking" −23,000 looks a lot less shocking, which is exactly why the market didn't panic.

3. Everything now hinges on tomorrow's inflation print, and the nowcasts say it will be soft. (Pundit forecasters and bank economists.) On Macro Mondays (August 10), Real Vision's Andreas Steno said his nowcast has headline CPI at "0.1 on the month and the core inflation is 0.2," below the market's 3.4% year-over-year expectation, because "the tariffs being paid back to corporate America" have removed the urgency to raise prices, and "the energy shock is over." His conclusion: "our data probably tells the Fed that it should cut, not hike." Nomura's Aichi Nemeya, on Nomura – The Week Ahead (August 7), put a number on the same view: core CPI around 0.18% and core PCE (the Fed's preferred gauge) at "0.211% in July," consistent with the Fed staying on hold "through the end of 2027." One important wrinkle: on The Morning Market Briefing (August 7), the hosts walked through how the government is about to change the way it calculates core PCE, re-benchmarking software prices (worth roughly a 10 basis-point reduction) and computing portfolio-management fees off what people actually pay rather than off asset values. Those changes hit August data in mid-September and should mechanically lower reported inflation, though the move is already politically charged, with Senator Elizabeth Warren "summoning all of her people to have an inquest into the Fed on how they're changing the PCE calculation." Why it matters: the labor number took the hawks' fuel away; tomorrow's CPI is the swing vote for September.

4. AI is now roughly half of all business investment, and the bill is starting to show up in the credit market. (Strategists and allocators.) On Excess Returns (August 8), David Rosenberg said "about 50% of business capex is AI and AI related," growing "like 18%" in real terms, while "the ex-AI capex is running negative year over year," so "without the AI boom, we probably would be in a recession." He expects the credit market, not the Fed, to end the party: "the credit market will lead the ultimate rolling over of this AI trade," pointing to CDS spreads (the cost of insuring company debt) "widening out pretty dramatically." Morgan Stanley's Mike Wilson made the same structural point on RiskReversal Pod (August 7): the buildout is bigger than the 1990s internet boom and there "will be a massive hangover... malinvestment... disappointment," but we're only "mid-innings." The tell he's watching is funding: "up until about six [weeks ago]... it was all equity funded. Now it's being debt funded... the clock has now started," so "when the credit markets are choking" that is the early-warning sign. You can already see the flood: on Key Wealth Matters (August 7), KeyBank's Rajeev Sharma noted the five largest hyperscalers have issued about $190 billion of bonds in the first half of 2026 alone, versus roughly $100 billion in all of 2025, helping push his full-year investment-grade issuance forecast toward "close to $2 trillion." Why it matters: the single engine holding up growth is now being financed with debt, right as the long end of the bond market sits near multi-decade highs.

5. The consumer is still split in two, and the bottom half is where the credit stress is. (Strategists and an operator.) On Excess Returns with Tian Yang (August 6), the framework was crisp: lower-income consumers "face poor wage growth and weak personal finances," with New York Fed delinquency rates "very high," while upper-income households stay strong and carry overall growth, a split he called "dystopian." His most striking chart shows Americans saying jobs are easy to find yet their personal finances are struggling, "a gap... unprecedented in the history of the data," explained by price levels resetting faster than real incomes. The bigger picture came from Viktor Shvets on Macro Voices (August 6): the wage share of national income is "now at the lowest point since 1947" while the profit share is "at the highest level ever... about 16%, 17%," and the top 0.1% of households, just 135,000 of them, average "almost $200 million" in net worth and control "almost 15% of national wealth," up from 8% in the 1980s. Even the AI-driven equity crowd sees the same fault line: on Squawk on the Street (August 7), Trivariate's Adam Parker argued the K-shape "is maybe a balance sheet thing, not an income statement thing," the low end doesn't own the stocks that are ripping, but "the unemployment rate's really low... they're paying their 90-day credit card [balances]." Why it matters: the same bifurcation that keeps inflation contained (the bottom can't absorb price hikes) is also what makes the expansion fragile if the top ever blinks.

The Debate

The week's podcasts support all three camps we've been tracking, but the jobs print handed a clear win to the "leave rates alone" side.

Camp A: Hold, and probably done hiking (soft landing). Supported, and it dominated. The steel-man got stronger by the hour after Friday's number. J.P. Morgan Asset Management's David Kelly, on Squawk on the Street (August 7), summed it up: "American business can't find workers and American workers can't find a raise... the labor market's not strong, it's just tight," which is why we have "Teflon inflation" and "the Fed is absolutely right not to raise rates right now." Add Sahm's and Hunter's "warrant to hold," Steno's "should cut, not hike," Nomura's on-hold-through-2027 base case, and the break-even reframe, and the near-term data all points one way: no hike is needed, and the bond market is doing the tightening anyway.

Camp B: Growth is hollow beneath the AI boom (stall speed). Supported. David Rosenberg on Excess Returns (August 8) argued four-quarter real GDP is tracking about 1.4% versus potential of 2.0%, "the demand side of the economy is flowing below the potential supply," which is disinflationary, with real disposable income "fractionally negative year on year" and "all the growth on the income side... coming out of corporate profits." Snider's collapsing labor force and Jared Bernstein's "real fragilities under the hood" (see Read-throughs) sit in this camp. The steel-man: strip out AI capex and the economy is already contracting in housing and autos; the extra spending is borrowed; and when the hyperscalers decide to save, the income loop unwinds.

Camp C: The Fed is behind the curve and should be hiking (reflation). Supported, but a minority. Rich Bernstein, sharing the mic with Rosenberg on Excess Returns (August 8), noted every one of the Atlanta Fed's "30 different versions" of the Taylor rule "suggest that the Fed should be hiking rates," and argued the AI bubble is "inherently inflationary" because it grossly misallocates capital, starving housing (where builder stocks are in a bear market) while flooding data centers. On Bloomberg Daybreak: US Edition (August 7), Bloomberg's Edward Harrison warned "we're not on the disinflationary track... we've stopped disinflating," pointing to super-core services PCE at 3.8%, "well above the baseline." The steel-man: underlying services inflation is sticky, nominal demand is hot, and if the soft print is a head-fake the Fed is dangerously late.

The tell: all three camps can cite real numbers because the economy genuinely runs at two speeds and the inflation signal is genuinely noisy. But after Friday, the burden of proof shifted hard onto the hawks, and tomorrow's CPI is the first clean chance for them to make the case.

The Trades in Play

Positioning was unusually concrete this week, and it split cleanly: bond desks quietly adding duration on the soft data, and a broad rotation into hard assets: gold, silver, and copper all had a blockbuster week (gold up roughly 7%, silver near 10%, copper to record highs) as the "no hike" read weakened the dollar.

Operators (actually positioning money):

  • Adding duration and leaning on high-grade credit (KeyBank). On Key Wealth Matters (August 7), Rajeev Sharma walked through the post-print bond rally, the 2-year snapping down to 4.18%, the 10-year to 4.63%, and reiterated a preference for investment-grade corporate bonds over Treasuries "in this kind of rate environment," on a "higher for longer, prolonged Fed pause" view.
  • Rates go higher from here, so stay short duration (Citi Wealth). On CNBC's "Fast Money" (August 7), Citi Wealth's Alalu Agonga said that even after a weak report, "our view is that the next direction for rates is higher," and that "the break-even payroll trends have actually been relatively stable," so the desk leans to shorter duration and is broadening beyond the Mag-7 into cybersecurity, financials, gold, and natural resources.
  • Gold and silver as a delta-one trade, not options (macro fund managers). On The Macro Trading Floor (August 9), Alfonso Peccatiello and Brent Donnelly said the January silver/gold blow-up has fully cleared, "nobody gives a crap anymore about gold and silver," positioning is clean, so gold "relit 4%" the instant the Fed-won't-hike and Hormuz-reopening headlines hit. Their advice post-bubble: buy the metal outright with a stop, don't pay up for options. And notably, they see the real "animal spirits" not in the fad trades but in "value stocks... Japanese banks... European banks" (European banks +22.5% year-to-date). On duration, they flagged that shorting the long end costs you: the 30-year near 5.2% against front-end cash near 3.5% is "negative carry," so you need a real catalyst, and their models see "no major inflation impulse."
  • Long the AI complex and the reflation hedges (fund manager Jordi Visser). On The Pomp Podcast (August 8), Visser said aggregate payrolls (hours × wages × headcount) are "at the weakest level since 2012 on a six-month rolling basis, except for during COVID," which convinces him the Fed is out of tools and must reflate. He's positioned accordingly: a "good-sized position in NVIDIA" plus optical names and Marvell and Micron on the AI side ("the lows are in for the AI names"), and long gold, silver, platinum, and Bitcoin as the reflation hedge.
  • Copper over gold and silver (portfolio manager Nick Hodge). On The KE Report (August 7), Hodge said copper is his top call of 2026, trading around $673/lb with "July the highest monthly close copper has ever put in," as U.S. COMEX warehouses pull in metal on tariff uncertainty and Codelco output keeps "stair-stepping down." He bought Ero Copper ("up like 15% in... six days"), sold HudBay and rolled it into Arrow (Brazil copper), and holds Cotec (rare-earth recycling) and Energy Fuels (uranium/critical minerals).
  • Buying commodity equities into the summer dips (Olive Resource Capital). On Company Interviews (August 8), CIO Sam Peles said the firm used July's volatility to deploy capital into oil & gas producers, uranium, copper/base metals, and gold names, making Prospector a top-10 holding, on the view that "commodity weakness represents buying opportunities." He also flagged that refining margins in the U.S. are running "up... fourfold over the averages," a boom for refiners even as crude chops around.
  • Long silver and gold on a proprietary signal (Gold Fix trader Vince Lanci). On Kontrarian Korner (August 10), Lanci said he got long silver on Friday and took off half the position Monday, and sees gold reaching about $4,500 (its 200-day average) but "not new all-time highs... until November," with a floor near $4,000 held up by relentless Chinese buying, the PBOC adding 9–10 tonnes plus "another 38 tonnes" on the OTC market in a single month.

Pundit (a view, not a disclosed position):

  • The Fed should cut, not hike (Andreas Steno / Real Vision). As above, on Macro Mondays (August 10) Steno argued the September staff projections, built in June on roughly $115 oil, now far above spot, will look "a lot softer," removing the case for a hike entirely.

Read-Throughs

  • The AI bill lands in the credit market first. The through-line from Rosenberg, Wilson, and the KeyBank issuance figures is the same: the buildout has flipped from equity- to debt-funded, so the earliest crack won't be in the stock market but in spreads. Wilson's framing, the market now "punishes" companies that announce more capex, a switch that flipped about six weeks ago, is the tell to watch. Second-order: the most rate-sensitive thing in the economy is now the debt-funded data-center boom, precisely as the long end of the bond market sits near multi-decade highs.
  • The GDP headline flatters the AI story less than it looks. On Thoughts on the Market (August 4), Morgan Stanley's Sarah Wolfe noted that national data-center investment "can actually overstate the contribution to GDP" because so much of the hardware, chips, servers, networking gear, is imported, which nets out in the accounting. Locally the multiplier is real but front-loaded: a 250,000-square-foot Virginia data center supports "more than 1,500 workers during construction" but "only about 50 full-time workers once it's operating." The team's three bottlenecks, power, people, politics, are hardening, with a projected "38 gigawatts" power shortfall through 2028 and "more than 300 local moratoria" across 40 states. On Peter St Onge Podcast (August 10), St Onge made the more bullish cut of the same 1.5% Q2 number: consumption grew 3.2%, an import surge ahead of new tariffs "knocked a point and a half off GDP," and investment is now "accounting for more than half of GDP growth," spending that "makes you richer in the future."
  • Immigration is quietly rewriting how every jobs report reads. This was the missing piece last week; this week it's central. On CNBC's "Fast Money" (August 7), Tim Seymour tied "a decline in the labor force that we've seen every month since January" directly to immigration policy, "it's working," which mechanically pushes the unemployment rate lower even as hiring stalls. Combined with the break-even math, it means a low payroll print signals far less than it used to.
  • The consumer split is showing up in real income and real behavior. On Balance of Power (August 7), former White House economist Jared Bernstein called it a "no-hire, no-fire job market... with real fragilities under the hood," with 3.2% wage growth "almost certainly below inflation" and the "labor share of national income" hitting a historic low. His read on the Fed's dilemma: the inflation that remains "is largely coming from the supply side," tariffs, the war, and "AI picks and shovels," so the question is whether the hawks can look through it or whether repeated target misses force a hike to defend credibility.

What Changed

Friday's recap ("Inflation Broke Lower. Jobs Day Breaks the Tie.") was written before the jobs number, with the July print teed up as a genuine coin flip. Four days later, the tie is broken and the story has moved:

  • The coin flip landed, downward. The read went from "nobody can call it" to a decisive miss: payrolls −23,000 against +80,000 expected, with wages soft. The market's response was the opposite of panic, record highs, because "weak jobs" now reads as "no hike."
  • The break-even reframe moved from footnote to center stage. Last week immigration featured only as an inflation input; this week Hassett's "120–130k down to about 40,000," Deutsche Bank's "as low as zero," and BofA's "very close to zero" turned the scary negative print into something the Fed was explicitly told to expect.
  • A September hike went from near-certain to a coin flip or worse. One host tracked the odds falling "102% to 54%... 42" on the print; BofA now sees roughly 30 basis points of hikes priced by year-end, with September itself about 50-50 at best.
  • The inflation-measurement fight got a date and a politician. Last week's "core PCE is overstated" argument (RenMac, PIMCO) is now concrete: the methodology change lands in August data in mid-September and should shave reported inflation, and it's drawn a threatened Congressional inquest from Senator Warren.
  • Hard assets took over the leaderboard. Gold, silver, and copper all had a standout week, and the positioning conversation shifted decisively from the crowded AI trade toward metals, duration, and value, a rotation that barely registered a week ago.