Newsletter · · Ashutosh Agarwal

Cheap Gas Cooled Inflation but the Fed Isn't Buying It - US Macro Recap - Week of July 17, 2026

Global macro newsletter for the week of July 17, 2026. June inflation fell to a six-year low as cheap gasoline did most of the work and the market slashed its July rate-hike odds, yet new Fed chair Kevin Warsh refused to declare victory as oil climbed back and the labor market showed cracks underneath the headline.

US Macro Recap

Week of July 17, 2026: Cheap Gas Cooled Inflation but the Fed Isn't Buying It

The number everyone was waiting for on Tuesday's recap finally landed, and it came in cool. June consumer prices actually fell, headline inflation dropped to its lowest in six years, and within minutes the market tore up its bets that the Fed's next move would be a rate hike. Then the new Fed Chair, Kevin Warsh, sat down in front of Congress and refused to celebrate any of it. That gap, a soft print on one side and a stubbornly hawkish Fed on the other, is the whole story this week, and underneath it the podcasts kept circling the same three questions: is the cooldown real or just cheap gasoline about to reverse, is the job market quietly breaking, and is the whole economy now leaning on one giant AI bet whose bills are coming due.

TL;DR

  • June inflation came in surprisingly soft: consumer prices fell 0.4% on the month (the biggest monthly drop since April 2020), the yearly rate slid from 4.2% to 3.5%, and the "core" reading that strips out food and energy actually ticked down for the first time since 2020, to 2.6%. The market's odds of a July rate hike collapsed from about 40% to 12% in a matter of hours.

  • New Fed Chair Kevin Warsh spent two days testifying on Capitol Hill and pointedly did not take the win. "It's one data point," he said, vowing a "regime change in policy" to end what he called "63 months of inflation above target." The catch everyone flagged: gas is a big reason the print looked good, and oil is already up around 20% again in July, so next month could easily flip back.

  • The bottom half of the household economy keeps showing strain even as the top holds up: seriously late student-loan payments climbed to 10.3% of balances with 2.6 million people more than 120 days behind, and the job market's growth is being propped up almost entirely by low-wage elder-care work, while a shrinking labor force and near-zero immigration flatter the headline unemployment rate.

What's New

1. The inflation print fell off a cliff, and cheap gas did most of the work. On Marketplace (July 14), host Kai Ryssdal walked through a June report that "actually fell": year-over-year inflation "down from 4.2% in May to 3.5% last month," with even core inflation easing to 2.6%. BlackRock's Gargi Chaudhuri noted the feared World Cup travel surge never showed up, airfares and hotels "haven't generated a meaningful inflationary impulse." But the report split forecasters cleanly: Jay Hatfield of Infrastructure Capital called it "very positive," while William Blair's Richard de Chazelle warned it "doesn't take a rate increase off the table for September." On Real Vision Macro Mondays (July 13), fund manager Andreas Steno Larsen had even called it two days early, "3.7% for headline and 2.7% for core," a hair below consensus. Why it matters: this was the single most-anticipated data point of the month, and it landed soft, but almost entirely because oil sold off in June, which is exactly what makes it fragile.

2. Warsh got the good number and refused to spike the football. Across two days of Capitol Hill testimony, the new Fed Chair stuck to a hawkish script. On Marketplace (July 14), Warsh told lawmakers the "63 months of inflation above target has been an unfair burden and has been a tax on the American people and businesses… That means we need a regime change in policy… And we're just getting started." He also made clear he's stepping back from telling markets in advance what he'll do: giving a projection, he said, risks the Fed "taking information that's consistent with our priors and rejecting information that's inconsistent." By day two, on Squawk on the Street (July 15), his one notable line on the soft print was a splash of cold water: "It's one data point. The reason I created a task force on data is I do not want to overread or cherry-pick data. There might be some that look at this morning's data and say, mission accomplished… That is not my view." Why it matters: with Warsh deliberately refusing to give forward guidance, the market is left guessing, and every data point now swings expectations harder.

3. The market flipped from "hike coming" to "never mind" in a single session. On RiskReversal Pod (July 15), Ben Emons of FedWatch Advisors put a number on the whiplash: "We had the hike priced for almost 40% yesterday in July. We dropped right down to 12% after the CPI number. That's how quickly that changes." He was careful to separate the good headline from the messy underneath: the -0.4% monthly drop was "a lot of energy decline," but "energy prices are up 13% or 15% on the month already… the next monthly print may actually be positive again." And he coined the week's best word for the sticky part, "funflation," his term for categories like "recreation or food away from home, or… airfares and entertainment, sports events" that are "all up significantly." Strip the report down to the categories still running hot, he said, and "CPI is actually at 18% annualized currently. Just on those." Why it matters: the market treated one soft print as a green light, but the person watching the internals says the disinflation is thin, uneven, and possibly about to reverse on gasoline alone.

4. The bond market is now openly betting the Fed is done, breakevens fell below target. A "breakeven" is just the inflation rate baked into bond prices; when it drops, the market is saying it expects less future inflation. This week it dropped hard. On Kudlow (July 15), Larry Kudlow pointed out that "the two-year CPI break-evens have dropped all the way to 1.89 percent. That's below the Fed's 2 percent target," and declared rate hikes "off the table… for the rest of the year." He also flagged a genuinely strong data point for households: "real average hourly earnings rose 0.8 percent in June. That is the best monthly real wage gain in 11 years, excluding the pandemic." On Eurodollar University (July 15), Jeff Snider read the same tea leaves far more darkly, calling core CPI turning "fractionally negative for the first time since 2020" a "marching band full of red flags", in his view the softness isn't healthy disinflation but "demand destruction," businesses unable to pass costs on because "their customers can't afford it." Why it matters: the bond market and the Fed are now openly disagreeing, the market says the inflation fight is basically won, and Warsh says don't be so sure.

5. The consumer-credit strain got a fresh, hard number. On On The Market (July 16), BiggerPockets' Dave Meyer walked through the latest household-debt data: total consumer debt hit nearly "$19 trillion in Q1 of 2026," with credit cards at $1.25 trillion, student loans at $1.7 trillion, and auto loans at $1.7 trillion. Overall delinquency is "not a crisis" at about 4.8%, and credit-card trouble has actually "flattened out" at roughly 8.6% (concentrated in subprime borrowers). But student loans are the sore spot: seriously delinquent balances (90+ days late) rose to "10.3%… up from 9.6% the prior quarter," and "the number of people who are more than 120 days past due is 2.6 million." The knock-on effect he stressed: a survey found "43% of first-time homebuyers say that student loans are their number one obstacle to saving for a down payment," on an average federal payment of about "$382 a month." Why it matters: this is the concrete, current follow-up to last week's rent-delinquency scare, the stress is real, spreading through student loans, and now visibly denting housing demand for younger buyers.

The Debate

Three distinct camps showed up in this week's podcasts. All three are worth steel-manning, because the honest read is that they're each partly right about a different slice of the same economy.

Disinflation, the fight is basically won (bond market, market-watchers). This camp got the most fuel this week. Kudlow's "Has Goldilocks returned?" framing (above) is the cheerful version: falling prices, rising real wages, breakevens below 2%, and rate hikes off the table. Ben Emons's read on the collapse in hike odds (40% to 12%) is the market-priced version. And on Rebel Capitalist News (July 16), the host detailed the bond-market reaction as a classic "bull steepener", the 2-year Treasury yield dropping about 10 basis points and the 10-year about 4, calling the June drop "the largest one-month decrease since April 2020." The steel-man: the actual data moved decisively toward target, and the market is simply pricing what it sees.

Sticky / reflation, hold and be ready to hike (the Fed itself, plus sell-side). This is the camp with the institutional weight behind it. On Bloomberg Surveillance (July 14), the show quoted Warsh's prepared line, "the members of our committee have no tolerance for persistently elevated inflation", and noted Fed Governor Chris Waller "laid out the case that we could need some modest adjustments higher if inflation continues to rise." PIMCO's Tiffany Wilding said the print "buys them a little bit more time" but won't "close the door" on hikes, and that the Fed still has to figure out whether inflation is supply-driven or demand-driven. On The Julia La Roche Show (July 11), analyst Chris Whalen took the hard end of this view, warning of "double-digit inflation this fall" driven by diesel prices "up 30% year-to-date" and Middle East supply disruptions, and pointing to New York Fed data showing one-year consumer inflation expectations at 3.7%, near a three-year high. The steel-man: inflation has been above 2% for five-plus years, energy is already reversing higher, and the Fed can't afford to be fooled by one cheap-gas print.

Stall speed, the labor market is quietly breaking (data-focused analysts). The most granular bear case ran through the labor data. On Excess Returns (July 14), data specialist Eric Pachman argued the government "downplays a labor force participation crisis", the participation rate fell 0.4 points to 61.5%, and for men specifically it's down to "54.7%, the lowest level by far on record" (the prior low was 57.5% in the 1990s). His sharper point: nearly all recent job growth is coming from low-paid elder-care work, "social assistance… services for the elderly and disabled and home healthcare aides", jobs that pay around $35,800 a year and are mostly funded by Medicaid, which the new budget law is set to cut by "$1 trillion over 10 years." Jeff Snider's "demand destruction" reading (above) is the inflation-side version of the same worry. The steel-man: the headline 4.2% unemployment rate looks fine only because the labor force is shrinking, not because hiring is healthy, and the one part of the economy still adding jobs is the part about to lose its funding.

Worth naming the tension for readers plainly: the same week produced "best real-wage gain in 11 years" (Kudlow) and "the recession the unemployment rate can't see" (Pachman). Both can be partly true, real wages are rising for the people who still have good jobs, while the number of people in the labor force keeps quietly falling.

A note on the counter-camp that didn't show up much: the cheerful soft-landing case had one clear, credible voice this week. On Facts vs Feelings (July 15), Carson Group's Ryan Detrick and Sonu Varghese argued the job market is "better than the headline payroll numbers are telling us", payrolls averaging 110,000 a month year-to-date (versus about 120,000 for all of last year), unemployment back down to 4.2%, continuing jobless claims falling, and, a data point that surprised them, 531,000 annualized new business formations in June, "one of the highest levels ever." Their explanation for the thin hiring is telling: "we don't have a lot of labor supply… there are fewer people to hire." Which brings us to the immigration thread.

On immigration and the labor supply. No episode this week put a precise "break-even payrolls" number on the immigration effect, the specific figure that tells you how many jobs the economy needs each month just to keep unemployment flat. But the mechanism came through clearly from an insider: on Bloomberg Surveillance (July 14), former New York Fed President Bill Dudley argued the weak housing market isn't mainly about interest rates, it's that "we're not having any more immigration into the United States. And so the growth rate of household formation has collapsed… When you're not creating much in the way of job creation, the housing demand is going to be a lot weaker." Eric Pachman's care-economy data adds a second-order twist: he estimates 30–40% of the home-health aides doing nearly all the country's job growth are themselves immigrants, meaning tighter immigration could eventually choke off the one job engine that's still running.

The Trades in Play

Instrument-level positioning was thinner and more scattered than last week, but a few traders and fund managers laid out real, current bets.

  • Short precious-metal miners into specific downside targets. On The KE Report (July 11), technical trader Dana Lyons said the only thing he's been short "the last two or three months" is metals and miners, and it's "worked out nicely." His levels: he's looking for the silver ETF (SLV) around 48, the gold ETF (GLD) around 353, and the gold-miner ETF (GDX) down to about 65, where he'd "be covering a good portion or all of our shorts." He's a long-term metals bull just waiting for the correction to finish, and sees the critical-minerals ETFs (LIT below ~$70, REMX around $80–81) as closer to a bottom.

  • Stay long the AI/semiconductor complex, the cycle isn't near its peak. On Real Vision (July 13), fund manager Andreas Steno Larsen said he holds "9-10% of portfolio in memory stocks and over 25% in semiconductor companies total," dismissing the "buy peak earnings" bears, "the discussion is whether we've peaked or not," and he sees "nothing pointing to peak right now." He cited memory spot prices still climbing (up "another 20% or so" since Micron/Hynix updates), a Taiwanese supplier expecting the big players to hike prices "30% to 40% in Q3," and SK Hynix's order book running "until 2031." His guess for the actual top: "late 2027, early 2028."

  • Buy energy and gold stocks for reflation; be wary of the new IPOs. On Bloomberg Surveillance (July 10), George Noble, formerly head of overseas investment at Fidelity, called energy stocks his top idea, warning "we're sleepwalking into the biggest energy dislocation in history" with "very little downside and potentially a lot of upside" after crude's selloff. He also likes gold stocks ("a huge buy," citing SSRM at "7 times earnings") and copper names as reflation plays, while lambasting the wave of mega-IPOs, SpaceX at "120 times revenues" with a staggered lockup that unlocks fully by December, as a setup where "grandma's 401 is the exit liquidity."

Read-Throughs

  • The AI boom is turning into a credit story, and it may be why long-term rates won't fall. Two voices sharpened last week's "AI is the whole market" theme into a bond-market worry. On The Peter Schiff Show (July 15), Schiff argued the hyperscalers' "big cash cows have turned into cash vacuums", Meta, Amazon, Google and others have "borrowed tens of billions, hundreds of billions… over 100 billion already" to fund data centers, which "is going to drive up the interest cost of what everybody has to pay." Notably, even after Tuesday's cool inflation print, he pointed out the bond market gave back nearly its entire rally: the 30-year yield was "back up to 5.1" and the 10-year "back to 4.59" by the close. That refusal of long rates to fall on good inflation news is the read-through worth watching.

  • AI's contribution to growth is real, but it's decelerating. On Stansberry Investor Hour (July 14), strategist Marco Papic made the subtle-but-important point that growth depends on the rate of change of AI spending, and that rate is slowing: capex "went from $200 [billion] to $400, from $400 to $800, and then to $900. That's a slowdown", and it's "hitting some limiting point of copper, electrons, steel, and labor." He thinks the economy's real engine "has not been the consumer… it's really just been this pace of AI data center buildout," which makes 2027 look wobblier. Facts vs Feelings put a figure on how big that engine is: about "45% of [recent] growth came from investment in AI-related hardware and software," excluding data centers.

  • The next inflation print could undo this one. Almost every show flagged the same trap: June looked good because oil fell, and oil is climbing again. On Squawk on the Street (July 15), Bespoke's Paul Hickey offered the calmest framing, WTI crude averaged "$98.50 in May," "$82 in June," and was "$80 this morning," so even up 20% off the lows, today's oil is still below the June average the print was based on. Producer prices (PPI) told a similar story: down 0.3% on the month but still "elevated at 5.5%" year-over-year. The balance of risk for July's print is clearly to the upside on energy.

What Changed

Tuesday's recap ended on a cliffhanger: the June inflation report was "still ahead of us as this goes out," and the market was bracing for a possible rate hike. This week resolved it, and the resolution was a genuine plot twist.

First, the print itself: instead of the feared reflation, June inflation fell, headline down 0.4% on the month, the yearly rate from 4.2% to 3.5%, and core inflation dipping to 2.6%, its first monthly decline since 2020. Second, the market completely repriced: the hike bets that dominated last week's recap (a market pricing "two hikes" in September and December) collapsed within hours, with July hike odds falling from about 40% to 12%. Third, the disinflation-vs-reflation argument that was raging inside the bond market last week resolved decisively toward disinflation, two-year breakevens are now below the Fed's 2% target, versus the muddled, tariff-driven rise we flagged a week ago. And fourth, the consumer-stress story moved from a single shocking data point (last week's 50% jump in rent delinquencies) to a broader, harder set of numbers on student-loan and household credit strain.

The one thing that did not change: the Fed. Warsh took the good number and refused to blink, keeping the door to hikes open and the "regime change" rhetoric intact. So the standoff we described last week is still on, it's just that this week the market pulled decisively ahead of the Fed, and the big new known unknown is whether rising oil hands the hawks their ammunition back in the July print.