US economy — K-shaped / “run-it-hot” 2026 (new) Wealth-effect-funded spending on a record-low savings rate — behaviour over attitudes, with AI growth breaking before the consumer

Hay (Jan 5 '26): UMich consumer sentiment sits at 65-yr recession lows even as asset prices set records — a K-shaped “Special K” economy where the top 10% of earners drive ~50% of spending. Into the mid-terms he expects the “mother of all manipulations”: a pliable Fed pick (Trump wants 1% funds vs ~3% inflation, ~−2% real) plus a fiscal flood (bonus depreciation, ~$100B refunds, ~$500B tariff rebates, cash-for-clunkers, housing) → an inflationary boom, durable first then fading; long bond/mortgage rates keep rising despite 175 bps of Fed cuts. Hunt “run-it-hot” beneficiaries (Jacobs, J). Echoes Polomny (Oct 5 '24): the Fed is cutting for the fiscal hole, not a recession — “inflate or die” → own hard assets. Hay (Jul 1): housing has become structurally unaffordable and is a core driver of the two-speed economy — it now takes ~5× median household income to buy a median-priced home (down from the 5.8× 2022 peak, above the 4.7× 2007-bubble top) and mortgage payments eat ~35% of median income vs a ~20% 1984–2021 norm; median new-home prices are +44% since 2016 and existing +79%, converging as builders subsidize sales — cutting D.R. Horton / Lennar margins by a third to a half even with the homebuilder ETF near all-time highs — while home-insurance premiums (+21% 2021–24) and the average buyer's age (29 in 1980 → 40 today) compound the squeeze; regulatory red-tape reform is his main policy lever. Paulo Macro (Jul 8, the rollover read): the economy is rolling over after inflecting into Run-It-Hot late last year, and the issues run deeper than gasoline squeezing the bottom of the "K" — large employers shedding job openings, ISM orders:inventories (a 4-month PMI lead) rolling over after its May peak, construction "all about data centers," and real estate mired in weakness (he's warming to the Boomer-ageing residential→nursing-home shift flooding supply over the coming decade). NFIB hiring plans say private payrolls cycle either side of zero for months with a possible negative print late summer — which caps how much tightening the Fed follows through on and "could very well touch off the classic 4Q rally" of the midterm-year seasonality (bad data = easy liquidity); he's watching September opex/quarter-end. Hay (Jul 10, the lower-K stress datum): US 90-day+ auto-loan delinquencies hit 5.6% in Q1 2026 — the highest since 2010, above prior peaks (next release Aug 4) — the counter-cyclical repo-supply tailwind behind the Copart POW pick. App Economy (Jul 11): the lower-K read straight from the C-suite — PepsiCo's Laguarta: "the consumer is worse than what we had anticipated, driven mainly by gas prices" (the pullback concentrated in convenience/impulse channels; Frito-Lay volume flat), and General Mills' COO expects NO consumer recovery in FY27, shoppers still buying on promotion and trading down pack sizes. Eisman (Jul 31): the split in one earnings week — Visa payment volume +10% and Mastercard +8% ("no signs here that the consumer is slowing down") against Procter & Gamble at 0% organic revenue growth and EPS −3%. His mailbag counterpoint via John Cassidy's GDP-per-capita chart (flat from year 1 to ~1800, vertical since): "we can all complain about the K-shaped economy, but let's not forget that we are still all better off." AUG-16 (Jay Singh, SSR): the structural number behind the sentiment-vs-spending paradox — the top 10% of US consumers now do 60% of all spending, up from 30% two decades ago, so aggregate consumption tracks asset prices, not wages. Sentiment is near record lows on personal finances and the labour market while the savings rate sits at one of its lowest levels on record against record household net worth: a falling savings rate in a strong economy is "a sign of high consumer confidence, even though the surveys don't say that." He explicitly discards the survey as a forecasting input ("we are most interested in consumer behavior, not consumer attitudes") and cites the Minneapolis Fed finding little evidence of a K-shaped pattern in real spending volumes — households on all rungs buy more goods and services than in 2019. The clock he now watches is nominal GDP growth (AI-driven) racing ahead of aggregate weekly payrolls: "the first thing to break will probably be AI growth, and then second will be the consumer, because these savings rates are not sustainable." Housing is the source of the anger without being the source of a spending crack — frozen turnover, no forced sellers, and job mobility suppressed with it. 2026-AUG-17 — Jay Singh (David Lin Report): the top 10% of consumers now do ~60% of all spending vs ~30% two decades ago; essentials 20–25% above pre-pandemic; job mobility cooled and "Meta is firing people to invest in these data centers." Half of GDP growth is AI/data centres — B2B — and "that can continue maybe for a few quarters, but you need the consumer to come back." Luxury outperforms, trade-down benefits Walmart/Target, undifferentiated brick-and-mortar suffers. "This is not a main street economic boom." 2026-AUG-22: App Economy: demand is holding but the prints need adjusting — Walmart's 2.6% US comps (weakest in 6+ years) carried a ~125bp federal drug-pricing drag (ex Health & Wellness +3.4%, transactions +1.5%); Target's comp was +3.8% on +3.6% traffic; Klarna cut FY26 GMV guidance on softer German discretionary spending — the one clearly weak consumer datapoint of the week. (App Economy Insights 2026-AUG-22) 2026-AUG-23 — Jay Singh (SSR call, 2026-AUG-23): the consumer finally printed the slowdown. Walmart fell 8% — its worst day since May 2022 — on the slowest US comparable-store sales growth in more than six years (2.6% vs 3.7% expected). Goldman cuts second-half real consumer spending growth to 1-1.5% from 1.8% and full-year real GDP to 2.1%, with the Strait of Hormuz still closed and crack-spread-driven gasoline prices an added risk to lower- and middle-income households. "The majority of GDP growth is AI related spending." Kedrosky (Meb Faber #648, Aug 28): the statistic that pulled him into the whole subject is that AI/data-center spending has been more than 50% of US GDP growth, and that share has grown and persisted for six straight quarters — "for probably the sixth time in Western history, we have some force that's non-governmental that's actually so large as to shift the tides of GDP growth." The consequence is an attribution error he calls the dog-and-mailman problem (the dog barks, the mailman leaves, the dog takes credit): "your tariffs did not cause the US GDP to grow in that period. What caused it was this remarkable thing called data centers and hyperscalers" — and a broken causal model invites policy "actions that are actually really consequentially negative." It also makes the whole growth path hostage to one capex cycle. Steve Eisman (The David Lin Report, 2026-AUG-28) restates the K and names its single support: "we do have a K-shaped economy. The bottom of the K is struggling. There's no question about that. But as long as AI CapEx keeps powering higher, the economy and the market probably goes higher." Given AI capex is ~half of ~2% GDP growth, the top of the K is the AI trade. He locates the damage precisely — not in consumer goods generally but in shelter, and in the running costs rather than the purchase price: "the area that's been really, really hurt is housing. And it's a lot of hidden costs — the cost of taxes, the cost of property and casualty insurance, the cost of utilities. Because of that inflation there's a whole swath of Americans who've been completely priced out of the housing market. Leaving aside the price of the house, just to bear the annual costs are beyond their means." Scope deliberately limited: "that impacts housing-related stocks, but I don't think it really impacts much beyond that," and a housing correction "would just be a normal correction. I don't think it would be calamitous." On geopolitics: "I don't think the Iran war has real long-term implications for the US economy." 2026-SEP-01 (Paulo Macro, Substack chat): a pre-registered, bounded call now reporting in — “recall I have discussed at length since June the prospect of a soft patch in US data between midyear and mid autumn, and the Citi Economic Surprise index would suggest this is happening.” Note the construction: a window rather than a date, a shape rather than a magnitude, and a single named gauge that measures data against expectations rather than against its own history. No payroll estimate, no threshold and no recession call is offered; August NFP (Fri 4 Sep) simply falls inside the window he called in June, which is what makes it a test rather than a narrative. 2026-AUG-26 (Dowd/Phinance, WTFinance): without the AI build-out "the GDP would be a lot weaker — we may even already be in a recession." Walmart's quarter is the tell: beneath the drug-pricing noise "their same-store sales was the weakest it's been in 6 years and they said the consumer is struggling." "Joe Sixpack, 80% of the country is not doing well," and 10–20% of consumers cannot carry the rest of the economy. (2026-SEP-01, Paulo Macro) The 3Q soft patch he called in June is confirmed and spreading: "The Citi Economic Surprise Index seems to confirm this as the data has been progressively missing since June (note Japan rolling over now as well — Europe next? And what then?)." The day's US prints are the wrong kind of weakness — "it screams of stagflation." Dillian & Farley (2026-SEP-03): the payrolls-vs-unemployment divergence, unresolved on air. Dillian: “the rest of the economic data is actually terrible” — two weak payroll reports, 55,000 expected next, Chicago PMI ten points light, JOLTS “terrible,” ISM soft. Farley’s correction: payrolls are “quite correlated with immigration,” so 300,000 a month under high migration was never proof of strength and the break-even rate may now be “40,000 or 50,000” — which is why “payrolls … look abysmal, but the unemployment rate has actually gone down,” from 4.5% to 4.1% “and not only on an absolute level is it good, but it has momentum.” Dillian attributes it to participation (native-born participation running below immigrant participation) and calls the rate “a little bit of a head scratcher.” Both recall the false alarm: at 4.5% “everybody was saying that it triggered the Sahm rule … and Claudia Sahm came out and said actually technically it triggered the Sahm rule but it probably did not.” CNBC Halftime — Bill Baruch & Kevin Simpson (2026-SEP-04): the K read the desk actually traded. Baruch sells Amer Sports outright despite its top-of-the-K brands and buys Walmart as “a trade down for the consumer” into $100 technical support, 25% off the highs; pressed on why weakness should hit $300 tennis rackets, he redirects to margins rather than volumes. Simpson trims TJX after a double beat because the core Marmaxx comp fell from +6% to +1%, while conceding the K argument cuts for the name — “if there is a K consumer or a K economy, both sides of the K go to TJ Maxx.” And the whole athleisure complex — Lululemon, Nike, On Holding — is declared “no touch” on a secular, not cyclical, read. 2026-SEP-07 (Nomi Prins) — the August labour-market decomposition: of the +162k headline, food services & drinking places contributed 59k and local-government education 42k (school-year seasonal); the rest of the private economy managed 61k. She reads Trump's simultaneous claim that the report beat estimates “by double and triple” and his demand that the Fed “Lower the Rate” as a tell that the White House knows the print was weaker than its headline. 2026-SEP-07 (Joe Brown) — the opposite pole to the Prins note above, on the same print: a direct counter to the “wages never kept up” reading — average hourly earnings went from under $30/hr in 2020 to over $37/hr, ~25% in six years, and the trend growth rate is steeper than pre-2020, not a reverted spike. He reads the labour market from the openings side: 7.4m JOLTS openings, from which he draws the deliberately provocative inference that “you can ignore unemployment statistics” because the unemployed are declining available work. Flag as contentious and his own. 2026-SEP-07 (Jay Singh) — the borrowing half is cracking: 60-day delinquencies on US subprime auto loans at ~5.2%, the highest on record and more than double four years ago. Job growth is concentrated and low-quality (~77% of August's 162k in leisure/hospitality, government and healthcare); JOLTS fell to 7.271m; UMich sentiment stays weak against a strong services ISM at 55.4. App Economy Insights (2026-SEP-12), Chewy Q2: food and medication spend holding while discretionary categories like treats "weakened more sharply"; Chewy assumes no consumer recovery or pricing benefit for the rest of the year, and shares fell 11% despite a slight guide raise. Mike Taylor (Hedgeye, 2026-SEP-15): outside debt-funded AI, 'the consumer stinks. Housing stinks. Auto loans stink. Student loans stink' — plus a coming white-collar drag (he cancelled a 60-lawyer, $800/hr firm after using Claude). Barron's (Rivas, 2026-SEP-09): Walmart US comps slowed to a multiyear-low 2.6% as the energy spike squeezes lower-income shoppers, yet it beat and raised — “American shoppers are still more resilient than many investors expected, even with the albatross of inflation.” Puplava (2026-SEP-11): K-shaped — asset owners gain from higher rates while the paycheck-to-paycheck consumer is "really getting hurt"; strip out AI and data centers and non-AI private investment is shrinking in nominal and real terms (partly tariff uncertainty) — a "fragmented economy," though he sees no recession signal. 2026-SEP-21 (Spencer Jakab, WSJ Markets A.M.): rising rates and gasoline prices are squeezing lower-income drivers into deferring car maintenance, the same first phase seen in 2008-09 when oil hit its all-time high. AutoZone is down 31% in a year and O'Reilly and Advance Auto Parts have sounded downbeat. Jakab expects the pattern to reverse as unaffordable new cars force owners to keep repairing an aging fleet (cars 4-11 years old are the parts sweet spot), and notes unemployment is half the 2011 level, so the squeeze is milder than the last recession. 2026-SEP-19 (Jeff Weniger): he expects a pleasant economic surprise over the next 2-4 quarters. His 2027 arithmetic is 2.5% real GDP + 3.5% inflation = 6% nominal, which keeps debt/GDP flat near 120%, still below the COVID peak. Atlanta Fed wage growth (+3.8%) is beating CPI (3.4%), which favours discretionary over staples. Improved fleet fuel economy means the 2008 gasoline shock equals ~$9.50/gal today, so it takes $6-7 gasoline to squeeze the paycheck-to-paycheck consumer. 2026-SEP-18 (CNBC Halftime): Raskin - real income has been negative for most of the last ~18 months, with the wealth effect propping up spending while gasoline and mortgage rates rise; Sechan - "the lower end of the K is very stretched," so consumer brands (Nike, On, Deckers) must wait for inflation to cool and real disposable income to rise: "now is not that time." Harley Bassman on MacroVoices #550, 2026-SEP-17: corporate profits as a share of GDP doubled from 6% to 12%. The S&P 500 and Michigan sentiment have opened an 'alligator jaw' (stocks up, sentiment down), a measure of public distrust.