That the equity market is in boom mode is not up for debate. The S&P 500 is now up nine weeks in a row, a streak we last witnessed in late 2023. It is now up about 11% year-to-date. The fact that U.S. energy costs are up 47% from where they were at the end of February seems to be of little consequence to an investor base mesmerized by the AI trade that has morphed into an epic memory chip shortage, which, in turn, has sent the PHLX Semiconductor Sector index up a ripping 80% just since March 30th.
Real interest rates, measured by 10-year Treasury Inflation-Protected Securities (TIPS), have popped by over 30 basis points since late February, and that in the past reflected a tightening in financial conditions and an impairment to the U.S. stock market’s P/E multiple, but not this time.
The CAPE multiple has jumped to 39.6x in May from 36.9x in March, and is now tied for the second-highest level in the past century. At just under 2.1%, that real rate stands where it was in late October 2007, just over one month prior to the Great Recession, which came as a big surprise at the time, if you recall. Not even the shift of view in the money market from Fed easing to Fed tightening has upset the apple cart one iota.
All the while, the equity risk premium (the excess return investors demand for holding stocks rather than risk-free assets like government bonds) has vanished as investors now believe the stock market has become a riskless asset class.
One of the most glaring anomalies is that, as investors are brimming with confidence, businesses have joined the consumer sector in terms of losing theirs. The Conference Board’s CEO Confidence Survey for Q2 took a dive to 47 from 59 in Q1 — and again, right around where it was in the mid part of 2007. Six-month expectations slid to 46 from 57 and are lower than they were in the second quarter of 2020, when COVID-19 was raging. More than 30% of companies are now looking to cut their staffing in the next 12 months (another 40% say they are holding their staffing levels flat), a near record.
And if you don’t think AI is exerting a secular impact on hiring plans, 32% intend to reduce their headcount over the next three years, which is an all-time high for this data series. Only 7% of CEOs claim they are having any problem at all finding qualified workers — that number was 66% in early 2022, when the labour market was really heating up, and the primary reason why inflation did not end up proving to be all that transitory at the time — because the supply shock on prices quickly seeped into wages. That is not happening this time around at all. The grand total of a 4% share of the business sector plans to raise wages by more than +5% in the coming year — again, that number in the middle part of 2022 was more than four times larger, at 23%. Totally different kettle of fish today.
That said, despite the overall downbeat tone to the survey, one positive item that stood out was capex spending intentions. The share stating they are boosting their plans has gone from 15% in the third quarter of last year to 22% in the fourth quarter, 35% in Q1 of 2026, and now to 37% as of the current quarter. That is virtually unprecedented. Only 8% are cutting their one-year-ahead capex budget.
So, let’s get this straight. We have 37% of companies saying they are boosting their capex plans, which is near all-time highs, and only 28% state they are going to expand their workforce, which is very near a record low. How this ongoing divergence between the capital and labor share of national income (as we saw hit previously unseen levels in last week’s second quarter GDP revision) manages to generate anything but a fundamentally disinflationary backdrop in the future is a legitimate question. And it will be more evident once this war-related oil shock, at least partially, reverses course, which seems likely at this point since President Donald Trump, despite all the rhetoric, is aching for an exit strategy in Iran.
Now, there are some caveats to this market rally worth discussing, and one is the lack of breadth. Almost every day when the S&P 500 goes up, declining stocks are vastly outnumbering the advancers. Only one sector has made a new one-month high relative to the S&P 500. By contrast, six are at, or near, multi-year relative lows. While the overall market is at all-time peaks, only two of the eleven sectors have managed to reach that status.
It is indeed a very rare situation to be talking about a stock market at record highs at the same time that the Financials, of all sectors, are very nearly in correction mode (down almost 10% from their record highs).
One reason why we can be comfortable with the view that the bond market correction has seen its peak is because the run-up in interest rates over these past two months is really beginning to bite into the credit-sensitive sectors like housing. We got the U.S. mortgage applications data for the May 22nd week, and they plunged by 8.5% and have declined sharply in four of the past five weeks to the lowest level since last August. Refinancing activity has fallen like a stone to levels we last saw in July of 2025.
The new purchase index has gone nowhere this year, and this means the May data for the real estate market are going to look even worse than what we saw in April when new home sales cratered by 6.2% sequentially. At 622,000 units annualized, April new home sales were lower than they were in late 2007 when the unexpected housing-led recession was just getting started — a recession, like so many, that caught everybody by complete surprise.
Meanwhile, the number of homes available for sale bumped up by 1.7% month-over-month, and the gap between supply and demand has caused the inventory glut to expand to 9.4 months from 8.7 months (a balanced market is typified by a 6-month backlog of unsold homes). It has become so tough for the builders to lure in customers that the median length of time it is taking to make a sale upon completion has gapped up to 3.6 months from 3.5 months in March, 3.3 months in February, 3.0 months in January, and 2.8 months in both November and December. We have not seen a number this high since August 2021 — and it helps explain why average prices have deflated by 1.1% on a YoY basis and by 5.8% from the late 2025 cycle peak. A pretty important antidote (even if forgotten by practically everyone) to the oil-induced price shock.
I should add that if the U.S. Bureau of Labor Statistics had stuck actual home and rental rate numbers into the CPI instead of the methodology it currently uses, even with the energy price shock, headline inflation would be right near the Fed’s 2% target right now, and core inflation would be under 1.5%. Something for the Fed hawks and bond bears to consider.
In fact, the really important piece of inflation news last week was not just that the core PCE deflator came in less than expected, at +0.2% in April, but that the Jerome Powell super-core measure, which is PCE services net of energy and housing, slowed sharply to +0.1% MoM from +0.3% in each of the prior two months, and that was the lightest increase in a year. Powell may now be just a Fed governor, but the reason why he favored this metric is because it is one that is most closely linked to the contours of the labour market, and just as we are seeing in terms of decelerating nominal wage growth (and contracting outright in real terms), this attests to its continued loosening.
As for the U.S. economy, I have been on the receiving end of many questions on the Atlanta Fed’s Nowcast model calling for +3.0% real GDP growth in the second quarter. Well, nobody seems to have noticed that over a week ago, this estimate stood at +4.3%. Or that this was nearly the same number the Atlanta Fed was publishing for Q1 midway through that quarter, and look what we ended up with: +1.6%. At the same time, the current projection for Q2 real final sales is +2.5% annualized, so this is largely a war-induced inventory story. The New York Fed’s model is at +2.5%, in line with our proprietary survey-based model, and St. Louis is only at +1.1%. Our bottom-up sector analysis is so far showing a build-in of +2% QoQ growth for the current quarter, and we see that the U.S. economic surprise index swung into negative territory for the first time since April 24th last Thursday.
Outside of AI, there is zero vitality in the U.S. economy. In fact, strip out the AI spending boom and the 3-percentage point plunge in the savings rate to a four-year low of 2.6%, and real GDP would be running at a -0.6% YoY rate, not +2.6%. I have never in my 40-year career seen such a lopsided economic backdrop, where everyone thinks the economy is in such great shape when industrial capex, housing, non-residential construction, and consumer spending on durable goods and cyclical services are in contraction mode.
In fact, when you look beyond the AI binge, the federal government reopening effect, the impact from the further savings drawdown, and the nonfarm inventory swing, the economy contracted at a 1.1% annual rate in the first quarter. When you add up…
- Housing
- Nonresidential construction
- Old economy capex
- Consumer cyclical services
… the economy contracted at a 1.9% annual rate in Q1 after a 2.6% falloff in 2025Q4. So a significant chunk of the U.S. economy is already in recession, and yet nobody seems to be aware of it. And this is not a blind attempt at data mining. Rather, it is analysis revealing the lopsided nature of this expansion.
Allow me to go a step further. If you look at the U.S. economy from an income standpoint instead of a leveraged spending one, you will see that the economy is actually struggling. What went unreported was that real Gross Domestic Income growth throttled back to a mere +0.9% annualized rate in Q1, from +1.6% in Q4 and +3.5% in Q3 of last year, and that is with the corporate profits boom. As for the data on personal income, they tell quite a different story. Nominal disposable incomes in April were flat, which means -0.4% in real terms. The only reason why consumer spending managed to expand by an as-expected +0.5% in April was because households drew down their savings rate even lower to a four-year low of 2.6% from 3.2% in March (3.6% in February and 4.3% in January). A year ago, the savings rate was sitting at 5.5%.
Think about what a drawdown to 2.6% means. It means that had consumers been compelled to spend against their organic real after-tax incomes, real consumer spending would have declined by -1.1% from year-ago levels — not the actual +2.1% growth number. That 3-percentage-point-plus gap between spending based on incomes, and spending based on the equity wealth effect for the high-end household and stepped-up credit card and “buy now, pay later” usage for the low-end (and now middle-class), is absolutely epic. But this has separated what is still a rather moribund economy, despite the consensus narrative, from a recessionary economy (even with the massive AI spending).
I realize it is not fashionable to be talking about recession, but keep in mind that one of the critical cornerstones deployed by the National Bureau of Economic Research (NBER) is real personal incomes excluding government transfers, which fell by 0.4% MoM in April and were down for a third straight month (having peaked last September).
Finally, I recommend a read of the excellent column penned by Greg Ip in the Wall Street Journal titled Gap in Profits, Worker Pay Widens. To quote Pope Leo XIV on this file: “Income from capital risks replacing income from labor” — which he just wrote in Magnifica Humanitas, his encyclical letter devoted to the effects of AI. The question I have been asking for some time is whether capitalism is now acting as the wolf that bites off its paw — the relative share of profits to the size of the economy has never been this extreme at a peak, and the same can be said for the labour share of national income in terms of its depressed state.
My biggest fear is that this wave of left-leaning political leaders, at the U.S. state and municipal level, is going to serve as a mean-reverting template for the future (but is anyone paying attention to the fact that we are now up to two-thirds of America’s 100 largest cities being governed by Democrats??). These are things that are not on the radar screen but deserve to be, especially now that the Senate, not just the House, is in play for this November’s midterms, which means we are going to have a totally different fiscal policy backdrop on our hands than what we have seen unfold these past two years.
David Rosenberg is founder of Rosenberg Research.