by Hubert Marleau, Market Economist, Palos Management
Last week I wrote: “There are several cross-currents at play like crude prices and the Iranian war, monetary policy and the yield curve, doubts about the capex cost of the AI trade, the huge funding needs stemming from large fiscal deficits and heavy corporate financing, and real yields that could change the stock market’s speed limit. On the other hand, there are many encouraging signs that make me believe that the market has much more room to run, given the prospects of a friendlier monetary stance and the widening effect of the AI boom.
“It is true that the AI boom is getting bigger, but it's no longer doing all the work: its impact is spreading into all sectors of the economy, widening the market’s profit base, and making earnings a bigger story than originally thought.
“In this connection, S&P 500 forward earnings per share, which are currently at $400, should progress toward $415 by the end of 2026 and then to $484 by the end of 2027. Assuming that 10-year Treasury yields will range between 4.50% and 5.00%, I think that by the end of this year it is highly probable that the benchmark could touch 8,300 as a minimum.”
The week of August 16
Stock operators started the week pondering what the Fed's next move would be, and where the rising debt load stemming from relentless fiscal spending deficit and the AI capex boom would take them. The rise in bond yields combined with hawkish minutes from the last FOMC meeting explained why the S&P 500 tumbled 100 points by Thursday as a result, registering a weekly drop of 1.3%.
First, the street belief of two months ago that incoming data would support a rate hike has changed to one that no longer holds. A recent cooling of inflation, labour conditions and consumer spending has changed the dynamics, leaving business investment as the only thing holding the economy up. The Atlanta Fed economic tracker is currently projecting a real growth factor of 4.0% in Q3, versus 6.2% just a few weeks ago, while the Citi Group Economic Surprise Index has plunged to 15.4 from a high of 58.7. While I acknowledge that the recent series of dramatic oil prints associated with Trump’s threat of economic warfare against Iran has caused havoc in both the bond and stock markets, I do not adhere to the thesis that oil prices are likely to veer out of control in spite of the $15 rise in West Texas Intermediate to $89 a barrel in less than two weeks. As a matter of fact, the long end of futures is not buying the permanency of this price increase: indeed, interestingly, 1-year oil futures are $15 lower than the spot price. The fact of the matter is that the Americas, particularly the US, Canada, Brazil and Guyana, are experiencing an oil boom, surpassing the Middle East in production, which is expected to continue beyond 2027, thereby creating a headwind against price pressures. This is happening while China is slashing oil imports, world demand has dropped by 1.6 million barrels a day in 2026, a lot more oil is flowing with covert methods through the Strait of Hormuz than is commonly accepted, and traders have discovered and/or invented new oil routes. Perhaps surprisingly, US commercial and strategic stockpiles of crude and product inventories have actually risen over the past few weeks, according to the International Energy Agency.
Second, the year-to-date increase in 10-year Treasury yields of 62 bps to 4.72% was not about higher inflation expectations, which has been remarkably stable around 2.25%, or the performance of the US Dollar Index (DXY), which has consistently traded around 98.50 throughout the piece, because these two highly sophisticated markets are conscious that when it comes to where the underlying inflation trend is, what matters is unit labour costs. This measure, which has been historically the key driver of inflation, is currently running below the Fed’s 2% target because wage inflation is falling while productivity is rising. In other words, rising computer chip and electricity prices are a reflection of the AI-linked boom, and not general inflation.
Moreover, contrary to popular opinion, high-grade hyperscalers' insatiable demand for capital to build their data centres has not had a direct impact either, otherwise Baa credit spreads would have widened; instead, they have fluctuated narrowly around 165 bps for the entire period under consideration. In my humble estimation, the culprit has been the crowding effect of an immense issuance of corporate bonds on an already overly extended load of $40 trillion of government debt, almost 25% more than nominal GDP ($32.5 trillion), at a time when the national savings rate is incapable of funding the capital needs of the nation without the help of private foreign investors who seem to be more interested in corporate debt and equity. Put simply, Bessent is being forced to change some of the fundamentals, particularly the Federal spending deficits, to arrest the relentless increase in real rates, which has resulted from a different buyer base insistent on getting more yield to compensate for long-term uncertainty risks. This is why the “Bessent Put” buyback intervention in the long end of the government bond market to reduce borrowing costs did not prove to be effective other than tweaking a few illiquid maturities.
This explains why the White House is about to embark on a major consolidation effort to cut waste, combined with a plan to address its fiscal deficits. Given that 10-year US Treasury yields (4.70%) are presently 150 bps higher than the comparable world average yield of 3.20% and also at about the same level as the year-over-year percentage change in the money supply, the program may work out favourably and land long rates near their current level.
The Return of the Fed’s Valuation Model
Thirty years ago, the then-Fed Chairman Alan Greenspan introduced the idea that a good way to judge whether stocks were overvalued or undervalued was to compare the 12-month S&P 500 forward earnings yield to the 10-year US Treasury bond yield, where a zero difference was regarded as fair value. On Thursday, the benchmark closed at 7,641, with Yardeni’s forward earnings forecast for the next 12 months of $391 per share, generating an earnings yield of 5.12% versus a 4.70% closing yield for 10-year Treasuries.
Based on these numbers, the S&P 500 is undervalued for as long as 10-year Treasury yields remain below 5.12%. In this connection, 8,300 for the S&P 500 is still a rational prediction. On Friday, the S&P 500 rose 40 points to close at 7,675 on its way to target.
Copyright © Palos Management
