Santoli: Why investors should be on high alert heading into September

Get Mike Santoli's Market Memo in your inbox first. SUBSCRIBE HERE

In markets as in movies, suspense builds through the slow, quiet scenes of the second act.

The stock market is doing just enough, through slim trading volumes and narrow index trading ranges, to escape August with its uptrend intact.

The S&P 500 is within half a percent of where it closed three weeks ago. It has slouched and equivocated in the two weeks since it hit the last record high just above 7800.

But the index has stayed within 2% of the peak, the pullbacks so far halting a handful of points above what had been the top of the former multi-month range.

Sure, the upwelling of relief that greeted Nvidia's strong results and bold guidance was dramatic. But for the full week, the stock was up just 1.3%, back to levels from three months earlier.

Semiconductors as a group held right where they "should" have, preserving the path of their post-July rebound.

The market, in other words, has done its best not to do anything that would require disturbing the portfolio managers at the beach.

This is where one is conditioned to expect a discussion of calm before storms, of late-summer hurricanes and the punishing realities of September markets.

Yes, as is noted everywhere, September historically is the worst month of the year for stocks. But it's utterly unclear what one is meant to take from this fact, other than to keep expectations muted - which is always advisable in my book. Given most late-summer midterm-election-year weakness has been more than recovered right afterward, the proper course of action now becomes even more obscure.

There is massive variation around historical monthly return patterns. When stocks have already been strong in a given year, as they have in 2026, September has been less scary.

The ranking of monthly performance changes significantly if one goes back 20, 80 or 100 years. None of these spans truly represents a statistically significant sample. And of course, this is all about calendar months, not every possible 30-day slice of market history.

The Vix, 10-Year Treasury yield at key levels

Given all this, I'd argue the reason to be alert now is not purely the turn of the month, but the fact that - much like the S&P and the semis in August - various key market metrics are coiling near consequential thresholds, which if they're crossed could imply a change in market character.

The CBOE S&P 500 Volatility Index (VIX) has slipped below 15. Appropriately so, given the placid recent range and the clockwork mechanics of sector rotation and low-correlation restraining index-level volatility.

Still, much below 15 gets away from "comfortable stability" and toward "eerie complacency" territory. Historically, this time of year, vol is biased pretty clearly higher. For now, the VIX futures curve is sloped healthily upward into coming months, but things can shift in a hurry.

The 10-year Treasury yield has nudged back above 4.7%, reacting in part to the clear message from Federal Reserve Chairman Kevin Warsh at Jackson Hole on Friday that for the near term, he shares his committee's view that short-term rates are the tool to employ against stubborn inflation and that it might need to be used soon.

As discussed last week , there is no known tripwire level where bond yields kneecap equities. Bursts of fixed-income volatility tend to do more damage to stocks than a steady ratcheting higher of rates. (Just because everyone says this doesn't mean it's wrong.)

Still, the suspense over the next Fed move might itself color the tape. After Warsh's Friday speech, market-implied odds for a September hike were a bit above 50%.

Near-coin-flip probabilities less than three weeks before a Fed decision is a kind of "What if?" proposition that can hold risk appetites in check, and perhaps might become more a feature of a less-communicative Fed facing an economy running on two speeds - corporate-capex aggression and housing/consumer caution.

I'll say again, a 10-year Treasury yield just under 5% is not misaligned with the present 5-6% nominal-growth economy. Nor is it unusually far above the current Federal funds policy rate.

Sure, as many tech bulls today point out, the late-90s tech boom and full-employment jubilee occurred with yields between 5-6%. But back then, 5-6% was experienced as "low" rates.

Treasuries were more than a decade into a massive secular bull market as the '90s bubble inflated. The decade had begun with 10s yielding near 9% and they were just below 8% as late as December 1994, eight months before the Netscape IPO touched off the Internet-stock frenzy.

The rise in yields toward 5% today feels more intrusive to the current investing generation, and leaves most debt issued in recent years underwater in terms of price.

(On the bright side, a buyer of high-grade debt today enjoys a buffer of decent nominal and real yields, and bonds would gain more in value from a 1-percentage-point decline in yield than they'd lose from an equivalent rise.)

Aside from yields pushing the upper end of the range, the broad commodity indexes are rising toward five-year highs, corporate-debt spreads are remarkably tight and one quirky risk-appetite gauge I watch - the relative performance of lower-quality/cheaper Citi vs. defensive/pricey JPMorgan - has retreated back toward its early-2026 breakout level.

A key feature of the turn into September, which begins a year's third and final act, is that analysts and investors start to calibrate their views of next year. Corporate-conference season gets rolling and earnings models need freshening after summer.

Among the questions now: How will stocks metabolize an inevitable deceleration in earnings growth from this year's heroic, if overstated, pace. Charles Schwab noted last week that Nvidia and Micron together are providing one-third of aggregate 2026 earnings growth and the top-ten earners account for two-thirds.

Yes, the median company has returned to profit g...

Indexed and credited by AIPROPX. Originating outlet: CNBC. Open at source →