Buying the Weakness Is Really Two Trades, and Only One Was Already Cheap

The S&P 500 is sitting a hair off its record, and the advice you're reading this week is the oldest line in retail investing: buy the weakness. The index is up about 12% this year, stacked on top of a 24% year, a 23% year, and a 16% year, so a two-week slide that knocks it down roughly 2% off the high looks, to most people, like a gift you'd be a fool to leave on the table.

Here's the thing about this particular dip. It didn't start in stocks. It started in bonds — a repricing that has rippled into the index over the past several weeks. And that changes what you are actually buying when you "buy the weakness" in the S&P this month.

The dip is a rate, not a company

The mechanism is easier to see than the headlines make it. Oil — Brent, the global benchmark — went back over $100 a barrel in early September, after the U.S. and Iran started trading strikes around the Strait of Hormuz, the chokepoint that normally moves roughly a fifth of the world's oil. Oil above $100 is an inflation threat. Inflation is exactly what the bond market is pricing right now, and the bond market reacts to inflation by demanding a higher yield. So the 10-year Treasury yield jumped to roughly 4.9%, its highest level in nearly three years, and the 30-year hit 5.3%, a 19-year record.

Stocks and bonds are two sides of the same discount-rate coin. A stock is a claim on a stream of profits that is, well, less certain and more remote than a Treasury promise. When the yield you can collect on a bond that pays you a fixed amount for the next decade goes up, the price you'll pay for that less-certain, more-remote profit stream comes down. The stock didn't fall because companies got dumber. It fell because the alternative — "I can get 5% for a bond that barely has to try" — got better. That is a different animal from the kind of dip where earnings collapse, and the distinction matters, because the earnings aren't collapsing.

The "cheap" multiple is an artifact

This is where the "but the S&P is cheap!" argument goes wrong, and it's a good one to understand. The forward P/E on the index — the multiple you pay for next year's earnings — has actually drifted back down to around 20, below its five-year average, even as the price hit records. The only way a price and a multiple can move in opposite directions like that is if the earnings on the bottom of the fraction did all the work. And they did: asset managers put second-quarter S&P 500 earnings growth at roughly 50% year over year, the best since 2021, with the highest profit margins in over a decade. The engine is the AI buildout — the handful of mega-cap tech companies that now account for about a third of the index's profits and more than 40% of its market cap.

So the bargain multiple is an artifact of the AI earnings surge, not of the stock getting sold off. Pull that earnings line out and the multiple snaps right back up. Which means the cheap-looking stock and the risky-looking stock are the same stock. That's the whole ballgame in one sentence: this dip is not "the market got scared and marked a great business down to a discount." It's "the market got a little less patient about the future, and the future is exactly where all this AI money is supposed to show up."

The dip started in bonds, and it was cheaper there

Because the dip started in bonds, the cleaner version of "buy the dip" already happened — and it was a bond trade. Long-term Treasury funds have been sold down all year, more than 5% lower year to date, while stocks ran. The disciplined move the financial advisors keep making this month is not "add more S&P." It's rebalance: sell a little of the asset that went up and buy a little of the asset that went down, which now pays 5% for the privilege.

That version of "buy low, sell high" is worth pausing on, because it's a trade across the asset line rather than within stocks. You are taking profit off a position that's rich and putting it into a position the same oil shock just repriced. If oil cools and rates come back down, the long bond rallies hard — long-dated bonds are the most sensitive things in the whole system to a change in the rate. If oil stays up and the Fed keeps its foot on the brake, the bond bleeds slowly, but it pays you a coupon to wait.

The risk that the stock-only version of the trade carries points straight at the cause of the dip. The Federal Reserve is holding its main rate at 3.5% to 3.75%, and most economists expect it to stay there — but a meaningful minority now think it will hike, which would be the first increase since mid-2023, and the market has been pricing in hikes on the strength of the oil spike and a hawkish new chair. If the Fed raises and oil stays near $100, the discount rate keeps climbing into a market whose "cheapness" depends entirely on AI earnings that are, themselves, financed by borrowing at those rates. The earnings engine and the cost of money start to fight, and the multiple that looks like a bargain is the first thing to give.

That's the honest version of what "buying the weakness" is a bet on this month. It is not a bet that the S&P overcorrected and will snap back to a prior level. It's a bet that the earnings keep outrunning a discount rate that just started moving the wrong way. If you think the AI line keeps compounding and the Fed blinks, the dip is a gift and the stock is the play. If you think the rate repricing is only the opening, the earlier, cleaner dip was in bonds — and the stock is the second-hand version of it.

The two "buy the dip" trades are different securities doing different jobs. The one that started the move was already cheap. The one everyone's still talking about is a bet on an earnings line clearing a bar that keeps getting raised.