Executive summary
 

1. The common question for this time of year: How will stocks finish the year?
 

2. As fun as it is to pontificate, I’m not sure it matters.
 

3. Take a look at 2021. A great finish to the year. And then? A brick wall.
 

4. We have our own brick wall this year. And just like at the tail end of 2021, it centers on tech.
 

5. The problems this year are a bit more concerning, though: They pertain to fair value.        

We’re in the end-of-year stretch, and in nearly every media appearance I have, the question arises: How will markets finish the year?

Our Chart of the Month below gives a sense of the seasonal pattern.

But while research suggests trends in rebalancing, tax-loss harvesting, and window dressing can have an effect, the macroeconomic backdrop, sentiment, and fundamentals can override what seem to be predictable trading trends. (I’m looking at you, 2018.)

And frankly, I'm not sure how much the final two months of the year tell us anyway.

Take 2021. The S&P 500 finished the final two months of the year up 3.6%. Technology did even better, rising nearly 8%. The Magnificent Seven gained another 3.1%.

A perfectly respectable finish to a very strong year.

And then came 2022.

The S&P 500 Index lost roughly 19%, while technology and the Magnificent Seven suffered even greater declines. In other words, investors who took comfort in the strong finish to 2021 were in for a rather unpleasant surprise.
In my view, extreme valuations played a major role, as did accelerating inflation and the interest rate hikes that belatedly followed.

Which brings me to today.

As I reflect upon it, 2026 has some similarities with 2021. But there is also an important difference. In 2021, investors were paying very high prices for businesses whose fundamental economics were relatively well understood. Today, valuations are more compelling, but the fundamental economics of those businesses may be becoming harder to understand.

And that, to me, is a more troubling problem.

Echoes of 2021?

Three things marked the period leading into the 2022 sell-off: Aggressive inflation, a Fed playing catch-up, and a tech trade ready to tumble.

Each of those variables is in question today, though none in quite the same manner.

Take inflation. In 2021, it was surging. In 2021, the Consumer Price Index averaged a 4.7% year-over-year climb each month. Today’s inflation is more modest — the latest data we have is September, which clocked in at 3.7% — but still frustratingly high, particularly considering we are rolling into year six of above-normal inflation levels.

But there’s a difference between inflation that is stubbornly high and inflation that is accelerating rapidly. And that difference matters for monetary policy.

In late 2021, the Fed was preparing to tighten monetary policy to get inflation under control. The rate hikes that followed in 2022 were remarkable, particularly for an economy and its markets that had grown complacent with rock-bottom yields.

I believe we sit at the beginning of another hiking cycle today, but this one is unlikely to reach the same heights. A more moderate, gradual cycle is less likely to spook markets. That said, rates are high across the curve, with structural forces (government spending, hyperscaler debt) that may keep the longer end of the curve higher regardless of what the Fed does.

So, some echoes of 2021, but a macroeconomic backdrop that isn’t quite as dicey.

Which brings us to tech. And here, I’m less reassured.

The tech trade: A different kind of risk

By November 1, 2021, the Bloomberg Magnificent Seven Index had returned 269% from the market low on March 23, 2020. And while the earnings trajectory looked compelling, the forward price to earnings was in the 10th decile of its own history; in other words, quite expensive.

After peaking on January 3, 2022, the Bloomberg Magnificent Seven Index fell 49%, marking an unusually painful collapse for the U.S.’s most prominent companies. And of course, putting it in a position to begin a remarkable rally, climbing 409%, peaking on October 13, 2026.

At first glance, today’s tech trade looks less vulnerable. The Bloomberg Magnificent Seven Index sits only in the lowest decile of its own valuation history. By comparison to 2021, that’s downright reassuring.

But I’m not sure the comparison is as reassuring as it seems.

The Magnificent Seven isn’t simply a collection of asset-light, high-cash flow, extremely profitable businesses anymore, but an amorphous group of businesses transitioning to an AI world. They remain highly profitable, with some encouraging signs around their ability to grow AI revenue, but they’re spending like wild to do it, and investors are leery of the potential return on investment (or lack thereof).

AI may well prove to be a transformational technology, but the capital required to build it out is enormous — and it isn’t yet clear whether the returns on that capital will justify the spending, or what will happen to legacy businesses, like Alphabet’s search business, that are less relevant in an AI world.

If businesses that historically generated extraordinary amounts of free cash flow now need to reinvest substantially more of their earnings to maintain their competitive positions, their fundamental economics may be changing.

We could also go half full here, since the opposite is also possible. AI could open enormous new revenue streams, and these companies could emerge as even more formidable compounders than they were before. The problem is that we don’t yet know which outcome we’re heading toward, or how much capital it will take to get there.

Which brings us back to valuations.

In 2021, investors were paying very high prices for businesses whose extraordinary economics were relatively well understood. Today, those prices look more reasonable. But the economics themselves may be becoming harder to understand.

And that’s an important distinction.

High valuations are, in part, a sentiment thing. If the fair value is stable, prices can move ahead or behind, and investors could use those price movements to help manage risk. But if the fair values themselves are in question, then we can’t be sure whether we’re looking at value or excess.

And today, the range of possible fair values may be unusually wide. These businesses could emerge from the AI buildout with extraordinary new earnings streams, or they could find themselves spending substantially more capital to generate the same returns. Both outcomes seem plausible. And that makes a stock’s historical valuation a less reliable guide to what it’s worth today.

None of this is to dismiss the hyperscalers. These remain extraordinary businesses, and even if we lower our estimates of their fair values to reflect greater capital intensity and uncertainty around returns, there may still be a discount in their share prices.

But the exercise of determining that discount has become considerably more complicated.

Of course, that question refers to a longer term than just these next two months. Which brings me back to where I started.

Frankly, predicting two months of returns is a heads-or-tails call. Probably we could lean toward higher; there have been more up days than down days, right? Seasonality is on our side, and the market has a remarkable capacity to look past longer-term concerns when earnings and sentiment remain supportive. Plus, we’ve seen some rotten market losses lately, which could embolden the #BuyTheDip crowd. (No judgment!)

Still. The experience of 2021 is a useful reminder that a strong finish to the year tells us very little about the risks that may be building beneath the surface.

And because everything worth saying is worth saying twice: The risks are large. Sell-offs based on valuation are one of my favorite kinds of sell-offs; they correct a misconception in markets and sometimes open the door for investors to get exposure at better prices. Sell-offs rooted in real concerns about fair value, however, are trickier; they require patience. And an appreciation for probabilities.

What I’m watching

Well, Anthropic, of course. I wrote about views around its valuation (maybe a touch high in my opinion) and impact on market structure (huge) back in the September edition. Looks like we are looking at an IPO before Thanksgiving, regardless of recent volatility.

Of course, I’m watching the AI trade broadly; the beatings it has taken recently all point to this concern I have around fair value — and whether the hyperscalers slow their spend, which hurts the broader AI supply chain. We’ll have to see what impact Nvidia’s numbers have on all of this; it reports on the 17th. On that note, the espionage angle looms large at the chipmaker; like everyone else, I'll be watching to see how that plays out.

And, of course, all things Iran, inflation, and rates. The more things change, the more they stay the same.

Explore previous editions

October 2026 | September 2026 | August 2026 | July 2026