Howard Marks talks in specific numbers, which is unusual for someone whose reputation rests on temperament rather than forecasting. His Wharton appearance is built around three of his memos, and along the way he gives figures precise enough to check. So we checked them.
The number that holds
Describing how the market got where it is, Marks runs the recent record:
"Then the market was up about 26% in '23, 27% in '24, 18% in '25 — one of the best periods, three very strong years, up 87% in those three years."
Against index data, on a price basis and a total-return basis:
| Year | S&P 500 price | S&P 500 total return | Marks |
|---|---|---|---|
| 2023 | +24.2% | +26.3% | 26% |
| 2024 | +23.3% | +25.0% | 27% |
| 2025 | +16.4% | +17.9% | 18% |
| Three years | +78.3% | +86.1% | 87% |
His aggregate is essentially exact on a total-return basis: 86.1 against his 87. Two of the three annual figures land within a fifth of a point. The 2024 number is the outlier — 27 percent against an actual 25.0 — and it is the kind of two-point drift you get quoting from memory in front of an audience.
Price return over the three years was 78 percent. Total return was 86 percent. Marks is quoting total return, correctly, but the two bases differ by eight points over three years — enough to change how expensive you think the market got. Whenever someone quotes a multi-year run-up, the first question is which basis.
The number that does not
Recounting his first job at Citibank in 1969, Marks describes what happened to the bank's Nifty 50 holdings:
"If you got there and bought the stocks the day I started work in late '69, and if you held them tenaciously because of resolve, because of dedication, because of intellectual commitment — if you held them for five years, guess what happened? You lost about 95% of your money."
That figure is far more severe than the record supports. The S&P 500 itself fell roughly 48 percent peak to trough in the 1973–74 bear market. For a fifty-stock basket of the largest, most liquid growth companies in America to have lost 95 percent over the same span, it would have to have fallen about twice as far in log terms as the index it dominated — and several of those companies were still standing, and large, at the end of it.
The better-known measurement runs the other way. Jeremy Siegel's study of the Nifty Fifty bought at the December 1972 peak — the worst possible entry — and held through 1995 found returns roughly in line with the S&P 500 over that horizon. Individual names were destroyed. The basket was not.
To be fair to the specific claim: Marks is describing a five-year window ending near the 1974 trough, which is the single worst window available, and he is describing what an investor felt rather than a portfolio he ran. We could not reconstruct the 1969–74 return of the actual basket from public sources, so we are flagging the figure as unsupported rather than false. But "about 95 percent" is doing rhetorical work that the historical record does not back.
The framework, which is the part that matters
None of this touches the argument, because the argument does not depend on the anecdotes.
The lesson Marks draws from Citibank is the one line most worth keeping:
"It's not what you buy, it's what you pay that counts. Good investing doesn't come from buying good things. It comes from buying things well."
With the corollary that makes it operational:
"There is no asset which is so good that it can't become overpriced and dangerous. And there are very few assets which are so bad that if they get cheap enough, they can't be a good buy."
On where we are now, he is careful to describe rather than predict — 22 times earnings against a historical average he puts at 16 to 17, and the standard rejoinder:
"The optimist always says, yeah, but this time it's different. In other words, history is not relevant."
And on the question every investor actually wants answered, he refuses it in the most useful way available:
"What is the bottom? ... The bottom is the day before it starts going up, right? And if that's true, then, by definition, you never know when you're at the bottom, because you can only tell the next day."
"One of the dumbest things you can do in the investment business is to say, I'm going to wait for the bottom."
On sentiment, which is his actual subject
The through-line of the whole session is that prices move further than facts do:
"In real life, things fluctuate between pretty good and not so hot. But in the minds of investors, they go from flawless to hopeless."
He applies it to the software credit market, describing a sequence where new coding models raised a question about software company economics, which spread to the debt those companies had issued, which spread to investors trying to exit instruments they had not asked exit questions about before buying. That is a mechanism, not a mood, and it is the most immediately actionable thing in the talk: the risk was not the AI news, it was owning something whose exit terms you had never read.
What to take from it
Take the framework and check the figures. That is not a criticism of Marks specifically — his three-year number was accurate to within a point, which is better than most people manage from memory. It is that a talk delivered from decades of experience mixes precisely correct data with anecdotes that have been retold enough to drift, and the two are indistinguishable in the room.
The parts of this session that will still be true in ten years — pay attention to price rather than quality, never plan on catching the bottom, read the exit terms before you need them — do not rest on any number in it.
Originally published at pickuma.com. Subscribe to the RSS or follow @pickuma.bsky.social for new reviews.
Top comments (0)