Michael Steinhardt: 28 years, one bad year, and the leverage behind it
For twenty-eight years Michael Steinhardt ran the best-performing hedge fund most people have never heard of. Then one year took a third of it. The interesting part is not that he lost — it is how small the market move was that did it, and what the two published numbers behind that move imply when you divide one by the other.
KEY TAKEAWAYS
- 28 years at 24.5% net turned $10,000 into roughly $4.8 million; the S&P 500 turned the same $10,000 into about $190,000 over that period.
- Fortune reported a $30 billion Eurobond position against $4.6 billion of capital. That is 6.5× leverage — our division of their two figures.
- At 6.5×, the bond move that took a third of the fund was about 5%. The fund fell six and a half times as far as the thing it owned.
- 1994 (−31%) and 1995 (+26%) combine to −13.06%. A very good year did not undo a bad one.
Who was Michael Steinhardt?
Michael H. Steinhardt was born in Brooklyn in December 1940 and graduated from the Wharton School in 1960. He worked as an analyst covering conglomerates at Loeb, Rhoades & Co., and in 1967, aged 26, co-founded Steinhardt, Fine, Berkowitz & Co. — later Steinhardt Partners — with $7.7 million under management.
It was a hedge fund in the original, literal sense: as willing to be short as long. “I really shorted a lot. I liked to short,” he told Forbes in 2014 — adding that this was “a very dangerous thing, because the short side is so tough.” The fund made money through the 1973–74 bear market, which almost nothing else did.
The record: 24.5% a year on average from 1967 to 1995, net of a 20% performance fee. Forbes put it in the form that lands hardest — $10,000 given to Steinhardt in 1967 was worth about $4.8 million the day he closed, against roughly $190,000 in the S&P 500. We checked that the two figures tie: $10,000 growing to $4.8 million over 28 years is a compound rate of 24.67%, which is the 24.5% headline rounded from a slightly different start date. The numbers are consistent, which is more than can be said for a lot of quoted track records.
He was also, by the accounts he himself reprints in his memoir No Bull, a brutal employer — staff used words like “mental abuse” and “rage disorder”. “When I did poorly, I was intolerant and difficult and had a bad temper,” he told Forbes. We mention it because this section prints the record, and because temperament and position size are not unrelated.
What was “variant perception”?
Steinhardt’s chapter in Jack D. Schwager’s Market Wizards (1989) is titled The Concept of Variant Perception, and that phrase is the one thing of his that every trader eventually meets.
"I try to develop perceptions that I believe are at variance with the general market view. I will play those variant perceptions until I feel they are no longer so."Michael Steinhardt — interviewed in Market Wizards, Jack D. Schwager (1989).
The idea is unglamorous and correct: if your view is the same as the market’s, it is already in the price, so there is nothing to be paid for being right. You only get paid for the gap between your view and the consensus — and only when the gap closes your way.
It is a good idea. It is also, on its own, an entry idea. It tells you what to buy. It says nothing whatsoever about how much of it to buy, and that is the seam 1994 opened.
What actually happened in 1994?
1993 had been extraordinary for leveraged bond funds — Fortune reported Steinhardt Partners had returned better than 60% in each of the previous three years. The engine was the carry trade: borrow at short-term rates near 3%, buy longer-dated bonds yielding 6% or more, pocket the spread, multiplied by however much you borrowed. Some funds were buying bonds on margin for as little as one or two cents on the dollar.
On 4 February 1994 the Federal Reserve raised the federal funds rate by 25 basis points, from 3% to 3.25%. A quarter of one percent. The 30-year Treasury yield jumped 40 basis points almost immediately, because leveraged holders had to sell to meet margin calls, and their selling pushed prices down, which produced more margin calls. Fortune quoted a name for it that stuck: “the great worldwide margin call”.
Steinhardt was not in US Treasuries. He was in European bonds, where rates were at 30-year lows and the consensus said they would fall further. That was the variant perception, and it was wrong. Fortune’s October 1994 account of the year reports the position:
- a Eurobond position reportedly around $30 billion;
- against $4.6 billion under management;
- losing about $4 million for every one-basis-point rise in European rates;
- by May, losses of roughly a third of the fund;
- still down more than 30% at the start of September.
The year closed at −31% — by a distance the worst of his career. He was not alone: Leon Cooperman’s Omega fell 24% in the first half on the same trade, Julian Robertson’s Tiger was down 7.5% by mid-September, and Askin Capital Management simply ceased to exist when it could not meet its margin calls.
How much leverage was that?
Here is the number that almost no retelling of this story prints, because it requires dividing one reported figure by another: $30 billion of bonds against $4.6 billion of capital is 6.5× leverage.
That single ratio explains the whole year. Work it the other way and it gets worse. If a third of the fund — about $1.53 billion — was lost on a $30 billion book, then the bonds themselves fell about 5.1%. A five percent move in a bond portfolio is a bad but entirely ordinary year in fixed income. It took a third of one of the great hedge funds in history.
The per-basis-point figure says the same thing in a form you can feel. $4 million against $4.6 billion of capital is 0.087% of the entire fund for every single basis point. One basis point is a rounding error on a screen. An unremarkable ten-basis-point day moved 0.87% of all the money in the fund. He did not need a crash. He needed a normal quarter going the wrong way.
Why a 26% year did not fix a 31% year
Steinhardt said he vowed to make his investors back as much as he could. In 1995 the fund returned +26%. On most Wall Street desks that is the best year of a career.
It did not get them back. $100 down 31% is $69. $69 up 26% is $86.94. The two years together are −13.06%, and getting from $86.94 back to $100 needs another 15.02%. That is the arithmetic every risk rule on this site exists to respect, and it is the same arithmetic that makes the big loss the only rule that has to hold.
Two more ways to size the damage, both ours. At his own career average of 24.5% a year, recovering a 31% drawdown takes 1.69 years of best-in-industry compounding just to be level again — nearly two years of the finest work in the business bought precisely nothing. And in the $10,000 terms above: had 1994 merely been flat instead of −31%, the terminal figure would have been about $6.96 million rather than $4.8 million. One year removed roughly $2.16 million per $10,000 invested — more than eleven times everything the S&P 500 produced across the entire twenty-eight years.
Then, after the good year, he stopped. “I thought there must be something more virtuous, more ennobling to do with one’s life than make rich people richer,” he told Forbes. He was in his fifties, and the hedge fund industry he had helped invent was about to become the most profitable business on earth without him.
The other bill: $70 million
1994 charged him twice. In 1991 the SEC began investigating four hedge fund managers — Steinhardt, Soros, Robertson and Kovner — over an alleged corner in two-year Treasury notes in concert with Salomon Brothers. Robertson and Soros were dropped. In 1994 Steinhardt Partners settled for $70 million, about 75% of it from Steinhardt personally; Salomon paid $290 million. He has maintained he did nothing wrong and settled to move on. We include it because a profile reporting only the trading loss tells you half of what the record says about that year.
What this means on a $2,000 account
You will never run $4.6 billion. You can reproduce the exact mistake before lunch.
Open a $2,000 account and take a position at 20× leverage — a default many exchanges will happily offer you. Your $2,000 now controls $40,000 of an asset. Run the same division Steinhardt’s numbers invite: 1 ÷ 20 = 5%. A 5% adverse move erases the account — and in practice the exchange closes you a little before that, because the maintenance-margin buffer bites first.
Now recall what took a third of Steinhardt Partners: a move of about 5%. The same 5% that cost the best trader of his generation a third of a $4.6 billion fund would cost you all of it, because his leverage was 6.5× and yours is 20×. Bitcoin moves 5% in a session regularly; a mid-cap altcoin can do it during breakfast.
The fix is not a better view — his view was the product of a career. The fix is the number in the leverage box, and sizing the position from the distance to your stop rather than from the profit you are picturing. Our Stage 1 lesson on leverage and margin works through where the exchange actually closes you, and the position size calculator does the arithmetic on your own balance.
PRACTICE CORNER
The lesson of 1994 is not about forecasting — it is about the leverage selector. Open the order screen on any major exchange, set the leverage to the lowest value it offers, and look at how much the liquidation price moves away from you when you do. That distance is the whole difference between a bad month and no account.
Referral links — they never change our assessment. Education only; most retail traders lose money.
Where this lesson stops working
6.5× on bonds is not 6.5× on crypto. Government bonds are among the least volatile assets that exist; Steinhardt’s leverage was aggressive for bonds. The same multiple on a mid-cap token is a categorically different bet. Leverage numbers only mean something next to the volatility of what you are levering.
The reported figures are reported figures. The $30 billion is described by Fortune as what he “reportedly” held, not as a disclosed position. Our 6.5× is an honest division of the best numbers in print, not an audited fact, and we have shown above where the source’s own arithmetic strains.
Low leverage is not automatically safe. An oversized position at 2× can hurt more than a small one at 10×, because what matters is the fraction of the account exposed, not the multiplier on the screen. That is the point of the 1–2% rule.
And a 28-year record is not a method you can copy. Steinhardt had an information edge in an era before electronic markets, a research operation, and clients who tolerated his temper. The lesson transfers. The career does not.
Common mistakes when learning from Steinhardt
Taking “variant perception” as permission to be big. A differentiated view tells you what to trade, never how much — and conflating conviction with size is precisely what 1994 punished. Assuming a bad year requires a bad call. His view on European rates was wrong by a couple of hundred basis points, which is small in forecasting terms. Leverage converted a small error into a career-ending one. Believing a big year cancels a big loss. +26% against −31% still leaves you 13% short — read the second diagram again before planning a recovery. Assuming the size grew because the conviction did. Three straight years above 60% reset what a normal return from the same capital feels like, and position size tends to follow the expectation rather than the evidence. Reading “diversified” off the number of tickers. His bond positions spanned several countries and were still, functionally, one trade — the same way a portfolio of eight altcoins is one bet on liquidity.
FAQ
How much did Michael Steinhardt’s fund actually return?
Steinhardt Partners averaged 24.5% a year from 1967 to 1995, net of a 20% performance fee. Forbes reported that $10,000 invested at the start would have been worth about $4.8 million when the fund closed, against roughly $190,000 in the S&P 500 over the same period. We checked the internal consistency: $10,000 to $4.8 million across 28 years is a compound rate of 24.67%, which matches the headline figure.
What went wrong in 1994?
He held a large leveraged position in European bonds, expecting rates to keep falling. The US Federal Reserve raised rates by 25 basis points in February 1994, leveraged holders everywhere were forced to sell into each other, and European rates rose instead. Fortune reported the fund was down about a third by May. The year finished at minus 31%, the worst of his career.
How much leverage was Steinhardt Partners using?
Fortune reported a Eurobond position of roughly $30 billion against $4.6 billion under management. Dividing one by the other gives about 6.5 times leverage. That ratio is why a bond move of roughly 5% cost a third of the fund: the fund fell about six and a half times as far as the assets it held.
Why did he close the fund after a 26% year?
He said he wanted to earn back what he could for his investors first, and did so in 1995, but that the work no longer justified itself to him. In his words to Forbes: he thought there must be something more virtuous and more ennobling to do with one’s life than make rich people richer. He was in his fifties and left the industry for nearly a decade.