My son asked me last week why we still have to learn about things that happened a long time ago. We talked about how there are lessons about life and from mistakes in history everywhere—in books like the Odyssey and ones about war. And particularly for me, in ones about financial history. How leverage and hubris can be found in most tragic endings. Then markets gave me some new material a few days later.

As many have talked about, the hedge fund Situational Awareness was forced to sell its entire public stock portfolio at a discount to Citadel to meet margin calls last week. The fund had been up a very impressive 439% for the year through June. But by the end of July it had gone from $45 billion to about $10 billion. Of note, 6 days before the forced sale, the firm sent investors a letter calling the drawdown a buying opportunity. 

Now, if you’ve studied financial history, you’ve seen this movie before. It might be from a different decade, and due to a different set of trades, but it has a similar ending. So this week felt like a good time to line up a few of the bigger ones of the last 30 years and see what the common threads, and lessons, are. (This is a longer one)

LTCM, 1998. 

• Run by two Nobel laureates and a former Fed vice chairman, the fund borrowed heavily to capture returns out of small gaps between related bonds. Leverage got up to as much as 130-to-1 by late September 1998, as losses took down the equity values faster than positions could be unwound, triggered by Russia’s defaulting on its debt, which had not been considered in the fund's models. 

• On September 2, 1998, the fund sent a letter to investors, disclosing the losses, but then pivoted to argue that market spreads had become "unusually attractive", asking investors for more capital. 3 weeks later, the fund needed a $3.6 billion emergency bailout from 14 banks to stay alive. LTCM lost most of its $4.7 billion in under 5 months, and its trades were so tangled up with Wall Street banks that the Fed organized a private rescue.

• The market recovered from the LTCM scare in about 3 months and kept climbing for another year and a half before the dot-com bubble popped in March 2000. It was an early sign there might be too much leverage in places.

Bear Stearns' credit funds, and the "Quant Quake" that followed, 2007. 

• Two Bear Stearns funds, leveraged roughly 10-to-1 and 13-to-1, were loaded with subprime-backed CDOs. 

• In May 2007, the firm told investors only 6–8% of assets touched subprime; the real exposure was closer to 60%. Both funds collapsed and filed for bankruptcy by the end of July.

• Then, a separate multi-strategy fund, holding both credit and quant-equity positions, covered its credit losses by selling its more liquid quant book. Many quant funds were running similar strategies with overlapping positions, so that one fund's selling looked to everyone else's risk models like a signal to sell the same things. In the first week of August 2007, an estimated $500 billion in assets was liquidated in a matter of days, while the stock market itself stayed nearly flat. 

• The market topped out about 3 months later, before the 2008 crisis. This was a warning sign of the crisis, in its earliest visible form.

Archegos, 2021. 

• Bill Hwang's family office was about $20 billion in equity, but controlled up to $160 billion in stock exposure through total return swaps. This structure let the fund build large, concentrated positions without the disclosures that owning the stock directly would have required. No single bank could see the full positions across all the others. 

• On March 24, 2021, with his most concentrated holding already down sharply, his team reportedly told several banks the issue was a "liquidity issue, not a solvency issue." Two days later, those banks sold roughly $20–30 billion of stock to cover their own exposure. Archegos lost about $20 billion in 48 hours.

• The market topped in January 2022 before that year's bear market, triggered by the Fed hiking rates into inflation. BUT, rates were so low before that, along with Quantitative Easing (QE), investors (and even some banks) were lulled into complacency, leading some to take major risk. It also drove inflation higher. When that started to unwind, we learned who took on too much risk.

The common threads, in order of how much they explain:

• Leverage. The ratios may have been different, but the mechanism is the same. Borrowed money means your lender could decide when the trade ends, instead of you. 

• Crowded or concentrated positions turned an isolated loss into a cascade. When investments are commonly held or linked and something goes wrong, a lot of people try to exit through the same door at once.

• A reassuring message tended to go out right before the end. "A liquidity issue, not a solvency issue." Spreads were “unusually attractive.” A letter calling a historic drawdown a buying opportunity. None of these were necessarily dishonest in the moment. They may have believed it right up until they couldn’t meet the next call.

Based on the timing of what we’ve studied: is a hedge fund blow up a warning that a bear market is coming?

It certainly can be.

In every one of them, an extended stretch of cheap money and/or easy credit came first. Low rates in the mid-to-late 1990s. Low rates and lax lending standards through the mid-2000s. Near-zero rates and stimulus checks in 2020–2021. That kind of environment inflates risk-taking broadly, and it usually shows up first in the most leveraged, most exposed corners of the market, because that's where excess collects fastest. 

What ends the party, in every case, is the same conditions reversing. As a very seasoned JPMorgan portfolio manager, who refused to own tech in late 1999, once told me: “you see who is not wearing a swimming costume once the tide is out.” (he’s British, adapting Buffet’s quote)

Maybe the best question to ask is: what conditions let this most recent fund get leveraged in the first place, and how many other places are those same conditions showing up?

Situational Awareness was a genuinely leveraged, single-thesis bet that became crowded and broke, which rhymes with all cited cases. Whether it's an early crack in something bigger depends on whether the conditions that let a levered AI trade grow to $45 billion are showing up elsewhere too. Based on our assessment in July (see here), and adding in the big financing bets of the largest companies in the world, we say yes they are. 

But timing is important. We believe this is more like 1998, where there are still returns and analysis to be had, than the summer of 2007, where the market peaked 3 months later. 

So a few habits worth keeping:

• If you're using margin, know the exact price move that triggers a call. If you are investing in levered funds, know that they manage for the day and not longer, so moves are amplified.

• A crowded trade has a small exit door. If a position is popular and everyone knows it, think about what happens to the price if a chunk of those holders need out at the same time.

• Hearing "There's nothing to worry about," while a position is already under pressure, deserves a bit more scrutiny. 

For the Robinhood Strategies team, we’ve lived this history and learned from it, influencing our investment process and how we consider risk and return over time. It’s led us to consider flows, margin book sizes, aggregate short positioning, in addition to fundamental and price factors. We believe this allows us to manage through rougher periods and compound growth over time.