Indeed, the situation was the same across the entire financial system. Banks, hedge funds, consumers, and even countries had been leveraging up and doubling down for years. In August 2007, the global margin call began. Everyone was forced to sell until it became a devastating downward spiral.
Near midnight, Rothman, luggage still in tow, hopped in a cab and told the driver to take him to the Four Seasons. As he leaned back in the cab, exhausted, he pondered his next move. He was scheduled to fly to Los Angeles the next day to visit more investors. But what was the point? The models were toast. Forget it, he thought, deciding to cancel the L.A. trip. I need to hammer out a call.
As the losses piled up at AQR, Asness continued to put in frantic calls to Goldman Sachs Asset Management. But GSAM was in radio silence. At the height of the convulsions, Robert Jones, who ran Goldman’s quantitative equities team, emailed Asness with a terse three-word note: “It’s not me.”
Asness wasn’t so sure. He knew GSAM about as well as anybody outside of Goldman, having launched Global Alpha more than a decade before. And he knew Global Alpha had cranked up the leverage. He looked with horror at how big his creation had become, a lumbering monster of leverage. Asness knew that if GSAM imploded, it would be a disaster.
AQR traders were running low on fuel, high on adrenaline. It was something like the energy of a firefight, full of both fear and grim exhilaration, as if history was in the making.
Asness decided to give a pep talk to his bedraggled quants. Rumors that AQR was on the verge of melting down were spreading. There was no central meeting area in the office, so employees huddled in a number of conference rooms and Asness addressed the troops over speakerphones from his office. Some of the traders thought the setup was strange. Why didn’t Cliff address them directly, face-to-face? Instead, he was just a voice, like the Wizard of Oz behind his curtain. Beside him were partners such as John Liew and David Kabiller, as well as Aaron Brown.
He acknowledged that the fund was seeing unprecedented losses, but told his crew not to panic. “We’re not in a crisis,” he said. “We have enough money to keep the trades on. We can handle the situation.”
He ended the call on an optimistic note, referring back to the dotcom bubble that had nearly crushed AQR. “The partners have been in this situation before. The system works. This is something that we’ll get through, although I understand that it’s difficult.”
But there was one cruel fact Asness couldn’t escape: AQR’s IPO would have to be put on hold, he said. “And it may never come back.”
At Saba, Alan Benson, the trader in charge of its quant fundamental book, was verging on collapse. He was putting in eighteen-hour days, trading like a rat trapped in a maze that seemed to never end. He and only a few other traders ran billions worth of assets, and it was impossible to keep track of it all. The fund had lost about $50 million or $60 million in two days alone, and Weinstein wasn’t happy. He kept pressing Benson to sell and cut his losses as fast as possible.
The losses were brutal throughout quantdom. Tykhe Capital, a New York quant fund named after the Greek goddess of good luck, was in tatters, down more than 20 percent. In East Setauket, Renaissance’s Medallion fund was getting pummeled, as was the Renaissance Institutional Equity Fund, the massive quant fundamental fund Jim Simons had once said could handle $100 billion in assets.
The losses in Medallion, however, were the most perplexing. Simons had never seen anything like it. Medallion’s superfast trading strategy, which acts as a liquidity provider for the rest of the market, was buying up the assets from quant funds that were frantically trying to exit positions. Medallion’s models predicted that the positions would move back into equilibrium. But the snapback didn’t happen. The positions kept declining. There was no equilibrium. Medallion kept buying, until its portfolio was a near mirror image of the funds that were in a massive deleveraging. It was a recipe for disaster.
The losses were piling up so quickly, it was impossible to keep track of them. The Money Grid was on the precipice of disaster and no one knew when it would stop.
At PDT, Muller kept ringing up managers, trying to gauge who was selling and who wasn’t. But few were talking. In ways, Muller thought, it was like poker. No one knew who was holding what. Some might be bluffing, putting on a brave face while massively dumping positions. Some might be holding out, hoping to ride through the storm. And the decision facing Muller was the same one he confronted all the time at the poker table, but on a much larger order of magnitude: whether to throw in more chips and hope for the best or to fold his hand and walk away.
Other managers were facing the same problem. “We were all freaking out,” recalled AQR cofounder John Liew. “Quant managers tend to be kind of secretive; they don’t reach out to each other. It was a little bit of a poker game. When you think about the universe of large quant managers, it’s not that big. We all know each other. We were calling each other and saying, ‘Are you selling?’ ‘Are you?’”
As conditions spun out of control, Muller was updating Morgan’s top brass, including Zoe Cruz and John Mack. He wanted to know how much damage was acceptable. But his chiefs wouldn’t give him a number. They didn’t understand all of the nuts and bolts of how PDT worked. Muller had kept its positions and strategy so secret over the years that few people in the firm had any inkling about how PDT made money. Cruz and Mack knew it was profitable almost all the time. That was all that mattered.
That meant it was Muller’s call. By Wednesday morning, August 8, he’d already decided. The previous day, he’d caught a flight out of Boston as it became clear that something serious had happened. At La Guardia Airport, he was picked up by his chauffeur, a retired police officer. Riding into the city in the backseat of his BMW 750Li, he placed a phone call to the office to gauge the damage.
The losses had been severe, twice as bad as on Monday. He knew something had to be done fast. There wasn’t much time. It was already late in the day. A decision had to be made.
After stopping off at his Time-Warner apartment, he walked down to the office. It was about 7 P.M. He’d come into Morgan’s office to meet Amy Wong, the trader in charge of the quant fundamental portfolio getting clobbered. They huddled in a conference room just off PDT’s small trading floor, along with several other top PDT staffers. Wong tallied up the damage. The quant fundamental book had suffered a loss of about $100 million.
“What should we do?” Wong asked.
Muller shrugged and gave the order: sell.
By Wednesday morning, PDT was executing Muller’s command, dumping positions aggressively. And it kept getting killed. Every other quant fund was selling in a panicked rush for the exits.
That Wednesday, what had started as a series of bizarre, unexplainable glitches in quant models turned into a catastrophic meltdown the likes of which had never been seen before in the history of financial markets. Nearly every single quantitative strategy, thought to be the most sophisticated investing ideas in the world, was shredded to pieces, leading to billions in losses. It was deleveraging gone supernova.
Oddly, the Bizarro World of quant trading largely masked the losses to the outside world at first. Since the stocks they’d shorted were rising rapidly, leading to the appearance of gains on the broader market, that balanced out the diving stocks the quants had expected to rise. Monday, the Dow industrials actually gained 287 points. It gained 36 more points Tuesday, and another 154 points Wednesday. Everyday investors had no insight into the carnage taking place beneath the surface, the billions in hedge fund money evaporating.
Of course, there was plenty of evidence that something was seriously amiss. Heavily shorted stocks were zooming higher for no logical reason. Vonage Holdings, a telecom stock that had dropped 85 percent in the previous year, shot up 10 percent in a single day on zero news. Online retailer Overstock.com; Taser International, maker of stun guns; the home building giant Beazer Homes USA; and Krispy Kreme Doughnuts—all favorites among short sellers—rose sharply eve
n as the rest of the market tanked. From a fundamentals perspective, it made no sense. In an economic downturn, risky stocks such as Taser and Krispy Kreme would surely suffer. Beazer was obviously on the ropes due to the housing downturn. But a vicious marketwide short squeeze was causing the stocks to surge.
The huge gains in those shorted stocks created an optical illusion: the market seemed to be rising, even as its pillars were crumbling beneath it. Asness’s beloved value stocks were spiraling lower. Stocks with low price-to-book ratios such as Walt Disney and Alcoa were getting hammered.
“A massive unwind is occurring,” Tim Krochuk, managing director of Boston investment manager GRT Capital Partners, told the Wall Street Journal. Pissed-off plain-vanilla investors vented their rage on the quants as they saw their portfolios unravel. “You couldn’t get a date in high school and now you’re ruining my month,” was one sneer Muller heard.
Amid the carnage, Mike Reed had an idea: stop selling for an hour to see if PDT itself was driving the action—a clear indication of how chaotic the market had become. No one knew who was causing what. But the desperate move didn’t work. PDT continued to get crushed. There was a deceptive lull soon after lunchtime. But as the closing bell neared in the afternoon, the carnage resumed. Mom-and-pop investors watching the market make wild swings wondered what was going on. They had no way of knowing about the massive computer power and decades of quant strategies that were behind the chaos making a hash of their 401(k)s and mutual funds.
Reed’s intuition that PDT’s decision to sell was hurting the market wasn’t completely wrongheaded. A source of the extreme damage Wednesday and the following day was the absence of high-frequency stat arb traders that act as liquidity providers for the market. Among the largest such funds were Renaissance’s Medallion fund and D. E. Shaw. PDT had already significantly deleveraged its stat arb fund, Midas, the week before. Other stat arb funds were now following suit. As investors tried to unload their positions, the high-frequency funds weren’t there to buy them—they were selling, too. The result was a black hole of no liquidity whatsoever. Prices collapsed.
By the end of the day on Wednesday, PDT had lost nearly $300 million—just that day. PDT was going up in smoke. Other funds were seeing even bigger losses. Goldman’s Global Alpha was down nearly 16 percent for the month, a loss of about $1.5 billion. AQR had lost about $500 million that Wednesday alone, its biggest one-day loss ever. It was the fastest money meltdown Asness had ever seen. He was well aware that if it continued for much longer, AQR would be roadkill.
And there was nothing he could do to stop it. Except keep liquidating, liquidating, liquidating.
Inside GSAM, the grim realization was taking hold that a catastrophic meltdown would occur if all the furious liquidations weren’t somehow stopped. Goldman’s elite traders were running on fumes, staying at the office fifteen or twenty hours—some not leaving at all.
Carhart and Iwanowski, like every other quant manager, were feverishly trying to delever their funds, trying to get their volatility-based risk models back into alignment. But there was a problem: every time GSAM sold off positions, volatility spiked again—meaning it had to keep selling. Higher volatility readings automatically directed the fund to dump more positions and raise cash.
The upshot was terrifying: GSAM was trapped in a self-reinforcing feedback loop. More selling caused more volatility, causing more selling, causing more volatility. It was a trap that had snared every other quant fund.
They were trying to make sure that their positions were liquid, that they could exit them whenever necessary without major losses. But every time they deleveraged positions, they were back to square one. The GSAM team realized, with shock, that they might be trapped in a death spiral. Talk about an LTCM-like meltdown, one that didn’t just take down one giant fund but dozens, started to make the rounds.
“There was a sense that this could be the end,” said one GSAM trader.
And if it continued much longer, it would make LTCM’s collapse in 1998 seem like a walk in the park.
What to do?
Matthew Rothman woke early on Wednesday, August 8, and walked to Lehman Brothers’ San Francisco office on California Street, just around the corner from the Four Seasons. He sent a stream of emails to his quantitative research team back in New York with essentially a single order: get to work on a report explaining the quant meltdown.
But several members of his staff were having trouble making it to Lehman’s midtown Manhattan office on Seventh Avenue. The massive storm had knocked out the city’s subway system, and no one could find a cab. Rothman told them they had to find a way to get into the office, no matter what. Walk, run, ride a horse. Whatever. They had to get this call out.
Rothman, in constant contact with his research staff in New York, spent all day collecting data, working the Street for insight, writing, putting together complicated charts. By the time the note was finished, it was midnight local time, 3:00 a.m. Eastern. Rothman stumbled back to the Four Seasons, exhausted.
“Over the past few days, most quantitative fund managers have experienced significant abnormal performance in their returns,” he wrote with classic Wall Street analyst understatement. “It is not just that most factors are not working but rather they are working in a perverse manner, in our view.”
The report continued with the scenario Rothman had worked out over sushi with Levin: “It is impossible to know for sure what was the catalyst for this situation. In our opinion, the most reasonable scenario is that a few large multi-strategy quantitative managers may have experienced significant losses in their credit or fixed income portfolios. In an attempt to lower the risks in their portfolios and being afraid to ‘mark to market’ their illiquid credit portfolios, these managers probably sought to raise cash and de-lever in the most liquid market—the U.S. equity market.”
The following pages of the report were a detailed examination of the specific trades that were blowing up. The oddest section, however, was its conclusion, a terse reiteration of the quant credo that at the end of the day, people—and investors—generally behave in a rational manner. The Truth, after all, is the Truth. Right?
“We like to believe in the rationality of human beings (and particularly quants) and place our faith in the strong forces and mutual incentives we all have for orderly functioning of the capital markets,” Rothman wrote. “As drivers of cars down dark roads at night, we learn to have faith that the driver approaching on the other side of the road will not swerve into our lane to hit us. In fact, he is just as afraid of our swerving to hit him as we are of his swerving to hit us. We both exhale as we pass by each other headed into the night in our respective opposite directions, successfully avoiding both of our destructions.”
The report, called “Turbulent Times in Quant Land,” was posted on Lehman’s servers early that morning. It quickly became the most highly distributed note in the history of Lehman Brothers.
As word of his report seeped out, he got a call from Wall Street Journal reporter Kaja Whitehouse. When asked to describe the severity of the meltdown, Rothman didn’t mince words.
“Wednesday is the type of day people will remember in quantland for a very long time,” Rothman said. “Events that models only predicted would happen once in 10,000 years happened every day for three days.”
He spoke as though the worst were over. It wasn’t.
Early Thursday morning, August 9, PDT held a series of emergency meetings in Peter Muller’s office. The situation was dire. If the fund lost much more money, it ran the risk of getting shut down by Morgan’s risk managers—a disaster that could mean the group would have to liquidate its entire portfolio. Reed advocated even more aggressive selling. Muller agreed but wanted to wait one more day before ratcheting up. Meanwhile, the losses were piling up.
By then, the quant meltdown was affecting markets across the globe. The Dow Jones Industrial Average tumbled 387 points Thursday.
The Japanese yen, which quant funds li
ked to short due to extremely low interest rates in Japan, surged against the dollar and the euro—an example of more short covering by quant funds as the carry trade fell apart. But the dollar rose against most other currencies as investors snapped it up in a panicked flight to safe, liquid assets, just as they had during Black Monday in October 1987 and in August 1998 when LTCM imploded.
On Friday morning, Muller came into the office early. The plan was to deleverage like mad before everything was wiped out. But before he gave the thumbs-up on the plan, Muller wanted to see what happened at the opening bell. You never know, he thought. Maybe we’ll get a break. But he wasn’t counting on it.
There was plenty of bad news to worsen the mood. France’s largest publicly traded bank, BNP Paribas, froze the assets of three of its funds worth a combined $2.2 billion. In a refrain that would echo across financial markets repeatedly in the coming year, BNP blamed the “complete evaporation of liquidity” in securitization markets related to U.S. housing loans, which had “made it impossible to value certain assets fairly regardless of their quality or credit rating.”
Late Thursday, Jim Simons had issued a rare midmonth update on the state of one of his funds. The Renaissance Institutional Equities Fund, which managed about $26 billion in assets, was down 8.7 percent so far from the end of July—a loss of nearly $2 billion.
In a letter to investors, Simons attempted to explain what had gone wrong. “While we believe we have an excellent set of predictive signals, some of these are undoubtedly shared by a number of long/short hedge funds,” wrote the white-bearded wizard of East Setauket. “For one reason or another many of these funds have not been doing well, and certain factors have caused them to liquidate positions. In addition to poor performance these factors may include losses in credit securities, excessive risk, margin calls and others.”
The Medallion fund, however, was doing even worse than RIEF. It had dropped a whopping 17 percent for the month, a loss of roughly $1 billion. Like Muhammad Ali getting licked by Joe Frazier in Madison Square Garden in 1971, the greatest fund of all time was on the ropes, and seemed at risk of getting knocked out.
The Quants Page 26