Risk of Ruin

Inside Long Term Capital

Half Kelly Media

The story of Long Term Capital Management as told by LTCM partner Eric Rosenfeld. Eric earned a PhD from MIT, then taught at Harvard Business School, then became a trader at Salomon, then was a founding partner of LTCM.

Resources:

HBS Study on LTCM dividend

Eric's talk at MIT in 2009

Michael Lewis article on LTCM

When Genius Failed, Roger Lowenstein

Follow the show on Twitter: @halfkelly
Email: riskofruinpod@gmail.com
Web: halfkelly.com

SPEAKER_02:

At this point, we know we're going under. Bear Stearns has already informed us they're not going to give us access to any of our positions or cash after Wednesday. So by definition, if you can't move cash around, you can't make margin calls, you've gone under. Thursday morning, if nothing happens, we're going under. And let me tell you, we've thought about that now for the last three to four weeks, and we know that it's going to be a catastrophe. Forget about whether we still have a business. Forget about our own personal capital that we have in the fund. That is long gone. It's not even close. We're more worried about the wherewithal of the financial system. We did not want to be the ones that take down the entire financial system.

SPEAKER_00:

Risk of Ruin is a podcast about gambling and life and their intersection. I'm John Reeder. This is Episode 12, Inside Long Term Capital. Even though the whole thing played out more than 20 years ago, that doesn't mean anyone has forgotten the story of long-term capital management. To see what I'm talking about, just search Twitter for LTCM and on any given day, there will be dozens of tweets relating it to that day's financial news. There's something obviously compelling about the saga of a startup hedge fund that raised a billion dollars, delivered blowout results for four years. In fact, they were so successful, they sent$2.7 billion back to investors at the end of 1997. And then just 10 months later, they were basically out of business. Further to the point that the story is inherently fascinating, here are some comments Warren Buffett made about the incident in October of 1998.

SPEAKER_01:

The whole story is really fascinating because if you take John Merriweather and Eric Rosenfeld, Larry Hillenbrand, Greg Hawkins, Victor Agani, the two Nobel Prize winners, Merton and Scholz, if you take the 16 of them, they probably have as high an average IQ as any 16 people working together in one business in the country, including at Microsoft or wherever you want to name. So an incredible amount of intellect in that room. Now you combine that with the fact that those 16 had had extensive extensive experience in the field they were operating. I mean, this was not a bunch of guys who had made their money selling men's clothing and then all of a sudden went into the securities business or anything. They'd had, in aggregate, the 16 that probably had 350 or 400 years of experience doing exactly what they were doing. And then you throw in the third factor, that most of them had virtually all of their very substantial net worths in the business. So they had their own money up, hundreds and hundreds of millions of dollars of their own money up, Super high intellect, working in a field they knew, and essentially they went broke.

SPEAKER_00:

But even if everyone remembers the story, that doesn't mean more than the CliffsNotes version has been remembered. For example, I think many people would be surprised to find out that LTCM had positive returns over the life of the fund, including the losses in 1998. This is Eric Rosenfeld. He has a PhD from MIT and was one of the partners at Long Term. The

SPEAKER_02:

average investor had a 19% internal rate of return, including the losses. Only 10 investors lost money, and only 5 investors lost more than, I think, like$5 million. This

SPEAKER_00:

episode has everything you would want in a gambling story. It's about looking for edges, figuring out how much to bet, and perhaps most importantly... having to deal with a heartbreaking loss. But advantage gambling and quantitative finance have more in common than parallel themes. They also share common DNA. Both were born out of the excitement over the power of computing. Eric was at MIT in the early 70s and saw this excitement firsthand.

SPEAKER_02:

This was my sophomore year. I had just bailed from astrophysics to a much more manageable major, econ. I was taking sort of the econ stat econometrics course from this new professor this year, Jerry Hausman. And one day at the end of class, Hausman said, I'm working on this model for predicting horse race numbers. betting. Is there anyone who has an interest? I went up to him after class and said, you know, I have a real interest in this, but I have no interest in betting on the ponies. What I'd like to do is work on a model to predict professional football. And he said, boy, that, you know, that that would work as well. But where are we going to get the data? And the next day I plugged down 15 years worth of data that I had been recording with with all the lines and scores of every game available. So he said, fine. And we set to work building the model. The model I thought was pretty innovative in that you can imagine kind of a linear regression model where you might try to predict the line on the game. So your Y variable could be the line on the game And then you would have all these various explanatory variables. Which team was home? What was their performance in the last games? You know, those sorts of variables. The approach that Hausman took was Rather than predicting the line, because you don't know if the actual line is six points and you're predicting 10 points, you don't know how much of an edge that is. He wanted to actually predict the probability of winning the bet.

SPEAKER_00:

Today we have Python and open source machine learning and the Amazon cloud. But this football model was built in the days of Fortran and punch cards and the IBM 370. Eric says that they won about 60% of their bets in the first season, which was good for about$1,000 at a time when minimum wage was just$2. It didn't do as well in the second season, and they moved on to other things. But they realized that this approach could be used on other problems. I think one

SPEAKER_02:

thing that came out of this might have been one of the first stat arb models that statistical arbitrage in stocks or high frequency trading, it's sometimes referred to. We came up with a model using this technology.

SPEAKER_00:

I love that a professor and a student at MIT were talking about a model to win at gambling. And the only question was, should they start in the horses or sports? I asked Eric if interest in gambling was common.

SPEAKER_02:

I belonged to a fraternity at MIT. It was SAE, and it was known at MIT for either being the sports fraternity, you know, if there was such a thing at MIT, and or the gambling fraternity. There was kind of a number of weekly poker games, lots of backgammon and gin were played. There was chess for money was played at the fraternity. It was pretty prevalent. I think at the time, one of the fraternity brothers had handed me Thorpe's book, which I loved, beat the dealer. And, you know, I had done, you know, after that, I did, you know, some rudimentary card counting in Tahoe. Everyone was looking for an edge. And, you know, I thought about it for pro football betting, but I just didn't know how to get started.

SPEAKER_00:

The NFL model wasn't the only time that Eric used his programming skills. on gambling. He also did some analysis for Manny Kimmel, who has a prominent role in the history of advantage play. In the early 60s, Kimmel bankrolled the real-world test of the card counting system invented by Ed Thorpe.

SPEAKER_02:

One of my backgammon buddies after graduating MIT went to work for this guy, Manny Kimmel, who was the famous Mr. X in Thorpe's book, a professional gambler. He wanted somebody to try to work out the probability for this card game that Manny used to play in. You know, hand one started with the ace of hearts, hand two with the king of hearts, hand three with the queen of hearts, and hand four with the jack of hearts. And then people would bet, what's the probability that the ace beats the jack? And Manny wanted to know what those probabilities were. So I think I just simulated hundreds of thousands of hands on this little computer in a box. I think it probably took about two weeks to get the results, but it actually worked out quite well for Manny. I didn't get the exact detail of how much he made, but for me, I think I might have made$1,000, which... There

SPEAKER_00:

were two concurrent revolutions that Eric was a part of at MIT. One was in computing, the other was in finance.

SPEAKER_02:

I just went down the hall to get a PhD in finance, working for this guy, Bob Merton, who had developed the option pricing formula in 1973, along with Black and Scholes. And, you know, just some amazing things were happening in finance. This whole new paradigm for, you know, looking at finance in a totally new way just had amazing results. And this was just the start of it. I mean, this was 1975 I went over, so this revolution had just started two years before.

SPEAKER_00:

Eric told me about a number of the computers that he was programming in the 70s, and they all sounded like something from the Flintstones. One of them didn't have a monitor, and it communicated with beeps. Another one wrote data to a tape drive. Some of them took hours to get results. One PC was allegedly portable, even though it weighed 50 pounds. But Eric was so mesmerized by the potential of these machines that when he eventually got his own computer, it was off to the races writing programs.

SPEAKER_02:

And I started writing all these finance and statistical programs. And I put an ad in one of the personal computer magazines that I had these programs to trade. And lots of Apple employees said, got back to me and said, boy, I'd love to trade. I've got these Star War games I can trade for your mortgage calculator or statistical package. And one guy who answered the ad was... This unemployed guy who was driving like a 10-year-old beat-up VW Bug. And he said, you know, we should form a local Apple users group. And I think we formed the first, you know, sort of Cambridge, Massachusetts Apple users group. And, you know, it was very early times. And that guy was Mitch.

SPEAKER_00:

The Mitch in the story is the technology pioneer, Mitch Kaeper. of Lotus 1-2-3 fame. One of their shared projects was a desktop version of a stats program called Troll that Eric had seen on the shared terminal machines.

SPEAKER_02:

He asked me, what would I like to do on sort of version two of the programs that I had written? And I said, well, I showed him sort of the troll program and just how easy it was to get results and how doing it in an interactive way allowed you to really understand the problem better. And he said, boy, you know, I could take your program and write the troll interface on top of it. And he did that. And we created this program called Tiny Troll, which was, you know, essentially like a little mini troll. And it was written in basic language and a combination of 6502 assembler. I remember because Professor Koo, who had written the Fortran version of real troll, which ran on the Cornell IBM 370, threatened to sue me and then i showed him our code and he goes oh okay and let matters be so this was the summer of 1979 we had written this program i had just graduated from mit and was going to teach at harvard business school and we decided that summer we were going to uh to sell the program I was making$19,000 a year as an assistant professor at Harvard Business School. That summer, we made$100,000 selling this program. It was unbelievable. And when I started the term at Harvard Business School, I gave a seminar with this guy, Dan Bricklin, who was the inventor of VisiCalc, which was the very first spreadsheet program. He had just graduated from Harvard Business School, and he had written the program along with another MIT guy as a solver for Harvard Business School case studies. So here it was. He had just graduated, and we were holding a seminar. He was going to show the students the power of this VisiCalc, and I was going to show the students Tiny Troll. So it was at the biggest auditorium at Harvard Business School. It held 500 seats. It was packed to the rim. We gave an hour demonstration. At the end of the demonstration, 497 students surrounded Dan and three students, all of which were my students, surrounded me. And I walked away and I talked to Mitch after this and I said, there really is something to this VisiCalc. We've got to pursue it. So we kind of took Tiny Troll and we rebranded it as VisiPlot, VisiStat, and we had the VisiCalc guys market it for us. Three months later, we sold the program to the marketing company for$1.2 million. And I'm making$19,000 a year as a professor. After that, Mitch came back and said, you know what? I want to do the next generation of these programs. I want to put them all together. I want to put spreadsheet, plotting, database, stat. I want to put them all together in one program. And he said, are you in? And I said, hey, look. I'm actually a finance guy. I'm not a computer guy. I got to get on with my career. And he said, okay. A year later, I saw the prototype. It was called 123, Lotus 123. I tried to beg my way back in, but it was too late. And three years later, he sold the company to IBM for$600 million. Wow. That was a bad trade.

SPEAKER_00:

Eric taught for a few years at Harvard, but eventually decided that academia wasn't a perfect fit.

SPEAKER_02:

So I was at Harvard Business School teaching introductory finance for five years. And I have to say, it was getting pretty stale for me. The teaching, I was really just going through the motion. My research kind of suffered. What really excited me was the consulting work, taking these new theories of Merton and others and applying it to real world problem. So one day in May 1984, I was kind of in a bad mood. I was grading exams and I got a call from a Solomon managing director who had been referred by Bill Sharp, the Stanford finance professor, He was looking for one of my students, you know, for a junior fixed income trading position. And Sharpe had said, you know, Rosenfeld's doing some interesting work. Maybe one of his students would work for you. So I told this guy that, you know, I had the perfect person for him, you know, the best guy I could recommend. And when he realized I was talking about myself, he said, well, how old are you? And I said, I'm 30. And he said, that's way too old. Well, two weeks later, I started at Solomon in their fixed income ARB US trading group. I remember my first day. I remember my first day on the job. I kind of sauntered in around nine o'clock. I had a big black coffee. I had a Wall Street Journal. I found my desk. I sat down on it, put my feet up, drinking the coffee with the Wall Street Journal. And this kid, about five years younger And about a foot taller came up to me and he said, I'm your boss. And, you know, I kind of sat up a little in the chair and he said, let me tell you your duties. At 730, you get me breakfast. And at 1130, you get me lunch. Any questions? And I kind of then sat up completely in the chair and said, well, how do you take your coffee?

SPEAKER_00:

When LTCM eventually formed, it was made up mostly of partners that had worked together at Solomon in a group run by the legendary bond trader John Merriweather. So to understand long-term capital, we have to spend some time talking about Solomon in the 80s. In the Michael Lewis book, Liar's Poker, the firm was likened to a jungle, but less dramatic descriptions have described it as a true meritocracy that rewarded people for working hard.

SPEAKER_02:

It was exciting. There was lots of testosterone around. But maybe it was the time of the markets. You had just come off this unbelievably extreme volatility. You had incredible financial intermediation taking place. There were new products that would help people accomplish things. what they were looking for. There was just incredible volume. And to make this work, you had to be fleet of foot. And you had to work together because it spanned multiple areas. And I don't know, for whatever reason, because of Solomon's culture, it was there. Hardworking people that knew they had to work as a team to get things accomplished. And These

SPEAKER_00:

guys were doing a number of different things, but their core strategy was convergence, or relative value trades. For example, a newly issued 30-year treasury might trade at a premium to a treasury that's six months old. So Merriweather's traders would short the expensive bond and buy the cheap bond. The profit comes if the yield on these two bonds converges. The risk is very low because the long and short position offset each other, and the yields tend to move together. But that means that the return is also low. So they would use leverage to increase the risk and increase the return. Then they did similar trades with other asset classes. There's a widespread perception of this group as being shackled to mathematical models, but their strategies required a deep knowledge of how the guts of the market worked. Eric said that a lot of their edge came the old-fashioned way. They were plugged into the best information. You know, but I think some

SPEAKER_02:

of the secret sauce was... Understanding the different parts of the trade. A lot of our trades really depended on financing. We sat right next to the people who traded the government bonds. And on the other side of that, of us on the other side, were the people who financed the firm's positions. So we knew where we could get a hold of those 5% bonds that we wanted to sell. You know, we knew that, you know, these particular bonds were going to get hard to borrow. And we needed to take that into account in terms of, you know, whether we would do the trade, or maybe we could borrow them for extended period of time, you know, to make sure that we, you know, didn't have to unwind the trade early because we couldn't finance the positions. You know, it was kind of a Combination of a lot of things, you know, being able to trade cross markets. Not everyone had that ability. You know, we could trade mortgages, government bonds. We could get involved in interest rate swaps. We could get involved in options. We could get involved in futures. The ability to span different marketplaces, you know, had serious value. You could always buy the very cheapest thing and sell the richest one.

SPEAKER_00:

Since their trades required leverage, it would be fair to wonder how the group did in times of turmoil. One test came in October of 1987. On

SPEAKER_02:

Tuesday, October 20th, when the bond markets opened, in the first hour of trading, we lost our entire year. All the money we had made up until then, we lost the entire year. But... To our credit, the trades we put on the rest of that week were the best trades we've ever done. So I give that credit to Solomon's risk tolerance. It didn't mean that we increased our risk. We took off trades that were still great trades, but put on even better ones.

SPEAKER_00:

In 1991, a scandal changed things forever at Solomon. The head of the government bond desk a guy named Paul Moser, admitted to submitting illegal bids in treasury auctions, and it was a huge deal. Solomon CEO John Goodfriend and President Tom Strauss both resigned. The firm might have been toast if not for Warren Buffett. He was the largest shareholder, and he stepped in as interim CEO to stop the hemorrhaging of the firm's reputation. Eric was promoted to take Paul Moser's place, and part of his job was to work with the government to to prove that the bad behavior had been limited, and it wasn't an entire bank full of bad apples. But this event also undermined John Merriweather's position. I

SPEAKER_02:

mean, we love Solomon. And frankly, even after the scandal... I could see staying at Solomon. Buffett did offer Meriwether, you know, to come back and work at Solomon. He just didn't offer him, you know, kind of anywhere near the position that he had before. And Meriwether decided he didn't want to come back. And at that point, I kind of felt a little more inclined to go with Meriwether than to stay at Solomon. When the crisis was over, I mean, it was clear at that point that Solomon would be fine.

SPEAKER_00:

LTCM started with 11 partners that included Meriwether, Eric, and a number of the other accomplished traders from Solomon. Their pitch was roughly as follows. They were putting in$100 million of their own money. Their proprietary trading group at Solomon had averaged$1 billion a year in profit from 1990 to 1992. And just in case that wasn't enough, the fund would also be advised by the two pioneering academics, Myron Scholes and Robert Merton. The partners took this story and went around the world to raise an additional billion dollars. We got a

SPEAKER_02:

call from Kerry Packer, who is this huge trader and gambler from Australia. He owns a number of media companies and he had heard about us. He wanted to be like our anchor investor and, you know, he tried to negotiate the terms. Our terms were a little steeper than normal because we thought we had something special. It turns out he didn't get, you know, any special terms, but he kind of committed to invest$250 million, which, you know, kind of having that in our pocket gave us a lot of confidence when we started to go around the world on this fundraising. Later, when we got closer, you know, we were just a few months away from people putting in their money and starting the fund. Kerry wanted to meet again just to, you know, go over the terms of his investment. We kind of felt that that was, you know, kind of another shot at the trough where he could try to negotiate special terms for himself. But of course, we didn't. We didn't do that for anyone. So he said, look, I'm in Las Vegas. Why don't you guys come come out and meet me? You know, we can do a little gambling and then we can sit down over dinner and, you know, talk about the terms. So we knew that, you know, Carrie was, you know, just this huge gambler. You know, so Meriwether kind of left it to me for, you know, well, how much money should we you know, what sort of a line should we take out and how much should we be betting? So it was Merriweather, myself, and one other person went out to Vegas. I negotiated us a$100,000, you know, credit line. You know, when we would go out to Vegas or Atlantic City, we were typically betting, you know,$100 or$200 a hand. And so we sit down at the table and I'm, I'm kind of like a first, first position. So I get dealt first and then there's Merriweather, then there's the other guy and then there's Carrie Packer. So it's on me to, you know, make the first bet. So I go, Hey, you know, I'm not going to be, you know, a total, it looks like we're going to play for about an hour. So, you know, I pretend like I'm a big shot and I put down$5,000. And Merriweather puts down$5,000 and the guy next to him puts down$5,000. Carey puts down$75,000. Then he proceeds to put$75,000 behind each of our bets. So, you know, I get a card. I get 10 and the dealer's showing a seven. And I go to Carrie, I say, well, you know, You know, normally I would just play my hand because I do know how to count cards. But in this case, you know, his 75,000 is sitting behind me. So I go, well, Kerry, normally I would, you know, double this down. But, you know, what would you do in this situation? And we proceeded to do that. We had a pretty good run for an hour. You know, we were up, I don't know, most of us were up 10 or$15,000. Kerry made, I think,$1.25 million. He kind of hooched about what I guessed as about$100,000 in chips to the dealer. And then we went to dinner. He reaffirmed his$250 million investment. And a week before the money was due, he reneged.

SPEAKER_00:

Despite the hiccup with Kerry Packer, The partners raised all of the money they were looking for. In 1994, they started trading and things went very well. In their first 42 months, they had just eight losing months. In 1996 and 97, the combined profits were about$3.5 billion. But the fund was so successful that eventually they ran into problems caused by success. Their convergence trades were getting crowded at exactly the time their capital base collapsed. was becoming oversized. In late 97, they forced their investors to take back$2.7 billion in a forced dividend. Then in 1998, everything fell apart. They lost$400 million in May and June. In August, Russia defaulted on its debt, and even though LTCM wasn't highly exposed to Russia, there was a flight to quality that wreaked havoc on long-term's positions.

SPEAKER_02:

So the crisis hit on Friday, August 21st. On that day, we lost 12% of our capital when I think our largest daily loss was before that was about 1.5%. Lots of people were on vacation. Everyone flew home. We met that weekend back at LTCM headquarters. What were we going to do? We knew this was serious and was life-threatening for us.

SPEAKER_00:

One of the core assumptions of LTCM's strategy was that diversification would keep their risk low. They regularly had thousands of individual positions that span geographies and asset classes. Pre-crisis, the correlation of the positions was very low. But in the crisis, everything got bad at the same time. When correlations went to one, the fund's leverage, which had juiced up returns... ended up becoming a major problem.

SPEAKER_02:

We decided that what we had to do was raise additional capital and we had to sell off some of our trades. And we knew we couldn't sell off our trades just around the edges. We had to sell large blocks of our trades. Over that weekend, we actually hired JP Morgan as our banker and we offered them a share of our management company to help save us. And they agreed to put$300 million into the fund as proof that they thought we would be successful. And then they were going to go out to their customers to try to raise additional funds. They were also going to try to market big blocks of our stock to typical players who could purchase it. That Sunday night, we called Warren Buffett, we decided that he was pretty familiar with our risk guard portfolio and we would offer our entire risk guard portfolio to him at attractive prices. We knew he was involved in risk guard, he understood the risks, and we knew that he would think the trade was attractive. He agreed that it was something he would consider and we sent Larry Hillebrand on the first plane on Monday morning to go over those trades and to fill Buffett in on what our situation was. You know, over the course of the next few weeks, JP Morgan was not successful fundraising. In early September, we hired Goldman Sachs to replace JP Morgan. We offered them 50% of our management company, but they too were unsuccessful raising additional funds. They would go out and, you know, in a given day, they would kind of circle, you know, two or 300 million, but we would end up losing that amount of capital in the market. So we made no ground. And And we could not take in, as a fiduciary, we could not take in sort of capital piecemeal. We could only take in a large block of capital from our investors that would be sufficient enough that we knew that the risk of us going under was going to be small. We wanted investors to be only betting on the positions, not investing. on the positions plus whether we would not fail. So we were unsuccessful. And then on Wednesday, September 23rd, it all came to a head. At the request of the president of the New York Fed, he called in all the major LTCM creditors, all the banks. He called them to New York Fed to figure out what they were going to do with the LTCM situation. We thought if we go under, for sure, Bankers Trust and Lehman are going under, and maybe it'll bleed through to Merrill Lynch, and then who knows, maybe it could go to Goldman and Citi. I mean, it would just be this domino effect. where we could be the original trigger. We saw it happening. We were telling lots of people that this was going to happen and how to prevent it, but we were totally unsuccessful, including the Federal Reserve. That's why they were having this

SPEAKER_00:

meeting. In the 11th hour, Long Term got an offer from a group that included Warren Buffett, Goldman Sachs, and AIG, but the bid had a one-hour expiration. And Buffett was barely reachable. He was in the middle of a trip for the birthday of Bill Gates. Here he is talking about the bid in 1998.

SPEAKER_01:

But I got the really serious call about long-term capital about four weeks ago this Friday, whenever it was. It was my granddaughter. I got it in mid-afternoon and my granddaughter was having her... birthday party that evening and then I was flying that night to Seattle to go on a 12-day trip with Gates to Alaska and a private train, all kinds of things where I was really out of communication. But I got this call on a Friday afternoon saying that things were really getting serious there. I'd had some other calls before that the article gets into a few weeks earlier. I know those people, most of them pretty well. A lot of them were Solomon when I was there. And the place was imploding and the Fed was sending people up that weekend. And so between that Friday and the following Wednesday when the New York Fed, in effect, orchestrated a rescue effort, but without any federal money involved. I was quite active, but I was having this terrible time because we were sailing up through these... through these canyons, which held no interest for me whatsoever, in Alaska. And the captain would say, you know, if we just steer over here, we might see some bears and whales. And I said, steer where you've got a good satellite connection. So it was, in fact, there's a picture, unfortunately, where I've got my old faithfuls going off behind me, and I've got my back to it. I'm on the phone, which was the people in the group thought it was kind of funny the way I was working the phone. But we put in a bid on Wednesday morning. By then I was in Bozeman, Montana. We made a bid. It was because it was being done at a long distance and everything. It was really the outline of a bid. But in the end, it was a bid for$250 million essentially for the net assets. But we would have put in$3.75 billion on top of that. And it would have been$3 billion from Berkshire Hathaway,$700 million from AIG, and$300 million from Goldman Sachs. And we submitted that. But we put a very short time fuse on it because when you're bidding on$100 billion worth of securities that are moving around, you don't want to leave a fixed price bid out there very long. Plus, we were worried about it getting shopped.

SPEAKER_00:

The Buffett bid wasn't structured in a way that the partners believed they could accept in an hour. But they tried to offer another way to make the deal.

SPEAKER_02:

Buffett's offer was attracted to us because it meant that the financial system would stay intact. It still left our personal wealth at zero because we had borrowed against our fund holdings and the value of what was left to us was going to be less than our loan. So we would not have the financial wherewithal. And we certainly didn't have a business because, hey, you lose 95%, you don't have a business. And, you know, Buffett didn't want us managing the positions, you know, after what had occurred. So his offer was attractive to us, and we wanted to do it. And we were not negotiating with Buffett. His lawyers were the people we had on the phone. And we thought we had presented a way that we could do it for them. But unfortunately, they were unable to get Buffett back on the phone to confirm it. And nobody was going to do a$4 billion trade without his explicit approval. Since then, students have heard this story and they've asked Buffett, if you have gotten the call, would you have done the trade? And And he says, yeah, of course I would have. And in fact, I've made the calculations. If Buffett had done the trade and he had done what he had said he was going to do, which was leave the positions intact for a year, he would have doubled his money in the first year.

SPEAKER_00:

The meeting at the New York Fed was successful at getting the banks to realize that to save themselves, they had to shore up LTCM. 14 banks agreed to put in$3.6 billion, and that effectively ended the crisis. So now let's back up a little and talk about the unique factors that went into this saga. One of the lightning rods for criticism directed at LTCM has been the forced dividend at the end of 97. Maybe that$2.7 billion could have provided an additional cushion in the crisis, but the issue is actually more complicated than that.

SPEAKER_02:

We saw that opportunities were much less, you know, at the end of 1997 than they were at any time kind of in Solomon or in LTCM. The opportunity set was much lower. And in fact, we went out to investors and we told them that. We said, you know, rather than earning sort of 40 and 50 percent returns, we think we're going to earn more. around 15% returns in 1998. And we asked them to take a forced dividend. And of course, they all came back to us. No one wanted to take a dividend. They all asked if they could invest more.

SPEAKER_00:

The partners considered other options, like raising a much larger fund that would have aimed for smaller returns. But ultimately, they decided to do what they had always done, which was to go for higher returns with a smaller capital base. If you see that the

SPEAKER_02:

opportunity set is less, maybe you shouldn't be fully invested. Maybe you should only be 50% invested and wait for the opportunity to enlarge and then scale into your portfolio. So we struggled with that question, and we decided that we would always be fully invested, independent of the opportunity set. You might say, well, that kind of doesn't make sense. You guys are kind of monitoring these trades. You know the best that's going on. You kind of see everything. You know the opportunity set's going to kind of go out. Why don't you... leaves some dry powder. And what we thought was it's not our business to be changing whether we're fully invested or not. That's really investor's decision. When they come to us, they're making an allocation, they're deciding how much they wanna allocate to these relative value investment opportunities. If they know that we're always gonna be fully invested, they can make the decision as to how much to invest.

SPEAKER_00:

The idea that LTCM had become too large is even further supported by their experience trying to exit trades in 97. Their sales were having enough impact on the market that profits were being eroded. In the crisis, their size was even more of a problem.

SPEAKER_02:

You know, so now you have to get back to sort of risk compliance. You can either do that by selling positions or by raising additional capital. The problem is when you're a large impact player and the markets are extremely illiquid as they were, you know, in these late days of August and you're a large impact player. So, yeah. So the downward spiral is there for sure. If you're a small player, you probably could get back to risk compliance and you'd be okay.

SPEAKER_00:

There's a baked-in problem for people that make money from short-term mispricings. You need the market to be a little irrational. You just don't want it to be too irrational. If

SPEAKER_02:

you sell 16% volatility vis-a-vis options, you're betting that the market moves less than 1%. 1% a day, kind of on average, you'll make money. If it moves more than 1%, you lose money. I mean, maybe it could move 1% a day. Maybe it could move one and a half percent a day, you know, maybe on the outset, you know, 2% a day. And maybe that would be for short periods of time. The marketplace for this, you know, kind of volatility type trades for a five year period was at between three and four percent today. It was just, you know, outrageous. But, you know, what could you do? You know, it was it was a somewhat esoteric trade and the market was frozen. Obviously, once the consortium came in and it was clear There wasn't going to be mass liquidation of these trades. It came back to a more normal number of, you know, one to one and a half percent per day.

SPEAKER_00:

The fund was up against the usual problems that you see in crises, like the total disappearance of buyers. But what started as a flight to quality eventually turned into a crisis that was specifically about LTCM. The other firms with similar positions could see these guys teetering, so there was a rush to get out of the way. Also, if you were a counterparty worried the fund was going under, you might fudge things a little. to try to get paid early.

SPEAKER_02:

These particular bonds were for hurricane catastrophes in southern Florida. You know, it's September for us. These bonds mature November 1st when the hurricane season is over. You know, kind of before the LTCM crisis, these bonds are kind of trading at... around 105, a little above par. So, you know, so the reason they're trading above par is because people like the ability to earn, you know, kind of three or 4% or over normal interest rates and they don't see a hurricane on the horizon. So we finance this position. The bond is, let's say it's a million dollar face bond. you know, of the bond, you know, we go to someone like Merrill Lynch and we lend them that bond and they give us a million,$50,000, you know, against that bond because it's trading above par, you know, as a loan. All of a sudden in the crisis with not a cloud in the sky, Merrill says, we don't think this bond is worth anything. a little over par. We think it's only worth 70 cents on the dollar. We're going to reduce the loan that we give you on that. Okay, so that's one thing. The other issue is Merrill is marking our position. So now you now mark our bond at 70 cents on the dollar. We immediately take, you know, kind of a 35 point loss on that bond.

SPEAKER_00:

even though there's been no trades. Long-term has also been criticized for falling in love with their trades and being unwilling to abandon losers, or in some cases, doubling down on trades that had gone against them.

SPEAKER_02:

We had a lot of experience in these trades. We've seen trades move against us and we've made mistakes and we have to take them off. or we've seen trades move against us and we think it's just a better opportunity and we scale up the positions. Usually, in addition to just trying to look at the models themselves, we try to determine if there was a qualitative reason as to why the spreads went against us. Maybe there was some short-term dislocation And we thought that it could easily erase. But it could be something like there was a tax change and it would be a permanent shift. And then, of course, we'd have to have to take the position off, you know, at a loss. The other thing I would point to is we didn't put traders in one particular set of strategies and leave them there. We tried to rotate people around in a lot of different strategies. So when we're now scrutinizing over a position that's gone against us, yes, we've got the people that are currently running the strategies, but lots of other people who have been involved in these strategies in the past are chiming in with their thoughts. So I think we were able to get a lot of information kind of quickly, trying to assess why these spreads went against us.

SPEAKER_00:

The thing that is often referred to as a bailout was really LTCM's creditors putting money into the fund so that the positions could be unwound in an orderly manner. They avoided a fire sale on long-term assets, which helped the creditors and the rest of the system more than it helped any investor or partner in LTCM. So not only did the creditors recover the collateral that would have been in doubt in an LTCM failure, But the additional money they put in actually made about 10%. You know, we thought it was

SPEAKER_02:

important that we clean up our own mess. And we did also still have investors in the fund. We wanted them to get, you know, the best they could out of this. You know, we fought hard with the consortium to... try and reap whatever value there was. We fought to kind of have two objectives. One was to liquidate as quickly as possible, but also to reap whatever inherent value was still left in the portfolio. And I think we were able to get a pretty good mix of that. And when we finished... I think the consortium basically said, hey, you know, we really appreciate what you guys did here and we're going to help you start your next venture, which was, you know, another fund, albeit, you know, much smaller.

SPEAKER_00:

The most well-known account of the history of LTCM is the Roger Lowenstein book, When Genius Failed. Considering that LTCM is still widely talked about today and considering that the Lowenstein book is the only account that most people know it's fair to say that it has been hugely influential as financial history

SPEAKER_02:

There were some serious risk management mistakes that were made at LTCM. We wanted people to learn from that. We thought the Lowenstein book was a good opportunity, you know, to teach those lessons. And he was much more, he was not interested in doing that. I think he was capable of doing it, but he didn't. It was pretty hard for us because when the group of banks came in, they prohibited us from dealing with the press after that. We were not really allowed to tell our story. We kind of secretly went to Michael Lewis and had him do it. And then the banks really clamped down on us. So the Lowenstein book was our first opportunity. And unfortunately, it wasn't very helpful. from our point of view. He had actually kind of made up his mind before, you know, he kind of liked the Buffett type investing, which is, you know, kind of, you know, really good intuition, a few key concepts, and, you know, kind of staying the course. And models just won't work. And we were unable to kind of convince him otherwise. And it was a real disappointment to us. Models, of course, have their limitation. But as I said before, we didn't swear by these models. We just tried to use these models to help us kind of understand the trade, understand what the risks were of the trade and kind of what were the events that might lead to convergence or divergence and what would happen in particular regions and what to watch out for. We would just never do a trade if it looked good with the model, and we didn't understand why it was a good opportunity.

SPEAKER_00:

One opportunity that I think this podcast offers is to hear the story of LTCM from someone that was there. If you're going to try to learn from what happened, then I think it makes sense to take seriously how smart these guys are and how much thought went into their decisions. By contrast, if the entire episode... is an exercise in schadenfreude, and the failure of long-term was just karmic retribution for the sin of overconfidence, then I think there's a limit to what might be learned. After things were wrapped up at LTCM, Eric and a number of the other partners joined John Merriweather again to start a new fund called JWM. The plan was to make similar trades with less leverage. I remember

SPEAKER_02:

calling on one former LTCM investor. This guy is, you know, a mega Brazilian billionaire. And he was kind of running a family office with his son. And I went to see them a few weeks before we closed the new fund. And I told them sort of the tale of woe of LTCM and what had happened and what we were going to do in the new fund to prevent this from happening again. And about three quarters of the way through the presentation, the son had to leave to go take care of something. And I just finished the story with the father. And when I was done, I said, you know, can you give me a little bit of feedback? Would you be willing to invest in our new fund? You liked the first one, and now we've made some changes. And he said, there's no way I'm investing with you guys. You're way too risky for me. I'd be out of my mind. I said, okay. You know, I hear what you're saying. You know, thank you for your time. So I went to leave and at the elevator, I saw the son. I said, you know, I just finished up with your father, but I'd be curious, you know, just if you could give me a little feedback, would you be interested in investing in the new fund? And the son looked at me and said, you know, you've been so beaten up. I don't see you taking any risks again. So there's no way I'm going to invest with you because you're not really going to lever up the positions at all. So it was kind of that story of, you know, everybody had an excuse as to why not to invest with us. So it was pretty tough going early on. Think of it as significantly reduced risk. and much more liquid. And we got the sort of returns that we were expecting. You know, they were on the order of, you know, kind of nine to 12% with much less risk. And then 2008 happened and it was, you know, an order of magnitude worse than 1998. You know, we had hit the portfolio hard and at the end of the day, The portfolio was down 40%, but it followed this liquidation procedure whereby we liquidated everything and the portfolio was down 40% and we gave back the money to investors and closed the fund down. So everything did work as planned, both the expected return sort of pre-crisis and in the crisis, we were able to handle it. Actually, when I say we, I wasn't there at the time. I obviously saw the crisis coming, so I got out early. Just kidding.

SPEAKER_00:

It's true that LTCM and JWM both closed in the midst of economic crises, but it's also true that both funds had positive returns over their lifetimes, even including the losses in the years they closed. And when you add those results... In retrospect, these positions should

SPEAKER_02:

not be run, you know, in something like an investment bank or in a hedge fund. They belong in the you know, very lightly levered portfolios, fixed income portfolios that are looking to get, you know, just a slight edge because you have to be, you can't run a business where, yeah, you're going to make good money for six years and then you're going to get cleaned in the seventh. You know, that's no business. The business is you need to be able to withstand these occasional crises.

SPEAKER_00:

When I think about the number of prominent and accomplished people Eric has been connected to over his career, it's sort of mind-blowing. But perhaps no one is more aware than Eric exactly how uncommon his life has been. I

SPEAKER_02:

feel so lucky. I mean, it's the mentoring that I got along the way, I... I just can't tell you how lucky I was that back as an undergraduate that somebody like Jerry Hausman took me under his wing And not only did he teach me things, but he really, you know, gave me some, you know, great life advice to go study under Merton. And then the time and effort that Bob Merton placed in me, just bringing me along, helping with my career. And, you know, even the advice that he gave in the midst of the LTCM crisis, it was great to have. And Meriwether as well. you know, willing to take a risk on an academic, which, you know, traditionally has been, you know, kind of a, you know, more macho, you know, trading position and, you know, teaching me along the way and living with my mistakes and, you It kind of reminds me a little bit of I taught with a former colleague. We taught at MIT and we taught at Yale. It was kind of my way of trying to give back for the mentoring I had. And You know, this professor that I teach with, you know, tells a story of... He worked, by the way, he worked at Salomon and then went to Deutsche Bank. So he worked in our group at Salomon and he was also an MIT PhD. And his son went to him and said, Dad, you've had such a successful career. How did you get to be so successful? And this guy said to his son, he said, son, there's really two ways you can go about being successful. You can either work really hard or you can be lucky. And the kid goes to the father, he says, so how do you get to be lucky?

SPEAKER_00:

Risk of Ruin is written and produced by me. Special thanks to Eric Rosenfeld. If you want to know more about this episode of Financial History, there are some additional resources you might not have seen. Harvard Business School did a case study on the issue of the forced dividend in 1997. There's also a Michael Lewis article that came out shortly after the consortium banks took over. Eric has given talks to students over the years and one of the talks is on YouTube. So I'll also post links to these things in the show notes. To contact the show, you can email us at riskofruinpod at gmail.com. You can also follow us on Twitter at Half Kelly. For more information about the podcast, you can find us on the web at halfkelly.com.