WEBVTT 00:00:00.000 --> 00:00:02.439 Welcome back to the Deep Dive. Today, we are 00:00:02.439 --> 00:00:04.379 not just talking about money. We are talking 00:00:04.379 --> 00:00:07.400 about what might be the single most expensive 00:00:07.400 --> 00:00:11.419 private education in the history of Wall Street. 00:00:11.769 --> 00:00:13.369 That's a great way to put it. We're going to 00:00:13.369 --> 00:00:15.529 unpack a story that involves, and I don't use 00:00:15.529 --> 00:00:18.050 this term lightly, the absolute smartest people 00:00:18.050 --> 00:00:20.969 on the planet. Truly. I mean, certified geniuses. 00:00:20.969 --> 00:00:23.510 And these certified geniuses managed to effectively 00:00:23.510 --> 00:00:28.309 set a pile of $4 .5 billion on fire. And not 00:00:28.309 --> 00:00:31.050 over a decade, but in a matter of, what, less 00:00:31.050 --> 00:00:33.170 than four months. It's the ultimate Icarus story. 00:00:33.549 --> 00:00:35.710 You know, it has absolutely everything, hubris, 00:00:35.810 --> 00:00:39.270 genius, what seemed like unlimited capital and 00:00:39.270 --> 00:00:43.229 just a catastrophic ending that almost, and this 00:00:43.229 --> 00:00:45.109 is not an exaggeration, almost took down the 00:00:45.109 --> 00:00:47.729 entire global economy. We are talking, of course, 00:00:47.750 --> 00:00:51.909 about long term capital management or LTCM, as 00:00:51.909 --> 00:00:54.149 everyone came to know it. And before we really 00:00:54.149 --> 00:00:55.969 get into the weeds here, I just want to frame 00:00:55.969 --> 00:00:58.149 the stakes for everyone listening. This is important. 00:00:58.570 --> 00:01:00.429 This isn't just a story about some hedge fund 00:01:00.429 --> 00:01:03.210 that had a bad quarter. This is a story about 00:01:03.210 --> 00:01:06.450 a fund that became so big, so leveraged and so 00:01:06.450 --> 00:01:09.790 interconnected that when it started to crack 00:01:09.790 --> 00:01:12.250 in 1998, the Federal Reserve, we're talking the 00:01:12.250 --> 00:01:14.329 Central Bank of the United States, had to step 00:01:14.329 --> 00:01:16.790 in, had to orchestrate a bailout. They were genuinely 00:01:16.790 --> 00:01:20.150 terrified that if LTCM fell, it would cause a 00:01:20.150 --> 00:01:23.400 complete systemic freeze. And that's not hyperbole. 00:01:23.400 --> 00:01:25.040 I mean, the real fear, the thing that kept the 00:01:25.040 --> 00:01:27.400 Fed chairman up at night was that the actual 00:01:27.400 --> 00:01:29.500 plumbing of the global financial system would 00:01:29.500 --> 00:01:33.500 just stop working. So if this one fund in Greenwich, 00:01:33.560 --> 00:01:36.480 Connecticut, went under, ATMs might stop dispensing 00:01:36.480 --> 00:01:39.019 cash. Corporate payrolls might bounce. I mean, 00:01:39.019 --> 00:01:40.760 that was the level of panic we're talking about. 00:01:40.780 --> 00:01:44.760 A true meltdown. Okay. So our mission today is 00:01:44.760 --> 00:01:47.620 to reverse engineer this disaster. We're going 00:01:47.620 --> 00:01:49.280 to look at the dream team they assembled. We'll 00:01:49.280 --> 00:01:52.239 get into the... the specific trading strategies 00:01:52.239 --> 00:01:54.159 that made them billions in the first place. And 00:01:54.159 --> 00:01:57.420 then the fatal flaw, the mathematical blind spot 00:01:57.420 --> 00:01:59.159 in their risk models that they just couldn't 00:01:59.159 --> 00:02:02.530 see. And then, yeah, the minute by minute panic 00:02:02.530 --> 00:02:04.930 of the bailout. We've got a massive stack of 00:02:04.930 --> 00:02:08.250 sources here. Biographies, Fed transcripts, trading 00:02:08.250 --> 00:02:11.030 logs, academic autopsies of the crash. I mean, 00:02:11.030 --> 00:02:13.610 this is one of the most studied financial events 00:02:13.610 --> 00:02:16.030 in history. It is. And I think, you know, if 00:02:16.030 --> 00:02:17.870 there's one big theme to look out for today, 00:02:18.030 --> 00:02:19.949 it's the difference between intelligence and 00:02:19.949 --> 00:02:21.710 wisdom. Oh, that's good. You're going to hear 00:02:21.710 --> 00:02:23.870 about Nobel Prize winners. You're going to hear 00:02:23.870 --> 00:02:27.550 about guys who literally invented the math that. 00:02:27.900 --> 00:02:29.939 you know, Wall Street still uses today. They 00:02:29.939 --> 00:02:32.919 were so incredibly intelligent. Beyond intelligent. 00:02:33.340 --> 00:02:35.099 But as we'll see, they lacked a certain kind 00:02:35.099 --> 00:02:40.419 of street smarts, a survival instinct. They mistook 00:02:40.419 --> 00:02:42.740 the map for the territory. I love that. The map 00:02:42.740 --> 00:02:44.560 is not the territory. Let's start at the very 00:02:44.560 --> 00:02:46.680 beginning then, the assembly of this team. Because 00:02:46.680 --> 00:02:48.699 like you said, this wasn't just some guys in 00:02:48.699 --> 00:02:50.479 a garage with a Bloomberg terminal. This was 00:02:50.479 --> 00:02:54.580 the 1927 Yankees of finance. I really was. And 00:02:54.580 --> 00:02:57.159 it all starts with one man, John Merriweather. 00:02:57.479 --> 00:03:00.360 Who is this guy? John Merriweather is or was 00:03:00.360 --> 00:03:04.080 a legend. Before LTCM, he was at Salomon Brothers 00:03:04.080 --> 00:03:07.180 in the 1980s. Now, if you've ever read Michael 00:03:07.180 --> 00:03:09.900 Lewis's book, Liar's Poker. Oh, a classic. You 00:03:09.900 --> 00:03:11.919 know the culture of Salomon at that time. It 00:03:11.919 --> 00:03:15.560 was aggressive. It was loud. It was full of, 00:03:15.580 --> 00:03:17.919 to use their terminology, big swinging dicks. 00:03:17.939 --> 00:03:20.659 Right. A pure alpha male environment. Absolutely. 00:03:20.919 --> 00:03:24.539 But Merriweather was. He was different. He was 00:03:24.539 --> 00:03:27.080 the exception to the rule. How so? He was quiet. 00:03:27.199 --> 00:03:30.479 He was cerebral, you know. He wasn't the guy 00:03:30.479 --> 00:03:32.419 screaming across the trading floor, throwing 00:03:32.419 --> 00:03:34.800 phones. He was the guy staring at the numbers, 00:03:34.939 --> 00:03:37.520 seeing these tiny patterns that nobody else saw. 00:03:37.680 --> 00:03:40.020 And he ran the bond arbitrage desk, right? He 00:03:40.020 --> 00:03:42.479 did. Yeah. And to give you a sense of just how 00:03:42.479 --> 00:03:45.539 important he was to that firm, in the late 80s 00:03:45.539 --> 00:03:47.919 and early 90s, Merriweather's specific group, 00:03:48.060 --> 00:03:50.580 just his desk, was reportedly responsible for 00:03:50.580 --> 00:03:53.310 something like 80 to 100 % of Salomon Brothers' 00:03:53.509 --> 00:03:55.669 total global earnings. Wait, wait, say that again? 00:03:55.710 --> 00:03:58.469 80 to 100 %? Of the entire bank's profits. That 00:03:58.469 --> 00:04:01.009 is just, that's wild. One desk was basically 00:04:01.009 --> 00:04:03.310 carrying a global investment bank on its back? 00:04:03.650 --> 00:04:06.430 Exactly. He was the golden deuce. Which is why 00:04:06.430 --> 00:04:08.530 it was such a big deal when he left in 1991. 00:04:09.270 --> 00:04:11.710 There was a trading scandal involving treasury 00:04:11.710 --> 00:04:14.449 auctions. Now, it wasn't something he did personally, 00:04:14.530 --> 00:04:16.810 mind you, but it happened on his watch. So he 00:04:16.810 --> 00:04:18.370 took the fall for it? He took responsibility, 00:04:18.569 --> 00:04:21.930 yes. And he resigned. But his reputation was 00:04:21.930 --> 00:04:26.009 so pristine, so legendary, that when he decided 00:04:26.009 --> 00:04:27.930 to start his own shop a couple of years later 00:04:27.930 --> 00:04:30.970 in 93, people were practically throwing money 00:04:30.970 --> 00:04:33.290 at him. They were lining up. But he didn't just 00:04:33.290 --> 00:04:35.569 bring his old traders with him, did he? This 00:04:35.569 --> 00:04:37.389 is where the story gets really interesting. He 00:04:37.389 --> 00:04:40.310 went out and recruited the academic heavyweights. 00:04:40.449 --> 00:04:43.389 This was his brilliant move. He decided to bridge 00:04:43.389 --> 00:04:47.009 the gap between the ivory tower of academia and 00:04:47.009 --> 00:04:49.509 the dirty fingernails world of the trading floor. 00:04:49.920 --> 00:04:52.639 He recruited the intellectual giants of finance. 00:04:52.980 --> 00:04:55.879 Specifically, Myron Scholes and Robert C. Merton. 00:04:56.180 --> 00:04:58.500 Yep. Now, for anyone who hasn't suffered through 00:04:58.500 --> 00:05:00.800 a university finance degree, those names are, 00:05:00.879 --> 00:05:03.339 well, they're effectively gods. These are the 00:05:03.339 --> 00:05:05.279 Scholes and Merton you read about in the textbooks. 00:05:05.800 --> 00:05:08.740 Precisely. They, along with a third academic 00:05:08.740 --> 00:05:11.639 named Fisher Black, created the Black -Scholes 00:05:11.639 --> 00:05:15.560 model. This is the formula, the math, that allows 00:05:15.560 --> 00:05:19.199 you to accurately price stock options. Before 00:05:19.199 --> 00:05:22.779 them, Pricing a derivative was. It was a guessing 00:05:22.779 --> 00:05:25.600 game. It was all intuition and gut feeling. They 00:05:25.600 --> 00:05:27.660 turned it into a science. They turned finance 00:05:27.660 --> 00:05:30.209 into physics. That's the key. They gave it the 00:05:30.209 --> 00:05:33.230 veneer of scientific certainty. And just to underscore 00:05:33.230 --> 00:05:36.329 how big a deal this was, these two guys, Scholes 00:05:36.329 --> 00:05:38.629 and Merton, they won the Nobel Prize in economics 00:05:38.629 --> 00:05:42.069 in 1997. While they were working at LTCM. Right. 00:05:42.149 --> 00:05:44.810 So imagine your potential investor. You're sitting 00:05:44.810 --> 00:05:47.410 in a meeting and the pitch is, we have the greatest 00:05:47.410 --> 00:05:49.970 bond trader of his generation, John Merriweather. 00:05:50.069 --> 00:05:52.709 And on our board, we have the guys who literally 00:05:52.709 --> 00:05:55.149 wrote the laws of physics for the market. Oh, 00:05:55.170 --> 00:05:56.790 and by the way, they just won the Nobel Prize. 00:05:56.970 --> 00:05:58.560 It created this aura of infallibility. availability, 00:05:58.779 --> 00:06:00.980 didn't it? It wasn't gambling anymore. It was 00:06:00.980 --> 00:06:03.079 science. They weren't betting on gut instinct. 00:06:03.319 --> 00:06:05.100 They were harvesting what they saw as mathematical 00:06:05.100 --> 00:06:08.180 certainties. And the structure of the fund itself, 00:06:08.420 --> 00:06:11.319 it sort of reflected that exclusivity, that arrogance 00:06:11.319 --> 00:06:13.579 almost. They weren't on Wall Street. No. They 00:06:13.579 --> 00:06:15.720 set up in Greenwich, Connecticut. They wanted 00:06:15.720 --> 00:06:18.480 to be physically removed from the noise and the 00:06:18.480 --> 00:06:22.100 emotion, the panic of New York City. It was designed 00:06:22.100 --> 00:06:25.040 to be very academic, very quiet, just supercomputers 00:06:25.040 --> 00:06:27.620 and silence. And very unregulated. Oh, absolutely. 00:06:27.899 --> 00:06:30.199 They used an exemption in the Investment Company 00:06:30.199 --> 00:06:33.279 Act of 1940. Basically, if you only take money 00:06:33.279 --> 00:06:35.980 from highly sophisticated, very wealthy investors, 00:06:36.180 --> 00:06:38.639 you don't have to tell the SEC, the regulator, 00:06:38.860 --> 00:06:41.620 much of anything. Which means they could be secretive. 00:06:41.879 --> 00:06:44.370 Extremely. They told their investors absolutely 00:06:44.370 --> 00:06:46.610 nothing about what they were actually trading. 00:06:46.730 --> 00:06:49.750 It was a total black box. The deal was, you give 00:06:49.750 --> 00:06:51.990 us the money, we give you incredible returns, 00:06:52.209 --> 00:06:54.790 and you do not ask how the sausage is made. And 00:06:54.790 --> 00:06:56.569 people lined up for that deal. They couldn't 00:06:56.569 --> 00:06:58.610 get the money in fast enough. By the time they 00:06:58.610 --> 00:07:01.769 launched in February 1994, they had amassed just 00:07:01.769 --> 00:07:05.829 over $1 .01 billion in capital. For a hedge fund 00:07:05.829 --> 00:07:08.029 launch at that time, that was a massive, massive 00:07:08.029 --> 00:07:10.730 number. Okay, so the dream team is assembled. 00:07:11.319 --> 00:07:13.720 They have the capital. They have the Nobel Prize 00:07:13.720 --> 00:07:16.680 winning brains. Now we need to understand the 00:07:16.680 --> 00:07:19.639 machine. How did they actually make those billions? 00:07:20.120 --> 00:07:23.819 Because the phrase bond arbitrage sounds incredibly 00:07:23.819 --> 00:07:26.839 boring. It does, doesn't it? But clearly it was 00:07:26.839 --> 00:07:29.839 incredibly lucrative. It's lucrative if you do 00:07:29.839 --> 00:07:33.620 it at their scale. The core strategy was something 00:07:33.620 --> 00:07:35.800 called convergence trading. All right, lay it 00:07:35.800 --> 00:07:39.019 on me. Simple terms. What is that? The fundamental 00:07:39.019 --> 00:07:41.759 belief of LTCM was that markets are efficient 00:07:41.759 --> 00:07:45.699 eventually, but in the short term, they can be 00:07:45.699 --> 00:07:48.220 inefficient. So things get out of whack for a 00:07:48.220 --> 00:07:50.819 little while. Exactly. Sometimes two things that 00:07:50.819 --> 00:07:52.379 are essentially the same that should have the 00:07:52.379 --> 00:07:54.220 same price, they trade at slightly different 00:07:54.220 --> 00:07:56.420 prices. Yeah. Maybe due to a temporary supply 00:07:56.420 --> 00:07:58.600 and demand quirk or a liquidity issue in the 00:07:58.600 --> 00:08:01.060 market. A convergence trade is a bet that eventually 00:08:01.060 --> 00:08:03.120 those prices will converge. They'll snap back 00:08:03.120 --> 00:08:06.740 together. OK, can we use the old versus new bond 00:08:06.740 --> 00:08:08.639 trade? I saw that in the notes and it seems like 00:08:08.639 --> 00:08:10.740 it was their bread and butter. It absolutely 00:08:10.740 --> 00:08:12.939 was. It's the perfect example. So think about 00:08:12.939 --> 00:08:16.040 U .S. Treasury bonds. The U .S. government issues 00:08:16.040 --> 00:08:20.019 30 year bonds all the time. When a brand new 00:08:20.019 --> 00:08:22.660 30 -year bond is issued, it's called the on -the 00:08:22.660 --> 00:08:25.060 -run bond. It's a shiny new thing. It's the shiny 00:08:25.060 --> 00:08:27.339 new thing. It's incredibly liquid. Everybody 00:08:27.339 --> 00:08:29.819 wants it. Pension funds, foreign governments, 00:08:29.939 --> 00:08:33.100 other banks, they all want the newest, most actively 00:08:33.100 --> 00:08:35.659 traded issue. It's like the new iPhone. Everyone 00:08:35.659 --> 00:08:38.139 pays a little premium for the new iPhone, even 00:08:38.139 --> 00:08:41.019 if last year's model is basically the same. That's 00:08:41.019 --> 00:08:43.360 a perfect analogy. Because it's the new iPhone, 00:08:43.559 --> 00:08:45.759 people are willing to pay a tiny, tiny premium 00:08:45.759 --> 00:08:48.120 for it. Now, you look at a bond that was issued, 00:08:48.220 --> 00:08:51.039 say, three months ago. Right. It's now technically 00:08:51.039 --> 00:08:54.019 a 29 and three quarter year bond. Right. The 00:08:54.019 --> 00:08:55.980 cash flows from these two bonds, I mean, they're 00:08:55.980 --> 00:08:58.179 almost identical. The credit risk is identical. 00:08:58.299 --> 00:08:59.879 It's the U .S. government. They're going to pay 00:08:59.879 --> 00:09:01.519 you back. So it's basically the same product. 00:09:01.779 --> 00:09:05.240 It is. But that older bond, it's harder to trade. 00:09:05.519 --> 00:09:08.259 It's less liquid. It's called off the run. And 00:09:08.259 --> 00:09:10.100 because of that, it trades just a little bit 00:09:10.100 --> 00:09:12.159 cheaper. So you have two things that are worth 00:09:12.159 --> 00:09:14.179 almost exactly the same, but one is expensive 00:09:14.179 --> 00:09:16.860 and one is cheap just because of liquidity. Correct. 00:09:16.919 --> 00:09:20.679 So LTCM's computer models would scan the market 00:09:20.679 --> 00:09:23.779 and spot this tiny discrepancy and they'd make 00:09:23.779 --> 00:09:26.240 a trade. They would short the expensive new bond 00:09:26.240 --> 00:09:28.299 betting. Its price would go down relative to 00:09:28.299 --> 00:09:30.299 the other. And they would simultaneously buy 00:09:30.299 --> 00:09:32.700 the cheap old bond, betting its price would go 00:09:32.700 --> 00:09:35.240 up. And the key is... They didn't care if the 00:09:35.240 --> 00:09:37.379 overall market went up or down. Didn't matter 00:09:37.379 --> 00:09:39.740 at all. Interest rates could go to the moon or 00:09:39.740 --> 00:09:42.820 to zero. All they needed was for the gap, the 00:09:42.820 --> 00:09:45.559 spread between those two bonds to close. They 00:09:45.559 --> 00:09:47.899 just needed the prices to converge. Because eventually 00:09:47.899 --> 00:09:50.960 the new bond becomes an old bond. Exactly. The 00:09:50.960 --> 00:09:52.940 moment the Treasury holds its next auction and 00:09:52.940 --> 00:09:55.620 issues an even newer bond, the on the run becomes 00:09:55.620 --> 00:09:58.639 off the run. The premium evaporates. The spread 00:09:58.639 --> 00:10:01.480 closes. It's a mathematical certainty, or so 00:10:01.480 --> 00:10:03.259 they thought. OK, but here's the catch, right? 00:10:03.500 --> 00:10:06.460 The difference in price between those two bonds, 00:10:06.559 --> 00:10:08.940 the spread, it can't be very big. We're talking 00:10:08.940 --> 00:10:11.700 pennies, aren't we? Or like fractions of pennies. 00:10:11.820 --> 00:10:14.659 We call them basis points. And yes, it is tiny. 00:10:14.799 --> 00:10:17.299 You are literally picking up nickels. If you 00:10:17.299 --> 00:10:20.440 or I did this with $1 ,000, we might make enough 00:10:20.440 --> 00:10:23.120 money to buy a stick of gum. It's very low risk. 00:10:23.909 --> 00:10:27.250 but it's also very, very low return. That's the 00:10:27.250 --> 00:10:29.789 problem with pure arbitrage. So how on earth 00:10:29.789 --> 00:10:32.909 do you generate 40 % annual returns from a stick 00:10:32.909 --> 00:10:38.070 -of -gum strategy? One word. Leverage. Massive, 00:10:38.070 --> 00:10:40.889 mind -boggling leverage. This is the scary part 00:10:40.889 --> 00:10:42.649 of the story. This is the fuel that they poured 00:10:42.649 --> 00:10:45.230 on the fire. Since the spreads were so small, 00:10:45.389 --> 00:10:47.250 like you said, like picking up nickels, they 00:10:47.250 --> 00:10:49.129 figured they had to use a bulldozer to pick up 00:10:49.129 --> 00:10:51.690 enough of them to make it worthwhile. And a bulldozer 00:10:51.690 --> 00:10:54.090 means borrowed money. Lots and lots of borrowed 00:10:54.090 --> 00:10:55.929 money. They borrowed from every bank on Wall 00:10:55.929 --> 00:10:58.570 Street to amplify their bets. I'm looking at 00:10:58.570 --> 00:11:00.929 the numbers here for the start of 1998. They 00:11:00.929 --> 00:11:04.830 had about $4 .7 billion in their own equity, 00:11:04.990 --> 00:11:06.750 their own capital, which is a lot of money. A 00:11:06.750 --> 00:11:10.590 huge amount. But they had borrowed over $124 00:11:10.590 --> 00:11:14.450 .5 billion. It's just an insane number. That 00:11:14.450 --> 00:11:17.210 is a debt -to -equity ratio of over 25 to 1. 00:11:17.980 --> 00:11:21.620 For every single dollar of their own money in 00:11:21.620 --> 00:11:24.059 the fund, they were playing with $25 of someone 00:11:24.059 --> 00:11:25.759 else's money. And that's just the stuff on the 00:11:25.759 --> 00:11:27.200 balance sheet, right? That's what they reported. 00:11:27.360 --> 00:11:29.360 What about the derivatives? Yeah, that's the 00:11:29.360 --> 00:11:32.500 shadow exposure. The really scary part off the 00:11:32.500 --> 00:11:35.279 balance sheet, they had positions, mostly things 00:11:35.279 --> 00:11:37.799 like interest rate swaps with a notional value 00:11:37.799 --> 00:11:41.379 of one point to five trillion dollars. Trillion 00:11:41.379 --> 00:11:44.100 with a T. I mean, it's hard to even wrap your 00:11:44.100 --> 00:11:46.940 head around a number that big. It is. To put 00:11:46.940 --> 00:11:49.759 it in perspective, they took an equity base of 00:11:49.759 --> 00:11:54.549 less than. $5 billion and turned it into a market 00:11:54.549 --> 00:11:56.970 footprint that rivaled the GDP of a country like 00:11:56.970 --> 00:11:59.870 Spain or Canada. It was immense. And for a while, 00:11:59.889 --> 00:12:01.909 it worked beautifully. Until it didn't. Until 00:12:01.909 --> 00:12:04.429 it didn't. Because leverage cuts both ways. If 00:12:04.429 --> 00:12:07.470 you're leveraged 25 to 1, a tiny 4 % drop in 00:12:07.470 --> 00:12:09.710 the value of your assets wipes out 100 % of your 00:12:09.710 --> 00:12:12.169 equity. You're at zero. Gone. And that's where 00:12:12.169 --> 00:12:14.289 that famous analogy comes from, isn't it? The 00:12:14.289 --> 00:12:16.889 strategy was compared to picking up nickels in 00:12:16.889 --> 00:12:19.850 front of a bulldozer. Yes. You make small, reliable 00:12:19.850 --> 00:12:22.730 gains day after day after day. But there is a 00:12:22.730 --> 00:12:26.029 very small, non -zero chance that one day the 00:12:26.029 --> 00:12:28.070 bulldozer will move and you'll get completely 00:12:28.070 --> 00:12:30.309 flattened. But for the first three years, from 00:12:30.309 --> 00:12:33.350 94 to 97, that bulldozer was parked in the garage. 00:12:33.669 --> 00:12:35.509 They were just crushing it. They were the kings 00:12:35.509 --> 00:12:37.450 of the world. I mean, look at the early returns. 00:12:37.570 --> 00:12:42.519 Year one. 21 % after fees. Year two, 43%. Year 00:12:42.519 --> 00:12:46.139 three, 41%. Unbelievable numbers. They were the 00:12:46.139 --> 00:12:48.460 envy of all of Wall Street. They felt completely 00:12:48.460 --> 00:12:51.059 untouchable. So what changed? We get to 1997 00:12:51.059 --> 00:12:53.580 and the seeds of destruction start to get planted. 00:12:53.759 --> 00:12:55.320 Things start to look a little different. Well, 00:12:55.340 --> 00:12:58.039 the first problem was, ironically, their own 00:12:58.039 --> 00:13:00.480 success. Arbitrage opportunities are basically 00:13:00.480 --> 00:13:03.580 market inefficiencies. Once a smart person finds 00:13:03.580 --> 00:13:06.440 an inefficiency and exploits it, it tends to, 00:13:06.580 --> 00:13:09.299 well... disappear. Because other people copped 00:13:09.299 --> 00:13:13.440 to you. Exactly. LTCM was so successful, so famous, 00:13:13.620 --> 00:13:16.960 that copycats started showing up. Other firms, 00:13:17.039 --> 00:13:19.100 other banks, they all said, hey, we can do that 00:13:19.100 --> 00:13:21.600 on the run, off the run bond trade too. So suddenly 00:13:21.600 --> 00:13:23.240 there were too many people trying to pick up 00:13:23.240 --> 00:13:25.600 the same nickels. That's exactly. And when that 00:13:25.600 --> 00:13:27.419 happens, the spreads get tighter and tighter. 00:13:27.559 --> 00:13:29.840 The opportunities get smaller. And you can see 00:13:29.840 --> 00:13:33.039 it in their numbers. In 1997, their returns dropped 00:13:33.039 --> 00:13:36.340 to 17 percent. Which for you or me, 17 percent 00:13:36.340 --> 00:13:38.799 is a fantastic year. We'd be thrilled. We'd be 00:13:38.799 --> 00:13:42.019 retiring. But for the geniuses in Greenwich charging 00:13:42.019 --> 00:13:44.179 these massive fees. It was an embarrassment. 00:13:44.580 --> 00:13:47.679 It was a failure. It barely beat the S &P 500 00:13:47.679 --> 00:13:50.139 stock market index that year. They were under. 00:13:50.379 --> 00:13:52.860 immense pressure, a lot of it internal, to keep 00:13:52.860 --> 00:13:55.740 the magic alive, to keep posting those 40 % numbers. 00:13:56.000 --> 00:13:58.360 And this leads to something called style drift. 00:13:58.639 --> 00:14:01.059 I love this term because it sounds so harmless, 00:14:01.059 --> 00:14:04.340 so academic, but it's actually incredibly dangerous. 00:14:04.720 --> 00:14:06.980 It's a beginning of the end for so many great 00:14:06.980 --> 00:14:10.259 investors. Style drift is what happens when you 00:14:10.259 --> 00:14:12.980 move away from your core competency, your area 00:14:12.980 --> 00:14:15.299 of expertise, because the returns just aren't 00:14:15.299 --> 00:14:17.799 there anymore. You start betting on things you 00:14:17.799 --> 00:14:20.220 don't understand quite as well. LTCM was the 00:14:20.220 --> 00:14:23.320 king of fixed income arbitrage. They knew U .S. 00:14:23.320 --> 00:14:25.639 bonds, European bonds, Japanese bonds, inside 00:14:25.639 --> 00:14:28.519 and out. But suddenly they're looking at other 00:14:28.519 --> 00:14:30.940 stuff. That's right. They started moving into 00:14:30.940 --> 00:14:34.809 equity volatility, merger arbitrage, and... this 00:14:34.809 --> 00:14:37.629 is the crucial part, emerging market debt. They 00:14:37.629 --> 00:14:39.809 were buying Russian bonds, Brazilian bonds, pie 00:14:39.809 --> 00:14:42.169 bonds. And this was a source of internal debate. 00:14:43.069 --> 00:14:45.210 Myron Scholes himself, the Nobel winner, actually 00:14:45.210 --> 00:14:47.710 warned the other partners about this. He said, 00:14:47.830 --> 00:14:50.090 and this is a quote, that they had no informational 00:14:50.090 --> 00:14:52.470 advantage in trading things like the Norwegian 00:14:52.470 --> 00:14:55.169 krona. But they did it anyway. They did. They 00:14:55.169 --> 00:14:57.190 needed the yield. They felt their models, their 00:14:57.190 --> 00:14:59.549 mathematical approach could be applied to anything. 00:14:59.970 --> 00:15:02.870 If the math works for U .S. treasuries, why wouldn't 00:15:02.870 --> 00:15:05.309 it work for a currency in Norway? It shows a 00:15:05.309 --> 00:15:08.409 certain level of hubris, a belief that math trumps 00:15:08.409 --> 00:15:12.059 local knowledge. It's pure hubris. And there's 00:15:12.059 --> 00:15:14.480 another story from this period, this UBS tax 00:15:14.480 --> 00:15:17.080 scheme, that I think really shows their mindset. 00:15:17.279 --> 00:15:19.220 It's a clear sign that they were getting too 00:15:19.220 --> 00:15:21.659 cute for their own good. What was that about? 00:15:21.980 --> 00:15:24.200 It was financial engineering at its absolute 00:15:24.200 --> 00:15:26.940 peak. The partners wanted to treat their income 00:15:26.940 --> 00:15:29.879 as long term capital gains to pay a lower tax 00:15:29.879 --> 00:15:32.720 rate, about 20 percent versus almost 40 percent 00:15:32.720 --> 00:15:35.980 for ordinary income. So they set up this incredibly 00:15:35.980 --> 00:15:38.740 complex call option structure with the Swiss 00:15:38.740 --> 00:15:41.159 Bank UBS. I'm not even going to pretend to understand 00:15:41.159 --> 00:15:43.960 the details. The details are a nightmare. But 00:15:43.960 --> 00:15:46.399 the takeaway is that they were using their immense 00:15:46.399 --> 00:15:49.320 brainpower not to find value in the market, but 00:15:49.320 --> 00:15:51.879 to game the tax code. It just showed a belief 00:15:51.879 --> 00:15:53.980 that they could outsmart any system, whether 00:15:53.980 --> 00:15:56.470 it was the market. Or the IRS. And it was a distraction. 00:15:56.730 --> 00:15:59.429 A huge one. And it brought UBS way too close 00:15:59.429 --> 00:16:01.149 to them, which would come back to haunt the bank 00:16:01.149 --> 00:16:03.450 later. Before we get to the final crash, we have 00:16:03.450 --> 00:16:05.649 to talk about one specific trade that went wrong. 00:16:05.750 --> 00:16:08.690 The Royal Dutch Shell trade. This one feels like 00:16:08.690 --> 00:16:11.269 the perfect microcosm of their entire model failing. 00:16:11.509 --> 00:16:14.629 It is the classic example. So Royal Dutch and 00:16:14.629 --> 00:16:17.889 Shell were at the time a dual listed company. 00:16:18.360 --> 00:16:20.440 Basically two different stocks, but for the same 00:16:20.440 --> 00:16:23.179 underlying company. The economic value of the 00:16:23.179 --> 00:16:25.539 whole enterprise was split 60 -40 between them. 00:16:25.700 --> 00:16:28.139 So theoretically, the two stocks should move 00:16:28.139 --> 00:16:31.120 in perfect lockstep. They should, right. If the 00:16:31.120 --> 00:16:33.399 company makes more money, both stocks should 00:16:33.399 --> 00:16:35.980 go up by roughly the same percentage. But for 00:16:35.980 --> 00:16:38.240 historical reasons, they didn't. They traded 00:16:38.240 --> 00:16:40.840 at different prices. Royal Dutch consistently 00:16:40.840 --> 00:16:44.240 traded at a premium to show. And LTCM looked 00:16:44.240 --> 00:16:46.860 at this and their model screamed inefficiency. 00:16:47.200 --> 00:16:49.960 Their model screamed, this is wrong. Matt says 00:16:49.960 --> 00:16:52.360 these prices should converge. So they put on 00:16:52.360 --> 00:16:55.259 a massive trade. They bet $2 .3 billion that 00:16:55.259 --> 00:16:57.379 the spread would close. They shorted the expensive 00:16:57.379 --> 00:16:59.799 Royal Dutch stock and went long the cheaper shell 00:16:59.799 --> 00:17:02.100 stock. A classic convergence trade. What happened? 00:17:02.279 --> 00:17:04.480 The market decided it didn't care about their 00:17:04.480 --> 00:17:07.400 math. The premium didn't shrink. It widened. 00:17:08.000 --> 00:17:10.599 Dramatically. It went from about 8 or 10 percent 00:17:10.599 --> 00:17:13.440 to over 22 percent. So the trade went completely 00:17:13.440 --> 00:17:16.740 against them. Completely. And they lost hundreds 00:17:16.740 --> 00:17:18.859 of millions of dollars on that one trade alone. 00:17:19.079 --> 00:17:21.619 It was a brutal lesson. It really is the old 00:17:21.619 --> 00:17:24.480 John Maynard Keynes saying, isn't it? The market 00:17:24.480 --> 00:17:26.740 can remain irrational longer than you can remain 00:17:26.740 --> 00:17:29.779 solvent. Precisely. Their model said, this must 00:17:29.779 --> 00:17:32.880 happen. And reality said, yeah, maybe, but not 00:17:32.880 --> 00:17:35.940 today. And not before you go broke. Which brings 00:17:35.940 --> 00:17:38.740 us to the really big question, the blind spot. 00:17:39.000 --> 00:17:41.519 We talked about the geniuses, the Nobel winners, 00:17:41.720 --> 00:17:45.539 the supercomputers. How did they miss this? How 00:17:45.539 --> 00:17:48.259 are their models so fundamentally wrong? It really 00:17:48.259 --> 00:17:51.960 comes down to one thing, data, or more specifically, 00:17:52.220 --> 00:17:54.819 a lack of it. Their primary risk model was something 00:17:54.819 --> 00:17:57.980 called value at risk or VAR. VAR? I've heard 00:17:57.980 --> 00:18:00.950 of this. Most banks use some version of it. It's 00:18:00.950 --> 00:18:02.849 a model that tries to calculate the maximum amount 00:18:02.849 --> 00:18:05.349 of money you might lose on any given day based 00:18:05.349 --> 00:18:07.509 on historical patterns. Okay, based on historical 00:18:07.509 --> 00:18:09.690 patterns, that sounds reasonable. It is reasonable 00:18:09.690 --> 00:18:12.490 if, and this is a very big if if you use enough 00:18:12.490 --> 00:18:16.029 history, the fatal flaw, the blind spot for LTCM, 00:18:16.089 --> 00:18:18.450 was that their models were largely built on data 00:18:18.450 --> 00:18:20.849 from the previous five years. Five years, that's 00:18:20.849 --> 00:18:24.599 it, from what, 93 to 98? Basically, yes. And 00:18:24.599 --> 00:18:26.839 what was happening in the early to mid -90s, 00:18:26.839 --> 00:18:30.440 it was a generally calm, steadily rising bull 00:18:30.440 --> 00:18:32.779 market period. Right. Not a lot of volatility. 00:18:33.039 --> 00:18:35.880 Very little. By using only five years of data. 00:18:36.279 --> 00:18:39.900 They completely, totally excluded the 1987 stock 00:18:39.900 --> 00:18:42.799 market crash. They excluded the bond market crisis 00:18:42.799 --> 00:18:46.799 of 94. They excluded the 1917 Russian default, 00:18:47.019 --> 00:18:48.440 which might have been a useful thing to have 00:18:48.440 --> 00:18:50.880 in their models. So their models basically assumed 00:18:50.880 --> 00:18:53.500 that a major global crash was impossible because 00:18:53.500 --> 00:18:55.640 it hadn't happened recently. Their models assumed 00:18:55.640 --> 00:18:58.960 it was a statistical impossibility. The historian 00:18:58.960 --> 00:19:01.200 Neil Ferguson had this absolutely brutal quote 00:19:01.200 --> 00:19:03.619 about it. He said, the Nobel Prize winners had 00:19:03.619 --> 00:19:06.079 known plenty of mathematics. but not enough history. 00:19:06.279 --> 00:19:09.059 Ouch. That's a perfect summary. It is. Their 00:19:09.059 --> 00:19:11.019 models predicted that the kind of daily loss 00:19:11.019 --> 00:19:13.519 they eventually suffered in August 98 was what's 00:19:13.519 --> 00:19:16.059 called a 10 -sigma event. It was an event that 00:19:16.059 --> 00:19:18.059 should happen only once in the entire life of 00:19:18.059 --> 00:19:19.900 the universe. And yet it happened on a Tuesday. 00:19:20.240 --> 00:19:22.779 Exactly. Because in the real world, financial 00:19:22.779 --> 00:19:25.660 markets don't follow a perfect bell curve. Real 00:19:25.660 --> 00:19:28.140 life has fat tails. Okay, fat tails. Explain 00:19:28.140 --> 00:19:31.259 that. So think of a normal bell curve. In a normal 00:19:31.259 --> 00:19:35.029 distribution, the tails... The extreme events 00:19:35.029 --> 00:19:37.890 way out on the left and right are very, very 00:19:37.890 --> 00:19:41.029 thin. They almost never happen. In financial 00:19:41.029 --> 00:19:43.849 markets, the tails are fat. Crashes, panics, 00:19:43.930 --> 00:19:47.150 defaults, pandemics. These extreme universe -ending 00:19:47.150 --> 00:19:49.730 events happen way more often than a standard 00:19:49.730 --> 00:19:52.390 statistical model predicts. And LTCM bet their 00:19:52.390 --> 00:19:55.869 entire firm and $120 billion of borrowed money 00:19:55.869 --> 00:19:58.769 that the tails were thin. They bet the farm on 00:19:58.769 --> 00:20:02.109 it. And then in August 1998, they found out just 00:20:02.109 --> 00:20:05.089 how fat those tails could be. The perfect storm 00:20:05.089 --> 00:20:07.390 arrived. And it started in Russia. It started 00:20:07.390 --> 00:20:10.329 in Russia. On August 17th, 1998, the Russian 00:20:10.329 --> 00:20:13.069 government defaults on its domestic debt. Now, 00:20:13.069 --> 00:20:15.569 hold on. Why was this such a massive surprise? 00:20:15.670 --> 00:20:17.690 I mean, governments go broke sometimes, right? 00:20:17.849 --> 00:20:20.890 They do. But they usually default on debt issued 00:20:20.890 --> 00:20:23.690 in a foreign currency, like U .S. dollars. The 00:20:23.690 --> 00:20:25.470 conventional wisdom at the time was that a sovereign 00:20:25.470 --> 00:20:27.630 nation would never default on debt issued in 00:20:27.630 --> 00:20:29.430 its own currency. Because they can just print 00:20:29.430 --> 00:20:31.670 more money to pay it off. Right. It would cause 00:20:31.670 --> 00:20:33.829 hyperinflation, sure, but they wouldn't technically 00:20:33.829 --> 00:20:36.630 default. Russia proved that entire school of 00:20:36.630 --> 00:20:39.329 thought wrong. They just said, nope. We aren't 00:20:39.329 --> 00:20:41.829 paying. And that one event sent shockwaves through 00:20:41.829 --> 00:20:43.730 the entire global system. It triggered a global 00:20:43.730 --> 00:20:46.910 panic. I mean, a real visceral panic. Investors 00:20:46.910 --> 00:20:48.769 everywhere woke up and collectively realized 00:20:48.769 --> 00:20:51.130 the world was a much more dangerous place than 00:20:51.130 --> 00:20:53.190 they thought. And when investors get scared, 00:20:53.349 --> 00:20:56.089 they all do the exact same thing. Flight to quality. 00:20:56.289 --> 00:20:58.589 Flight to quality. They sell anything that looks 00:20:58.589 --> 00:21:01.410 even remotely risky. Emerging market bonds, junk 00:21:01.410 --> 00:21:03.970 bonds, corporate bonds. Anything with hair on 00:21:03.970 --> 00:21:07.529 it. They sold it all. And they poured all of 00:21:07.529 --> 00:21:10.269 that money into the safest, most liquid asset 00:21:10.269 --> 00:21:13.670 on planet Earth. Brand new. On the run U .S. 00:21:13.670 --> 00:21:15.369 Treasury bonds. Wait a minute. That sounds familiar. 00:21:15.710 --> 00:21:18.690 That's the exact asset LTCM was shorting as the 00:21:18.690 --> 00:21:21.089 expensive side of their trade. You nailed it. 00:21:21.849 --> 00:21:24.450 LTCM was fundamentally betting that the spread 00:21:24.450 --> 00:21:27.049 between the risky stuff and the safe stuff would 00:21:27.049 --> 00:21:30.019 shrink. Instead, because of this global panic, 00:21:30.200 --> 00:21:33.019 that spread blew out to historic widths. The 00:21:33.019 --> 00:21:35.920 safe bonds they were short got massively, insanely 00:21:35.920 --> 00:21:38.519 expensive because everyone in the world wanted 00:21:38.519 --> 00:21:41.019 them. And the illiquid risky bonds they owned 00:21:41.019 --> 00:21:43.319 got dirt cheap because no one wanted to touch 00:21:43.319 --> 00:21:46.079 them. So every single one of their core convergence 00:21:46.079 --> 00:21:48.319 trades went the wrong way at the exact same time. 00:21:48.400 --> 00:21:50.700 Every single one. The fundamental assumption 00:21:50.700 --> 00:21:52.299 of their models was that all these different 00:21:52.299 --> 00:21:54.799 trades were uncorrelated. But in a crisis, that's 00:21:54.799 --> 00:21:57.500 not true. In a crisis, the only thing that rises 00:21:57.500 --> 00:22:00.099 is correlation. Everything you own goes down 00:22:00.099 --> 00:22:02.019 and everything you're short goes up. And to make 00:22:02.019 --> 00:22:04.660 matters worse, Salomon Brothers, Merriweather's 00:22:04.660 --> 00:22:07.539 old firm, had just decided to exit the arbitrage 00:22:07.539 --> 00:22:10.799 business right before this in July 98. That was 00:22:10.799 --> 00:22:12.759 just the worst possible luck. It was the nail 00:22:12.759 --> 00:22:15.680 in the coffin. Salomon had a portfolio that looked 00:22:15.680 --> 00:22:18.680 almost exactly like LTCM's. When Salomon started 00:22:18.680 --> 00:22:20.980 to liquidate their positions, they were selling 00:22:20.980 --> 00:22:23.700 all the same things LTCM owned, which pushed 00:22:23.700 --> 00:22:26.599 the prices down even further. faster it was like 00:22:26.599 --> 00:22:28.519 a run on the bank but the bank was the entire 00:22:28.519 --> 00:22:31.220 global bond market it was there was no liquidity 00:22:31.220 --> 00:22:34.720 no bids they were trapped so let's look at the 00:22:34.720 --> 00:22:37.700 damage the numbers are just breathtaking august 00:22:37.700 --> 00:22:42.019 of 1998 their losses were 1 .85 billion dollars 00:22:42.019 --> 00:22:45.750 in one month and it gets worse September starts 00:22:45.750 --> 00:22:48.869 and the equity in the fund just evaporates. It 00:22:48.869 --> 00:22:51.509 drops from $2 .3 billion at the start of the 00:22:51.509 --> 00:22:54.690 month to just $400 million by the third week. 00:22:54.869 --> 00:22:57.349 And remember the leverage. As their equity base 00:22:57.349 --> 00:22:59.809 shrank to almost nothing, their leverage ratio 00:22:59.809 --> 00:23:03.289 just skyrocketed. It went from 25 to 1 to over 00:23:03.289 --> 00:23:06.529 250 to 1. 250 to 1. That's not investing. That's 00:23:06.529 --> 00:23:09.130 not even gambling. That is just pure, desperate 00:23:09.130 --> 00:23:12.089 prayer. And by this point, the sharks were circling. 00:23:12.130 --> 00:23:14.109 The rest of Wall Street could smell blood in 00:23:14.109 --> 00:23:17.319 the water. Other traders knew LTCM was in deep 00:23:17.319 --> 00:23:20.099 trouble. They knew LTCM had to sell their positions 00:23:20.099 --> 00:23:23.039 to meet margin calls. So what do you do? You 00:23:23.039 --> 00:23:24.960 front run them. You sell the same things before 00:23:24.960 --> 00:23:27.240 they do, driving the prices down even more, forcing 00:23:27.240 --> 00:23:29.319 them to take even bigger losses when they finally 00:23:29.319 --> 00:23:33.140 have to sell. It's a brutal, vicious death spiral. 00:23:33.380 --> 00:23:35.019 Yeah. And they were caught right in the middle 00:23:35.019 --> 00:23:36.359 of it. So this is it. The fund is collapsing. 00:23:36.759 --> 00:23:39.180 And now we get to the bailout. And this is the 00:23:39.180 --> 00:23:40.819 part of the story that really highlights how 00:23:40.819 --> 00:23:43.559 dangerous this whole situation had become. It 00:23:43.559 --> 00:23:45.240 wasn't just that some rich guys in Connecticut 00:23:45.240 --> 00:23:47.779 were losing money. The Federal Reserve got involved. 00:23:48.079 --> 00:23:51.000 Why? Systemic risk. That's the phrase you'll 00:23:51.000 --> 00:23:54.700 hear over and over. LTCM was so deeply interconnected 00:23:54.700 --> 00:23:57.640 with every major investment bank on Wall Street. 00:23:57.779 --> 00:24:00.240 They had those trillions of dollars in derivative 00:24:00.240 --> 00:24:03.599 contracts. If LTCM just collapsed overnight and 00:24:03.599 --> 00:24:05.619 defaulted on all those contracts, they would 00:24:05.619 --> 00:24:08.180 have left massive. gaping holes in the balance 00:24:08.180 --> 00:24:10.519 sheets of Merrill Lynch, Goldman Sachs, UBS, 00:24:10.859 --> 00:24:13.339 JP Morgan, everyone. It would have been a domino 00:24:13.339 --> 00:24:15.359 effect. One falls, they all fall. The Fed was 00:24:15.359 --> 00:24:17.259 terrified it would freeze the global financial 00:24:17.259 --> 00:24:20.059 system. That credit would just stop flowing between 00:24:20.059 --> 00:24:22.880 banks, that markets would crash. It was a genuine 00:24:22.880 --> 00:24:26.319 fear of a 1929 style event. So on September 23rd, 00:24:26.319 --> 00:24:29.460 1998, the New York Fed calls a meeting. And it 00:24:29.460 --> 00:24:32.329 wasn't just a meeting, it was a summons. They 00:24:32.329 --> 00:24:34.990 basically locked the heads of all the major banks 00:24:34.990 --> 00:24:37.089 in a conference room at the Federal Reserve Bank 00:24:37.089 --> 00:24:40.589 of New York and said, we have a problem. But 00:24:40.589 --> 00:24:43.109 before that final meeting, there was this incredible 00:24:43.109 --> 00:24:46.750 moment where Warren Buffett almost bought the 00:24:46.750 --> 00:24:49.390 whole thing, right? Yes. This is one of the great 00:24:49.390 --> 00:24:52.849 what -ifs in financial history. Warren Buffett, 00:24:52.869 --> 00:24:55.750 along with Goldman Sachs and AIG, made an offer 00:24:55.750 --> 00:24:57.789 to buy out the partners and take over the fund. 00:24:58.190 --> 00:25:01.809 They offered $250 million for the partners' stake. 00:25:02.089 --> 00:25:04.529 $250 million? They had started the year with 00:25:04.529 --> 00:25:06.730 nearly $2 billion of their own money in there. 00:25:06.869 --> 00:25:09.069 It was pennies on the dollar. It was a total 00:25:09.069 --> 00:25:11.750 humiliation. But it was also the only offer on 00:25:11.750 --> 00:25:14.309 the table. And Buffett, being Buffett, put a 00:25:14.309 --> 00:25:16.670 very specific condition on it. He gave them less 00:25:16.670 --> 00:25:19.819 than one hour to decide. One hour. to decide 00:25:19.819 --> 00:25:22.099 the fate of their entire firm. He didn't want 00:25:22.099 --> 00:25:24.599 to get into a negotiation. He faxed them the 00:25:24.599 --> 00:25:27.480 offer and said, you have until this time to say 00:25:27.480 --> 00:25:30.400 yes or no. But apparently the structure of the 00:25:30.400 --> 00:25:33.059 offer was complicated, or Merriweather couldn't 00:25:33.059 --> 00:25:36.359 get the legal sign off in time. Whatever the 00:25:36.359 --> 00:25:39.839 reason, the hour lapsed. The deal died. So they 00:25:39.839 --> 00:25:41.759 were left with the Fed, with the government. 00:25:42.140 --> 00:25:44.440 Well, with the Fed playing chaperone. This is 00:25:44.440 --> 00:25:46.619 an important distinction. The Fed didn't want 00:25:46.619 --> 00:25:49.200 to use public money. This wasn't a taxpayer bailout 00:25:49.200 --> 00:25:52.220 like we saw in 2008. The Fed's role was to act 00:25:52.220 --> 00:25:55.680 as an organizer or maybe an enforcer. They basically 00:25:55.680 --> 00:25:58.099 told the banks, you guys created this monster 00:25:58.099 --> 00:26:00.920 by lending him all this money. You enabled them. 00:26:00.980 --> 00:26:03.180 So you guys need to fix it. If you don't all 00:26:03.180 --> 00:26:05.500 chip in to save this fund, you are all going 00:26:05.500 --> 00:26:07.779 down with it. That is some serious arm twisting 00:26:07.779 --> 00:26:10.650 from the central bank. It worked. 14 banks agreed 00:26:10.650 --> 00:26:14.329 to inject $3 .6 to $5 billion of new capital 00:26:14.329 --> 00:26:16.630 into the fund to keep it afloat. In exchange, 00:26:16.910 --> 00:26:20.250 they effectively took 90 % ownership of LTCM. 00:26:20.430 --> 00:26:22.809 But not everyone played ball, did they? There 00:26:22.809 --> 00:26:25.190 was a famous holdout. Bear Stearns. Bear Stearns 00:26:25.190 --> 00:26:27.349 declined to participate. They did. They were 00:26:27.349 --> 00:26:29.829 the clearing agent for LTCM, which meant they 00:26:29.829 --> 00:26:32.349 had a lot of exposure already. They refused to 00:26:32.349 --> 00:26:34.470 put any new money into the pot. They essentially 00:26:34.470 --> 00:26:38.440 said, we've got enough risk, thanks. Wow. That 00:26:38.440 --> 00:26:41.819 is just rich with irony, considering what happened 00:26:41.819 --> 00:26:44.059 to Bear Stearns almost exactly 10 years later 00:26:44.059 --> 00:26:46.339 in 2008. It is. A lot of people on Wall Street 00:26:46.339 --> 00:26:48.319 remembered that moment when Bear Stearns came 00:26:48.319 --> 00:26:51.500 asking for help in 2008. Let's just say they 00:26:51.500 --> 00:26:53.099 didn't have a lot of friends left in the room. 00:26:53.240 --> 00:26:55.599 So the banks put the money in. They stabilized 00:26:55.599 --> 00:26:58.799 the fund. What happened to the partners? The 00:26:58.799 --> 00:27:02.700 dream team. Meriwether, Scholes, Merton. They 00:27:02.700 --> 00:27:06.019 were completely wiped out. They had $1 .9 billion 00:27:06.019 --> 00:27:08.960 of their own personal money invested in the fund. 00:27:09.019 --> 00:27:11.500 All gone. All gone. They were left with a tiny 00:27:11.500 --> 00:27:13.819 sliver of equity that was basically worthless, 00:27:13.960 --> 00:27:16.559 consumed by the debt. They lost their personal 00:27:16.559 --> 00:27:18.619 fortunes. They lost their reputations. It was 00:27:18.619 --> 00:27:20.740 a total and complete financial ruin for them. 00:27:20.880 --> 00:27:23.119 But the banks, the ones who put up the bailout 00:27:23.119 --> 00:27:25.539 money, they eventually got paid back. They did. 00:27:25.619 --> 00:27:28.940 And then some. Once the immediate panic subsided 00:27:28.940 --> 00:27:32.160 and that flight to quality reversed, those convergence 00:27:32.160 --> 00:27:34.599 trades actually started to work again. The spreads 00:27:34.599 --> 00:27:36.880 started to close. The liquidation of the fund 00:27:36.880 --> 00:27:39.660 was orderly. By early 2000, the fund was completely 00:27:39.660 --> 00:27:41.859 dissolved and the banks got all their money back, 00:27:42.000 --> 00:27:44.960 plus a small profit. So in the end, disaster 00:27:44.960 --> 00:27:48.480 averted. The system was saved. In the short term, 00:27:48.559 --> 00:27:51.539 yes, the immediate fire was put out. But in the 00:27:51.539 --> 00:27:54.819 long term, no. Maybe not. The solution might 00:27:54.819 --> 00:27:57.019 have planted the seeds for an even bigger crisis 00:27:57.019 --> 00:27:59.720 down the road. Okay, this brings us to the analysis, 00:28:00.019 --> 00:28:03.119 the aftermath. What did we learn from this? Or 00:28:03.119 --> 00:28:05.559 maybe, what did we fail to learn? I think the 00:28:05.559 --> 00:28:07.339 single biggest concept that comes out of this 00:28:07.339 --> 00:28:10.140 is moral hazard. Explain that for us. Moral hazard 00:28:10.140 --> 00:28:12.380 is the idea that if you protect people from the 00:28:12.380 --> 00:28:14.599 consequences of their risks, they will naturally 00:28:14.599 --> 00:28:16.920 take even bigger and crazier risks in the future. 00:28:17.160 --> 00:28:19.359 By orchestrating this private sector rescue, 00:28:19.799 --> 00:28:21.920 The Fed, and specifically the chairman at the 00:28:21.920 --> 00:28:24.420 time, Alan Greenspan, signaled to the market 00:28:24.420 --> 00:28:26.619 that if a financial firm is too big to fail, 00:28:26.880 --> 00:28:29.740 the Fed will always step in and organize a lifeboat. 00:28:29.900 --> 00:28:32.359 This is the famous Greenspan put, right? Exactly. 00:28:32.740 --> 00:28:35.059 It's the belief that there's a safety net underneath 00:28:35.059 --> 00:28:37.400 the market provided by the central bank. And 00:28:37.400 --> 00:28:40.579 it arguably set the stage for the massive excessive 00:28:40.579 --> 00:28:43.319 risk -taking we saw in the housing market that 00:28:43.319 --> 00:28:46.490 led to the 2008 crisis. Banks thought we can 00:28:46.490 --> 00:28:48.750 lend a risky hedge funds. We could load up on 00:28:48.750 --> 00:28:50.930 subprime mortgages because if it all blows up, 00:28:51.029 --> 00:28:54.250 the Fed will save us. So the rescue itself created 00:28:54.250 --> 00:28:57.349 a bigger problem. Many people argue it did. And 00:28:57.349 --> 00:28:59.869 what about the players? Surely Merriweather and 00:28:59.869 --> 00:29:02.369 the Nobel winners retired to a quiet life of 00:29:02.369 --> 00:29:04.309 teaching and writing their memoirs. You would 00:29:04.309 --> 00:29:06.730 absolutely think so. You would think after a 00:29:06.730 --> 00:29:09.490 humiliation of that magnitude, you'd never show 00:29:09.490 --> 00:29:12.589 your face on Wall Street again. But no, John 00:29:12.589 --> 00:29:15.269 Merriweather launched a new fund. JWM Partners 00:29:15.269 --> 00:29:19.089 in 1999. You are kidding me. Less than a year 00:29:19.089 --> 00:29:21.569 after the collapse. Not even a year. And guess 00:29:21.569 --> 00:29:23.849 what strategy he used in his new fund? Please 00:29:23.849 --> 00:29:25.509 don't tell me it was the same one. It was the 00:29:25.509 --> 00:29:28.170 exact same strategy. Fixed income arbitrage, 00:29:28.250 --> 00:29:30.529 just with slightly less leverage. And how did 00:29:30.529 --> 00:29:32.430 that one go? It actually worked pretty well for 00:29:32.430 --> 00:29:35.049 a while. Yeah. Until the 2008 financial crisis 00:29:35.049 --> 00:29:38.690 hit. And then JWM Partners. lost 44 % of its 00:29:38.690 --> 00:29:41.210 value and had to shut down. It's unbelievable. 00:29:41.569 --> 00:29:43.170 It's like they were so convinced their model 00:29:43.170 --> 00:29:45.710 was right that 98 was just a fluke, a one -time 00:29:45.710 --> 00:29:48.210 thing. It suggests a real addiction to the model, 00:29:48.309 --> 00:29:51.390 a core belief that this time is different or 00:29:51.390 --> 00:29:54.369 we've fixed the bug. They just couldn't let it 00:29:54.369 --> 00:29:57.190 go. And what about the banks involved? UBS took 00:29:57.190 --> 00:30:00.250 a huge hit, right, from that tax scheme. They 00:30:00.250 --> 00:30:02.549 did. The chairman of UBS was forced to resign. 00:30:03.039 --> 00:30:06.339 The bank lost nearly $800 million, mostly related 00:30:06.339 --> 00:30:08.700 to those clever tax schemes and the put options 00:30:08.700 --> 00:30:11.319 they'd written for LTCM. It was a career ender 00:30:11.319 --> 00:30:13.700 for a lot of very senior people. Let's just tally 00:30:13.700 --> 00:30:16.039 the final bill on the fund itself. The total 00:30:16.039 --> 00:30:20.079 loss was $4 .6 billion. How did that break down? 00:30:20.299 --> 00:30:22.640 The sources show about $1 .6 billion was lost 00:30:22.640 --> 00:30:25.400 in swaps, $1 .3 billion in equity volatility 00:30:25.400 --> 00:30:28.380 trades. You know, what's really interesting is 00:30:28.380 --> 00:30:30.920 that only 430 million was actually lost in Russia 00:30:30.920 --> 00:30:33.779 directly. Wow. So the trigger event, the Russian 00:30:33.779 --> 00:30:36.359 default, was actually a pretty small part of 00:30:36.359 --> 00:30:39.700 the total loss. A tiny part. It wasn't the default 00:30:39.700 --> 00:30:42.599 that killed them. It was the ripple effect. The 00:30:42.599 --> 00:30:45.920 global panic, the flight to quality, and most 00:30:45.920 --> 00:30:49.000 importantly, the liquidity drying up. That's 00:30:49.000 --> 00:30:50.660 what killed them. You know, there's a saying 00:30:50.660 --> 00:30:53.099 that liquidity is like oxygen. That's perfect. 00:30:53.240 --> 00:30:55.119 You don't notice it at all when it's there. You 00:30:55.119 --> 00:30:57.099 take it for granted. But the second it's gone, 00:30:57.240 --> 00:30:58.940 it's the only thing in the world that matters. 00:30:59.140 --> 00:31:03.460 And LTCM suffocated. So we've unpacked the rise, 00:31:03.559 --> 00:31:06.440 the strategy, the spectacular crash, and the 00:31:06.440 --> 00:31:09.240 controversial bailout. When you step back and 00:31:09.240 --> 00:31:11.519 look at the whole story, what is the big takeaway 00:31:11.519 --> 00:31:13.740 for you? If you had to summarize the lesson of 00:31:13.740 --> 00:31:16.819 LTCM in just one thought. For me, it comes back 00:31:16.819 --> 00:31:19.019 to what I said at the start. It's the profound 00:31:19.019 --> 00:31:21.960 difference between intelligence and wisdom. These 00:31:21.960 --> 00:31:23.900 men were, without question, some of the most 00:31:23.900 --> 00:31:26.779 intelligent people ever to work in finance. They 00:31:26.779 --> 00:31:28.900 could calculate the price of a complex derivative 00:31:28.900 --> 00:31:32.259 to the tenth decimal place in their heads. But 00:31:32.259 --> 00:31:34.920 they lacked the wisdom to understand that the 00:31:34.920 --> 00:31:37.940 real world doesn't always follow the clean, elegant 00:31:37.940 --> 00:31:40.839 rules of mathematics. They mistook their map 00:31:40.839 --> 00:31:43.240 for the actual territory. I love that phrase. 00:31:43.460 --> 00:31:46.279 The map is not the territory. Just because your 00:31:46.279 --> 00:31:48.680 model, your map says a 10 -sigma crash is impossible 00:31:48.680 --> 00:31:51.420 doesn't mean the market can't crash. And more 00:31:51.420 --> 00:31:54.200 specifically, I think it's the ultimate cautionary 00:31:54.200 --> 00:31:57.859 tale about the danger of leverage. Leverage doesn't 00:31:57.859 --> 00:32:00.559 make you a better investor. It just magnifies 00:32:00.559 --> 00:32:04.220 your outcome. If you're right, you win big. If 00:32:04.220 --> 00:32:06.700 you're wrong, you get wiped out. They were smart 00:32:06.700 --> 00:32:09.259 enough to be right 99 % of the time. But leverage 00:32:09.259 --> 00:32:11.319 meant that the 1 % of the time they were wrong, 00:32:11.400 --> 00:32:13.839 it was fatal. And history. You have to look at 00:32:13.839 --> 00:32:16.359 more than just the last five years of data. Always. 00:32:16.559 --> 00:32:19.779 Yeah. History is a long, messy, complicated book. 00:32:19.960 --> 00:32:22.059 And if you only read the last chapter, you are 00:32:22.059 --> 00:32:24.240 absolutely going to be surprised by the plot 00:32:24.240 --> 00:32:26.589 twist. So before we wrap this up, I want to leave 00:32:26.589 --> 00:32:28.829 our listeners with one final provocative thought 00:32:28.829 --> 00:32:31.049 to chew on. We mentioned that Meriwether started 00:32:31.049 --> 00:32:34.349 a second fund that failed in 2008. But the source 00:32:34.349 --> 00:32:36.230 material notes that he launched a third fund 00:32:36.230 --> 00:32:39.569 in 2010. JM Advisors Management. Yeah. Unbelievable. 00:32:39.569 --> 00:32:41.930 Like using the same basic strategy all over again. 00:32:42.089 --> 00:32:44.480 He just can't quit it. It raises a really fascinating 00:32:44.480 --> 00:32:47.279 question, I think. Is the allure of this kind 00:32:47.279 --> 00:32:50.880 of easy, mathematical, scientific money so powerful 00:32:50.880 --> 00:32:55.900 that even a $4 .6 billion failure, a global public 00:32:55.900 --> 00:32:59.019 humiliation, can't break the addiction? And then 00:32:59.019 --> 00:33:01.039 you have to think about today's market. Right. 00:33:01.079 --> 00:33:03.920 It is completely dominated by algorithmic trading, 00:33:04.180 --> 00:33:08.319 high frequency trading, AI models making decisions. 00:33:08.779 --> 00:33:11.099 We have faster computers and way more data than 00:33:11.099 --> 00:33:14.099 they ever did. You do. But do we have more wisdom 00:33:14.099 --> 00:33:16.720 or have we just built a bigger, faster, more 00:33:16.720 --> 00:33:19.619 powerful bulldozer? If LTCM could almost break 00:33:19.619 --> 00:33:22.660 the world in 1998 with, you know, Excel spreadsheets 00:33:22.660 --> 00:33:24.900 and telephones, what happens when the algorithms 00:33:24.900 --> 00:33:27.200 that run our entire market today all decide to 00:33:27.200 --> 00:33:29.380 rush for the same exit at the exact same millisecond? 00:33:29.400 --> 00:33:31.539 That is a terrifying thought. The fat tail might 00:33:31.539 --> 00:33:35.049 be even fatter next time. On that cheery note, 00:33:35.130 --> 00:33:37.089 thank you for joining us on this deep dive into 00:33:37.089 --> 00:33:39.430 the rise and spectacular fall of long term capital 00:33:39.430 --> 00:33:42.509 management. It's a story of genius, ruin and 00:33:42.509 --> 00:33:44.930 the razor thin line that separates the two. Thanks 00:33:44.930 --> 00:33:46.430 for having me. It was great to talk through. 00:33:46.490 --> 00:33:47.609 We'll see you on the next deep dive.