Quants - The Alchemists of Wall Street Transcription:
Quants: The Alchemists of Wall Street - VPRO Documentary
When you're close to money ... As an oyster farmer ... I eat 100 oysters a week. How many people are doing this? Me, because they are there. When you are in the sector that earns money and circulates it ... and it's trillions, so of course we're going to make more money. You become isolated from the reality of the average worker ... who does not have this ability ... and who has trouble keeping his job, paying his mortgage ... feed and educate her children. This is the best part of my job. A white shell. No sickness. Making a lot of money is like taking drugs. We feel great from head to toe. It is like a drug. When you get a check for $ 1 million or $ 5 million ... we want more. "5 million now, then 50 million tomorrow ... because I am a genius. "There is no doubt of his own genius ... and that the rest of the world is ... We are so much better than the others. A bottle of beer costs 1 pound. But we're having a party, so we need a hundred bottles. How much is a hundred bottles of beer, if one bottle costs 1 pound? 100 pounds? No no no. Paul Wilmott, often described as the smartest of all quants. Mathematician and author of numerous textbooks. A quant guru who for years warns against the risk ... of a market collapse triggered by mathematicians. If we buy a hundred bottles in a store, we might get a discount. So a bottle costs 1 pound. A hundred bottles don't cost 100 but maybe 80 pounds. This is an example of non-linearity. You don't just multiply 1 by 100. It may also be that the store where you buy beer ... it's late at night, it's the only one open ... they know you need a hundred and they say: sorry ... we are only allowed to sell one bottle. But for 200 pounds ... we sell you a hundred. So it may be 80, and maybe 200. But not 100. And that is reality. But 99.9 percent of financial models don't take it into account. A quant uses quantitative techniques, mathematics ... statistics and IT to do ... valuation models for financial securities ... and to structure them, and cover the risks. Emanuel Derman, described as the Einstein of Wall Street. A former pioneer in ... who went to work for Goldman Sachs as a theoretical physicist. Now he is training new quants at Columbia University in New York. I started in 1985, when there were few quants. There were few rules. We were all amateurs. Now, you can be a professional quant, but at the time ... it was a disparaging term. Like "geek", a crazy computer in slang. A computer geek. We respect quants much more these days. Ten years ago... they were considered "geeks" who did not understand the industry. Now they are the sector. Manhattan, New York. A splendid and powerful place ... where I have a lot of experience and roots. But also a city where money marks everyday life. Now more than ever. I was a programmer for one of the main protagonists on Wall Street ... who is no longer with us. For obvious reasons. I created the plumbing of finance. I wasn't writing the models themselves, but the infrastructure around them. I could do it with my eyes closed. Some colleagues ... were content with themselves, smug, and lazy. Sometimes I worked on the trading floor. It was energizing. Most of my career I have been in a landscaped office. But I was located near the quants ... those who structured CDOs and other similar products, traders ... I saw what they were doing and it appealed to me. They seemed to solve thorny problems ... and work at the center of the action. It is as if I followed them through a window, but without being part of it. It made me want to do this job, and it's well paid. Each employer has a different view of these things and of the market. I don't want an agreement or disagreement like this ... influences my job search. This is why I have decided to remain off camera and anonymous. One of the first equations I derived in finance ... remains one of my favorites. The ideal is to have a good level of complexity, but not too much. The main thing here is that it looks simple ... but that there is a modulus, an absolute value, which makes it non-linear. And that is beautiful in itself. One of the problems right now in finance ... is that almost all equations are linear. It is... boring and unrealistic. Financial models mainly ask for data, such as: "What will the future interest rates be?" "What about rate instability?" "What about prepayment rates?" Then the model gives the present or future value of a thing. Models demand a vision of the future, which they translate into a price. They can't predict absolutely ... but only on the basis of your own vision. People call it forecasts, but they are the ones entering the data. So "forecast" is not the right word. -I do not think so. People like to say that this great global financial crisis ... results from bad models, but I don't think so. The crisis was caused by many things ... but the models are not at all the most important. Rather the incentives. Rather the whole system than a foolish person ... Finally, some have foolishly relied on the models. But no one can predict the number of prepayments ... or companies that cannot meet their obligations. A financial model is always based on a particular vision. My models ... I wrote software where we enter mortgage loans ... and expectations of default or prepayment. If we reach this fabulous moment, where everything works ... where the numbers fit ... and where it is easy for people to use ... well, yes, it's a beautiful time. We see them all using the software. We are perhaps the only ones to measure the effort provided, but it works. We were able to sell it to all the major investment banks. Including ABN Amro. He was one of our clients. And to all banks in Europe, Asia and the United States. Mike Osinski worked on Wall Street for 20 years. He was programming software ... that sliced subprime, high-risk loans into attractive products. Banks were putting billions into software ... and there were billions coming out of it. Initially it was so-called prime loans ... granted to people who were working. They were sure. People were paying. Then it became a huge business. Maybe already at the end of Clinton's term. I left in 2000. After that it was madness. They have given a lot of subprime loans. It was government policy. Both Clinton and Bush have ... forced banks to issue more subprime loans. The banks agreed, of course, because a subprime affair ... produced a hundred times more profit than a prime deal. They were happy to issue this type of debt. Corn... Did the buyers and customers of these titles really know ... what they were buying? Random CDOs are actually great instruments ... which consolidate mortgage loans and cut them into installments ... so that everyone can take the risk they want. It really is a find. To sell them you need a large profit margin. Because it is as much a profit margin as a margin of error. Its theoretical value is like that, but you have to sell it for that. And this covers possible errors. And when people compete ... the profit margin decreases. Because of the competition. And suddenly everyone is buying and selling CDOs. It is the largest of the instruments. But there is no profit margin, no margin of error. And the quants and their tools keep saying, Oh, no worries. And everything jumps in the air. I must apologize to the planet. I do it? I have warned against the risks of credit derivatives ... and mathematical models. "An in-depth reconsideration is called for ... to avoid a market collapse triggered by mathematicians. " When we want to warn people, we are supposed to do it in book form. We're supposed to write 300 pages about the dangers of ... etc. But I am a mathematician. A mathematician tries to compress 300 pages into a single equation. For them, it represents beauty. 6.30 hours. On the way to the library. I will stay there for twelve hours. Today's work: boring calculations ... the abstract manipulation of symbols, by hand, in pencil ... which could give 25 more handwritten pages. I like to learn, to solve difficult problems. And the challenge of representing part of reality with logical symbols. When I was 13, I understood that we can calculate the volume of a sphere ... by breaking it down into smaller and smaller rectangles. The flow of crowds. I like to observe the movement of crowds ... and feel the relationships and the forces that operate on them. A very busy public space. Crowds appear and disappear at the daily rate. I can see patterns in abstract mathematical symbols. What is the obsession with this formula? This is the formula that caused the Wall Street crash. As a mathematician, it's a combination of the sublime and the absurd. The sublime because it is mathematically beautiful. And the absurd because it must be used. It links the risks of default on two different things ... to the behavior of two companies, first independently ... then by way of a so-called "correlation". Is it serious if I do not understand what you are explaining to me? You do not understand? -No. We have already simplified the thing. -Indeed. Suppose you have a CDO. For example, there are a thousand mortgage loans in this contract ... and the risk is ... what it is. The main thing is a thousand different loans that you have to model. And this copula model includes a guess ... on the interaction between these loans. The relationship between Mrs. Jones' mortgage and that of Dr. Smith. How many correlations are there between these two things? The answer is 1000 times 999 divided by two. Which makes about half a million. How do you know those half a million digits? Impossible. So the followers of the copula say: Suppose they are all the same. It's too complicated. Let's say the number is 0.6. It doesn't mean anything. We have a very complicated model ... and complicated human interactions. So we go from sublime mathematical modeling to the absurdity of saying: They are all 0.6. Senior bankers don't necessarily know what they're doing. All of these quants have multiple doctorates. And the people above, the manager type ... do not understand the technical ideas that quants master so well. This can create problems: If a quant says something, what does the manager do? The manager has to believe the quant. And we hear stories of quants who base their models ... over five or six years of data. It is not really representative ... economic cycles in the real estate market. What effect will this have on the models? People above a certain age ... know that house prices can go down. FEAR I remember an occasion in 2007. A friend who works in the area paid me a courtesy call. She said: If the Fed doesn't lower its interest rates, we're all toast. I remember thinking: I sold all of my shares. Then: This is going to be a lot more serious than people think. Everyone had a home loan ... a car loan, and credit cards. If this market was about to falter ... And they continued to issue debt. People wanted to believe that everything was always going to keep growing. They were blinded by the remuneration they received each year. They are clean. Yeah yeah. These are huge numbers. Five, ten million dollars. A lifetime of income. No need to work anymore. Everyone wanted this. "I have gained so much in a year that I never have to work again." It seemed to me that this was everyone's goal. This is money. We earn money, and again, we earn it every year ... and then all of a sudden, we explode. The difference between my money and the money of a hedge fund ... it is because if it is my money, then I earn it, and there, I go bankrupt. But if it's someone else's money, if it's a hedge fund that's doing it ... they get a percentage of that every year as a profit. As a bonus, actually. They earn part of that, and that, and that. They put all that money into their bank account. And when they lose money ... it is not their money, but the client's money. It is therefore very easy to abuse this kind of construction. It's great that people who are at risk ... be compensated for that. But only if they really take the risks themselves. Take risks with other people's money ... does not deserve compensation. Sorry. Whatever economic theory says, you shouldn't be compensated for that. Unfortunately this is what this sector does. That's enough for now. At one point I wondered ... what I was looking for on Wall Street. Ice. Right now. Go hop. Ice. I had believed that people would be more judicious in their lending practices. I was involved myself, and I could tell they were on the rampage. Software is an inanimate object. It's true that people have used it ... but if they had entered good loans there would have been no problem. But when you enter home loans ... which we are relatively sure that people will not be able to pay them ... and that half of the loans issued in a year are like this ... so yes, the industry is unleashed. Trillions of dollars a year have gone through this model. Two or three years after their issue, these titles have gone from triple A ... to "no rating". A catastrophic collapse. Lots of brokerage firms with risky portfolios ... saw them lose their value. And given the debt ratio ... and everything the banks had borrowed ... there was a financial panic. Saturday, after midnight. I am still studying. I know I will continue to work long hours as a quant. I was mainly a technologist ... who did not understand everything that was going on. I think my motivation in becoming quant is precisely to understand ... now that I have experienced the ruin of my company. I looked again at my CV that I had circulated. The same headhunter called ... to offer me the type of job I had before ... as a financial technologist. I of course refused again. No proposal yet for a quant position. People in the area refuse to speak with the public. If they spoke to the press they would be fired. There are only a few people in the financial sector ... who have the opportunity to speak with the press. Once you enter this world, the phone starts ringing early in the morning. I remember guys I knew well ... who were trying to wake me up at 6 a.m. "Oh, I thought you were sleeping." And it lasted until 11 p.m. You have to be on edge and always vigilant. And absorbed in work. And you have to be perfect and be right all the time. If the software fails, people lose millions, billions. It is unthinkable. You can't go wrong. You have to be perfect. It is very stressful. My wife was in the same area, and we had the same nightmares. The phones were ringing, but we couldn't answer. Then it passed. We realized that we didn't know what day of the week it was. It's Lehman Brothers. The new Barclays. Barclays. It was Lehman. -They sank. Hey, Joe. Oysters. The banking sector no longer has anything to do with its original reason. Now it is getting dangerous. Some of these derivatives were invented ... to help, for example, a farmer insure the value of his crop. So that he does not speculate on the price of wheat, but cultivates it. Now those who trade these derivatives are more numerous ... than those who produce the goods. It's very weird. I know a lot of people in the banking industry who feel weary. People are starting to ask questions. My kind friends are wondering about their role in society. We make a good living, but how do we say to our grandchildren: "I was a banker. I was there when we caused the 2008-2009 crisis, etc." In fact, they spend their lives moving money. There is nothing to be very proud of. I think there are about 15 kilos of it. Accountability is not a one-way street. When a success goes wrong, you cannot become non-responsible. We necessarily have some responsibility. I could deny it, but I was involved. I made a good living with it. I was earning it very well. I was proud of what I had accomplished. I did not expect such a fiasco. They are rather big, the mussels. They are a bit wilder. They are yellow inside. A chef in town loves this wild taste. I only supply to one chef, otherwise there wouldn't be enough. The Hippocratic Oath of the quant: "I didn't create the world and it doesn't satisfy my equations." That is to say, you have to have the maturity to understand ... that whatever we do, the models will never be perfect. "I will never sacrifice reality for elegance without explaining why." Again, this competition between reality ... and the elegance of math. Sometimes the reality is just plain filthy. "I will not give false illusions about the accuracy of my model." "I will be explicit about his assumptions and omissions." A trader tells a quant: The risk you have estimated for this portfolio is too high. As for, go back to your drawing board and come back with a lower risk. We do not change the portfolio but the figures ... to make it seem less risky. Models can be used to hide risks."With the models I will estimate the value ... but I won't be unduly impressed with the math. " The financial sector has been made too mathematical. Lots of people who have to set up the models ... don't know what's going on. And when there is too much math, we no longer see the mistakes. "My work can have an effect ... on society and the economy beyond my comprehension. " So this is serious business. This is what it says. Quantitative finance in banking has become so huge ... that it is ahead of all other sectors. Whereas it should be a simple service to others. Now everyone is working ... to serve the interests of the banks. This is the impression we have. The banking industry has completely changed the nature of the world. There is a beautiful illustration in the book. It's me and Emanuel Derman with our Karl Marx beards. Because it's inspired by the Communist Manifesto. Combined with the Hippocratic Oath, this is the result. At first, I thought we could apply quantitative methods ... to financial markets. Now I don't think they are unnecessary but ... You cannot explain the markets with quantitative methods. Bohr, Einstein, Schrödinger, or Feynman ... discovered things that appear to be ... like the truth of God, even if they are not entirely correct. It doesn't seem possible to me in finance. It is an illusion. The financial world is changing all the time. History does not repeat itself. While in physics, experiences repeat themselves infinitely. After five or six years in this industry, I began to realize ... that it was not the same. In physics, if we wake up ... with an equation or theory in mind ... we have a very small chance of being right. But in finance, if we note assumptions ... they can be useful, but they will never be true in absolute terms. Because we're dealing with people, and people don't work that way. New weekend. There are so many parts of the city that I haven't seen ... since I started my training. I want to visit art galleries, eat in restaurants, live. The day at the library seems to have lasted over 12 hours. I study alone, among others who do the same ... like a monk in a monastery. It is peaceful. The library is quite old. Sometimes we have to cover the air conditioner ... with Soviet magazines from the 1960s. I dreamed of doing pure science, of building spaceships ... to work at a small start-up. There has to be a way to be creative as a quant. Like designing financial products and formulas. Do you think people can always express their concerns ... about the things they design or write? Yes, but at the end of the day these people are often employees. So we don't always listen to them. But you have to try ... and model users need to understand this too. But honestly I don't think the models are responsible ... of what has happened in the world. There has been a growing number ... a growing number of financial crises around the world since 1994. People are used to constant growth and acceleration. With each slowdown, the government has tried to revive growth ... by lowering interest rates, like now. It gives rise and fall, which we do not like. So we lend money again. These oscillations only increase. I don't know what to do, but now we're doing the same thing again. Every time the economy seems to be slowing, we try to get it going. It shocks me, as a business leader: If the economy contracts 1 percent, that is terrible. And if it grows by 1 percent, that's fantastic. That's a 2 percent difference. How can it have such an effect ... on the world around us? 1 percent more or less doesn't change my businesses. Economists see themselves as scientists. So they develop "laws". But these are not laws. The law of gravity is a law. Newton's formulations are laws. But economists offer frameworks, ideas. Who work or not. It is not about laws. They are wrong. So they build a great theoretical edifice ... on a very rickety foundation. And all kinds of nonsense comes out of it. I think the natural world is ... something that you learn to appreciate through a struggle. But in the financial world ... Money is an artificial phenomenon. It's like a game, and we make our own rules. But outside, the daily activity is not to earn money ... but to cultivate a healthy animal. Or a group of healthy animals. There is a big difference. This one is really beautiful. What is beautiful is that being neutral in the face of risk ... Remember, this is the risk neutral version, where µ = r. If this was reality, it would include dµ and dt. Let's do some manipulations. Some are simple. 6: Z. In this case there is no more Z at all. The d x dT version. Because we are looking for ... the stochastic differential equation ... not for log Z. And that gives µ - ½ sigma². Let's go back. If we have a stochastic differential equation ... for Z, then we can also write one for F. For a short period of time, the quants were frowned upon. They said: The bank has changed forever. No more quants and credit instruments. I said: You don't know history, or human nature. This will pass. And in a few months, return of the big bonuses ... everything went back to how it was before. If people don't complain now ... they will get what they deserve in the next financial crisis. Another 12 hours before evening classes. I feel connected to my classmates by the enormous workload. This is the commission that we are trying to maximize. Of course, you have to maximize profits. We have to manage their money well ... otherwise we will have difficulty in carrying out the commissions. I wanted a new challenge. I signed up for a quant training. $ 60,000 in tuition fees. So I am getting even more in debt. Given the structure of incentive bonuses ... maximizing the TWR is maximizing the commission. Think about it. It's rather risky. The course lasts a year and a half, full time. A year and a half without pay, with Manhattan housing charges ... and full-time tuition fees. So not a drop of alcohol until the end of the first semester at least. I can't afford a single hangover. I will see few people. Most of the other students are Asian or Eastern European. Math is their first language and our common language. We Americans are a minority. By maximizing the number of times we will exceed ... the previous climax, we are actually maximizing the bonuses. We can see that this type of optimization looks a lot like a hedge fund. Do you understand for the moment? Before, physicists split the atom. But who does this now? Who builds the bridges? Everyone wants to work in this industry. Scientific creativity becomes financial creativity. It's pipo. Quants are essential for modern banking ... which is based on techniques ... such as Algorithmic Trading and High-Frequency Trading. For that, you need math skills. Before we had traders and brokers ... who shouted on the stock exchange floors and traded stocks. An order was coming and they were rushing forward. They wore jackets of different colors. We know these images. Matthew Goldstein, investigative reporter for Reuters ... and one of the first to tackle the consequences of High-Frequency Trading. But the prosecution is only a small part of stock market transactions. The action is behind the scenes and in the models and software ... computer geeks. High-Frequency Trading analyzes a lot of data very quickly ... and produces negotiations that can last for milliseconds. I am very disturbed to hear that certain agencies are being awarded ... faster market access. Large corporations have a tenth of a second advantage. It is unfair. Hedge funds are trying to bring their black box closer to a stock market. The signal takes some time to go from the black box ... until the stock market, for the transaction. This is dictated by the speed of light. We are talking about negotiating at the speed of light. The classic stock market crash of October 19, 1987 ... happened in a day. The 20 percent drop in the S&P 500 index. The next crash could take a few minutes. What is a black box? -A black box ... contains a formula, which is sometimes secret. We put a lot of data ... like stock prices and other information. Then she tells you what to buy and sell. My favorite is the Google search terms. To negotiate... based on what people ... they search. It's not a black box in the sense that ... If we saw the algorithms, you and I wouldn't understand them ... but the heads, perhaps, and the computers surely. So it's a black box because the human brain struggles ... to understand what is really going on. Goldman herself has also been called a black box. How do they go about making all those billions and big bounties? The New York Stock Exchange is building a facility in New Jersey ... intended for High Frequency traders. Their equipment ... will be assembled in a small factory for High-Frequency Trading. Who can put their server where? Are they going to draw lots? Or pay more to be closer? It's absurd when you think about it, that we have come to this. The battle will focus on who has the most resources ... and who pays the most for the best buds. In fact there are only a dozen stars. We can't all be clients at Goldman Sachs, or Morgan Stanley ... or Barclays. With High-Frequency Trading people are more concerned with the price ... than by value. The value of a thing is what it is really worth. The price is just what you buy it for and sell it for. If we buy something and sell it two seconds later ... the main thing is that the sale price exceeds the purchase price. Value, who cares. It goes against our image of the markets. The companies themselves hardly matter. What matters is the movement of their actions. Which is pretty scary, from that point of view ... this is the generalized risk that this automated trading may involve. It goes so fast that the human element is disappearing more and more. Of course, human beings write software ... but they are not involved in the trading itself. We saw that with the rumor that United Airlines was going bankrupt. A press agency had transmitted ... an old story about filing for bankruptcy. All of these algorithms read the news and started selling. United Airlines' price halved. This is a good example of computers having lost their minds. The banking sector is taking power on the planet. It has a major effect on the man in the street ... and it shouldn't be. In fact the sector must take money from those who have too much ... and give it to those who don't have enough and want to start a business. But now the banking sector is playing with numbers ... without realizing that they represent people and jobs. We joke that the New York Stock Exchange will become a museum ... a show with brokers in a hurry: "We used to trade stocks like that ... How cute." But if there is a bad backlash in High-Frequency Trading ... we may see a return of the human element. People may say: We may be imperfect ... we don't want machines to control everything. I was walking in a Wall Street full of Chinese tourists ... who asked me if it was really the Stock Exchange. Wall Street doesn't matter as a location anymore. Many banks have moved their headquarters to the center ... and their administration to New Jersey where it is cheaper. Deutsche Bank is the only major bank still on Wall Street. Nearby is Goldman Sachs, which is also about to move. There are hardly any large company headquarters in the cradle ... of American finance. Only the New York Stock Exchange is still there. Its facade is one of the emblematic symbols of world capitalism. I try to interest young people in this work. It is for young people, it is physically quite tiring. They haven't grown so much. I may not have enough until Christmas. I prefer them larger. Usually we sell them bigger. What is more satisfying? With software, pleasure is quite intense mental work. We memorize a million lines of software. It's like the fun of Scrabble, the fun of rebuses. But it's not very healthy. We are glued to a machine, to a screen. It's not like this at all. Today it is idyllic. We're on the water. It's a good feeling to bring in food. We sell about 150,000 oysters. For $ 100,000. But of course I live off my interests, so ... It's nice to have, it's pocket money, etc. And the overheads are pretty low. You have to learn to accept nature, like these oysters. Every year I pick them in September and October ... and they are ready in November. But there is nothing to do. Whereas with software, we can always do. We can modify. It's a virtual world, you can do whatever you want. Here, we are limited ... by reality.