Financial Theory

ECON 251 - Lecture 19 - History of the Mortgage Market: A Personal Narrative

November 8, 2009 (recorded December 8, 2010)

Chapter 1. Fannie Mae, Freddie Mac, and the Mortgage Securities Market [00:00:00]

Professor Geanakoplos: Hi, this is John Geanakoplos again. Here to give a lecture, that I gave one evening, that we couldn't record. So I'm going to try and reproduce the lecture as faithfully as I can. And I think it's a historical lecture. I hope you find it interesting. It's about the history of the mortgage market. And I call it a personal history because by some accident, I participated at many of the key points in the recent history of the mortgage market.

So I started off in 1989. I was professor here at Yale. I was a mathematical economist. I thought of myself as using mathematics to study economics and staying really pretty much as far from the real world as I could. But for some reason, I decided that I wanted to see what was going on on Wall Street. The most interesting mathematical modeling of that day was being done on Wall Street. And mathematical modeling and finance, that is in economics. And so I decided. why not go see it?

And so I visited a bunch of investment banks. A number of my friends, including one from Yale, had worked at Goldman Sachs. So that was the natural thing to do. But, I had a little cousin who had just been hired recently, a little bit before that, at Kidder Peabody, which was sort of the number seventh investment bank at the time, in terms of size. And he introduced me to the fellow, Ed Cerullo, who ran fixed income at the time. And they persuaded me that it would be much more interesting to go to Kidder, Peabody and to talk to people like Ed Cerullo, than go to Goldman Sachs and be one of a hundred visiting professors.

So I decided that the firm was a little bit smaller, but I would see more of it. And why not do something a little bit different? So they never had an academic visitor, I think, like me spend a year there before. So I went in 1989 to 1990. And while I was there I talked a lot to Ed Cerullo and to the traders and to a bunch of other people. And at the end of the my sabbatical, Ed Cerullo came to me and he said, you know I've come to realize that our fixed income research department isn't very mathematical. Why don't you hire a research department for me? And I'll help you along. But, you find the people, you know the subject, you can judge, your business is judging people doing research, why don't you hire me a research department?

So I hired him a research department, which ultimately grew to 75 people. And I returned to Yale. And after I got back to Yale, he called me up and he said, now that you've hired the research department, you have heads of all these different groups, why don't you run the research department? You can do it from Yale, as a consultant. And so I became the head of fixed income research at Kidder Peabody from Yale.

And it was quite an experience. I hadn't realized, when I accepted the job, just how many complications would arise. How many people would get job offers from other places and want to leave. And how many models wouldn't work. And then there'd have to be a wild scramble to fix it. But anyway, that placed me at an investment bank, that would become one of the key players in the mortgage market.

Now in the mortgage market, well who are the players? They are the homeowners you all know about. They are the banks who are giving mortgage loans. But then there are a number of other players that are invisible to much of the public, which really dominate the market. There's the government agencies, Fannie Mae and Freddie Mac, which you'll hear a lot about. There are the investment banks, like Kidder Peabody and Goldman Sachs and a bunch of others. There are the hedge funds and there are other investors.

And this group of people creates a gigantic hierarchy, an invisible market, that's on the same size, the same scale, as the stock market. So I think in that time, 1989, 1990, if you'd asked anyone in America practically, what's an important financial market? They would have said, the stock market. What's another important financial market? Corporate bond market. What's another one? Well, foreign exchange market. What's another one? Options market. I don't think anyone would have said the mortgage market. Or at least they wouldn't have said it very early on in their list of important financial markets.

But in fact, as I'm going to try and tell you during this class, the mortgage market is not only on the same scale as the stock market, but actually quite a bit more complicated than the stock market. More mathematical than the stock market. And in some ways, more interesting than the stock market. And as we'll see at the end of the course, it was the mortgage market that led to the greatest crisis we've had since the depression. And in fact, caused several similar crises before that.

So mortgages appear at the bottom. You have homeowners living in their houses who need to borrow the money to buy the house. Before they can live in the house they have to get the money to buy the house. So they borrow the money by taking out a mortgage. And it's a bank or a thrift or somebody like that who lends the money. That's the first step.

So a mortgage is just a promise to pay the loan back over a long period of time using your house as collateral. An important part of mortgages is that you have an option to pre pay the mortgage. I mean, what if you move? The mortgage is, say, a 30-year promise. And after three years, you might want to move. So, if you're gone from the house, the house can no longer serve as collateral because you don't live in it anymore. So there has to be some way of getting out of the mortgage. And so there's an option to pre pay it.

Now, in the United States, that option can be used even if you stay in the house. And it turns out to be one of the more problematic aspects of valuing mortgages. And one of the most interesting aspects of valuing mortgages. We'll come to that later.

So this idea of a mortgage, those three ideas, were already known to the Babylonians more than 3,000 years ago. So the mortgage is not a recent invention. It wasn't invented after the industrial revolution. It was invented more than 3,000 years ago near the Middle East.

So, it's stayed pretty much the same for most of those 3,000 years until the 1930s when the amortizing mortgage was invented. So what happened in the 1930s, that was the time of the Great Depression and farmers had lots of mortgages and they would owe, $7.00 a year, say, as their interest payment. And at the end of 10 or 15 years, they'd have to pay $107, that is they make their interest payment. And they pay back the balance. What's called now a balloon payment. That's how a typical bond works.

Well, of course when things got really bad, a lot of them defaulted. And naturally, they chose to default just before the $107 payment. So seeing that, mortgage lenders decided that it would be much safer to make a flat mortgage loan where the payment was say, $8.00 a year for all of thirty years. Now, you pay a little bit more each year for 29 years, but then you continue to pay the $8.00 the 30th year. But you see, if you add up all the extra payments and you realize that there's discounting, the $107 way off in the end isn't really that much money when looked at from the beginning.

And so by paying $8.00 every year, you can get the same present discounted value of paying $7.00 29 years and $107 the 30th year. So that's called the amortizing mortgage. It makes the lender much safer because after a bunch of years have gone by, the house presumably has gone up in value. Or even if it hasn't gone up in value, the remaining payments are much less because so many have been made that the balance has been amortized. And so actually to get out of your mortgage, you need to pay less than $100 back. So this amortizing mortgage is something we're going to study mathematically in the next lecture. But now, I just mentioned that was one of the big innovations, which made the mortgage market much safer than before.

Well so things continued pretty much the same from the 1930s all the way to the 1970s when we had securitization. Like many of the great financial innovations in history, this one was created by the government. So Fannie Mae and Freddie Mac were government Agencies They were created by the government. They then became eventually separate from the government. But they were given the task, they were created for the purpose, of making mortgage pass-throughs. So we're in the 1970s.

So mortgage pass-throughs are the second tier of this hierarchy. So the banks who had lent the mortgage, remember when you take out a mortgage as a homeowner, you're selling your promise. You're getting the money by selling your promise to pay back later. So those promises are collected by the banks. And instead of just sitting on the promises, the banks now, with Fannie Mae and Freddie Mac could sell their promises to Fannie Mae and Freddie Mac. And Fannie Mae and Freddie Mac would put them together in gigantic pools called pass through pools.

Why were they called pass-throughs? Because the mortgage payments the homeowners made would go to the banks and the banks would just pass them on to the pools. And then the pools would collect the money and pass that money on to the shareholders. So the ultimate lenders to the homeowners are the people who buy shares in the Fannie and Freddie pools. So the banks appear to the homeowner to be lending the money. But actually, they're not lending it all. They are the middleman. And so they collect the money from the homeowner and they send it on to the actual lender, who's the shareholder of the pool.

And the banks are also the servicer, really. They're getting paid a fee for collecting the money and writing threatening letters if the homeowner stops paying. And as we'll see later, throwing the homeowner out of his house if the homeowner doesn't pay. So Fannie and Freddie played an extraordinarily important role in the mortgage market. First of all, not any mortgage could be sold into these pools. They had to meet strict criteria. You had to have a good credit rating. You had to have a job. You had to have all sorts of-- sorry about this, I'm going to have to shut off my cell phone.

So the loan to value of the mortgage had to be, that is if the house is worth $100, the loan could only be 80% of it. You had to have a record of the job you had and so on. So this was very standardized and very high performing loans. You could contrast with, say, to the world that you might have seen in the movie It's a Wonderful Life. So in the movie It's a Wonderful Life, you remember Jimmy Stewart runs a thrift and he makes mortgage loans. And people come to him and they say we want a loan. And you know one guy comes and says he wants a loan.

Jimmy Stewart says, well do you have any collateral? No, I haven't built the house yet, you know I'm trying to build the house. Do you have a record of employment? No, I just moved here, I don't have a job yet. Do you have someone who can vouch for you? No. Do you have a credit rating? No. There's no such thing as credit rating.

So then Jimmy Stewart looks into his eyes and realizes this is a good honest person and gives him a loan. Well, that doesn't happen in the Fannie Mae and Freddie Mac pools. They're very standardized criterion. And that guy, Martini who got the loan from Jimmy Stewart would never have gotten a conforming Fannie or Freddie loan in the 1980s or 1990s. Might have gotten one in the 2000s though, but we'll come back to that.

So that kind of market is very much like the kind of market in the Jimmy Stewart movie, It's a Wonderful Life, that gets created in his fantasy. Where you know, the town gets taken over by the evil banker and that evil banker, whose name I've forgotten at the moment, but that evil banker basically is creating the kind of loans, almost, that we're talking about now. Everything's mechanized and standardized. Of course with standardization you get tremendous advantages.

For one thing, these loans being pooled together and being of the same general good quality, they allow the lenders, instead of lending to a bunch of homeowners in Peoria like the bank would do, now all those Peoria loans are stuck with a bunch of other loans from all over the country in the same big Fannie pool. And so the lenders, who are the shareholders, have diversified their risk.

If the big businesses in Peoria, Illinois go under, it might be that all the homeowners in Peoria will default on their mortgages. But not in the Fannie and Freddie pools, because those are loans from all over the country. So it would have to be that businesses all over the country went bad for those loans all to go bad.

So secondly, if you're getting a whole pool and there's an automatic criterion for getting into the pool, you don't have to worry that you're getting the worst loans or the cheatiest loans or something. You know the quality, the general quality, of all the loans. And once you have shares in a pool, you can resell the shares. So a bank who has to study the homeowner and you know have meetings with them.

And you know, Jimmy Stewart had to look into his eyes. So Jimmy Stewart may have convinced himself the guy is a good risk, but how could Jimmy Stewart ever sell the loan to somebody else. The other buyer, who hasn't looked into Martini's eyes, he's never going to believe Jimmy Stewart that that's a good loan. So Jimmy Stewart is going to be stuck with that loan for 30 years.

In the Fannie and Freddie pools, the shareholder who buys the loans and knows they're standardized knows exactly the same thing as the next guy who might buy his shares from him. So if the shareholder will be willing to pay more for the loan, because he knows that if he needs cash, he doesn't have to wait 30 years, he can just sell his shares to somebody else.

So because of the diversification, because of the reduction in adverse selection, and because of this ability to resell the shares, lenders, that is shareholders, are willing to pay more for the mortgages. And so the mortgage rate went down. So, I've estimated that this operation together with the next one I'm going to talk about, has reduced mortgage rates by at least a percent.

So if you think the average loan is $200,000, you're talking about $2,000 a year that the average homeowners save by this financial innovation. So securitization seems to have been a great boon.

Now in 2002, when I made these slides, I'll just give you an idea of the size of the mortgage market. And you have to double all these numbers today, pretty much. So you see that you know the stock market was around $15 trillion and the mortgages around $7 trillion at the time. And you know that compares to $2 trillion for corporate bonds or $3 trillion for treasuries.

You see how big the mortgage market was then, and now it's twice as big. It's the same size basically as the stock market, which hasn't really grown since then. So, I don't have time to talk about this. But, of course, of the mortgage market, some of it is commercial, some of it is residential. The vast majority is residential. But there's a very big commercial mortgage market. I'm going to be talking mostly about the residential mortgage market.

Some of these mortgages are in fact held by the banks without selling them to Fannie Mae and Freddie Mac. But a lot of them are securitized just in the way we talked about. And they're going to be private securitizations later that we'll talk about. And so now that securitized part is $10 trillion. I'll give the numbers, recent numbers, later. So actually that may be a little bit big. The $7 trillion, the securitized part of the mortgage market.

So in 2002, the agencies dominated the security market. The securitized loans were almost all Fannie Mae and Freddie Mac. There's also another agency called Ginnie Mae, but there was another part of it, which were jumbo loans. So there was a size limit. The loans couldn't be too small and couldn't be too big for these Fannie and Freddie pools. The government was trying to appeal to the middle class. Establish homeownership in the reliable middle class.

And so the wealthy who were buying million dollar homes weren't able to get their loans sold into a Fannie or Freddie pool. And so those loans were securitized the same way, but by private agencies and not by banks, by investment banks, and not by the government. And that at the time, in 2002, was half a trillion. We'll come to all these numbers later. When we say today's numbers.

Chapter 2. Collateralized Mortgage Obligations [00:17:01]

So the next innovation after the 1970s, came in the 80s and that's the collateralized mortgage obligation market. CMOs, they are called. So the investment banks like Kidder Peabody and Lehman Brothers, for example, would buy some of these big pools. And then they would cut the pools, which were just pass-throughs. So the pools just passed through the money that the homeowners were giving them. So maybe they were passing through $1,000 a month as the promise.

Well, say Kidder Peabody, might buy a pool promising $1,000 a month to the shareholders. And then cut the promise into two pieces. Maybe a floater, which would pay $500 plus the interest rate. And so when the interest rate went up, the payments would go up, that's why it's called a floater. And maybe a second piece called an inverse floater, which is $500 minus the interest rate. So as the interest rate went up, the payments would get smaller. But as the interest rate went down, the payments would get bigger. That's called an inverse floater.

So that way the two add up to $1,000. But now, you can appeal to two different buyers. A buyer who needs the money when interest rates go up would buy the floater. A buyer who needs the money when interest rates go down, would buy the inverse floater. So by creating out of plain vanilla promise, two more tailored promises, you can target more sharply a clientele. And therefore get probably more than half the money for each of the two pieces.

And that way, of course competition in the CMO market raises the amount people are willing to pay for the pass-throughs. Because they can then buy them and split them up. And sell them for more. So that raises the price of the pass-throughs, which in turn raises the price that the homeowners can sell their mortgage, which in turn lowers the interest rate that the homeowners have to pay. So again, it's another reason why this whole operation of securitization improved the welfare of almost everybody.

So, the pieces gradually got more complicated So if there was default or somebody prepaid, as we talked about, used their option to pay early, instead of getting the money you expected, you get extra or less money than you expected. That created risk. And some of these buyers didn't want to bear the risk. So maybe you'd make the pieces $400 plus the interest rate and $400 minus the interest rate. And leave $200 to what might be called a residual piece or a derivative or something. And that $200 would bear the risk. So if someone defaulted, you take it out of the $200 piece and not of the first two pieces.

So that split up, again, made the floater and inverse floater safer and encouraged people to buy it. But of course, somebody had to buy the residual piece, which was more complicated. So, as I said, this whole operation was a way of accomplishing two things. It made homeowners able to sell their promises for more. So they effectively were paying a lower interest rate. So it made it easier to move into houses. And that's precisely what the government intended by creating these agencies. It also allowed buyers and investors to get money in the cases they needed it, in the states they needed it. Because the pieces were tailor made for them.

So I realized, while I was there at Kidder Peabody, that this whole multi trillion-dollar operation behind the scenes was, as I said, invisible to everybody and was worth writing about. It was bringing great welfare gain to the country and nobody quite knew about it. So for me, it crystalized what the essence of finance is. The essence of finance is you're trying to create promises, the financial system is creating promises, that deliver money to people in circumstances or states as I call it, that they really need the money.

But of course, you have to guarantee that the money is going to be paid to them. So in order to do that, you have to have collateral. So this entire system is a way of creating promises and backing them with collateral. So if you remember, here are the potential promises, and maybe some people want money down here. That's when the interest rates go up, they're buying the floaters. Way over there, the money's going down, the interest rates are going down, that's the people who buy the inverse floaters. That's when they're getting most of their money.

But these promises, you have a reason to expect to get paid, because they're backed by the collateral. OK, so this is not a very good picture. The houses backed the promises of the homeowner like in Babylonian times. If the homeowner doesn't pay, he loses his house. Then those mortgages themselves, those mortgage promises, are backing the pools, which are selling shares to the shareholders. But those shares are backing the CMOs, which are making more complicated promises.

So the collateral you see is used once by the homes, once by the pools, once for the CMOs. And then it will turn out, as we'll see in a few minutes, that the investors who buy are borrowing money, buying on margin, using the CMO pieces they buy, as collateral for their purchases. So the collateral is being used and reused. And so in fact, the entire system is stretching the available collateral as much as possible. So collateral is a very scarce resource. It's very important to running a financial system. Many developing countries don't have any collateral. And therefore they have a primitive financial system.

Here, there's a tremendous incentive to stretch the collateral as much as possible. You can stretch it by using it over and over again or by letting the same collateral back many promises. We saw both of those, the tranching, the different CMO pieces, and the pyramiding, the using the same collateral over and over again.

Chapter 3. Modeling Prepayment Tendencies at Kidder Peabody [00:22:44]

So I wrote my first paper, published it, in 1997. I had written it while I was at Kidder Peabody. And one of the questions I asked was, when you take out a loan, not only what interest rate do you have to pay, but how much collateral do you have to put up? And that was a question that nobody seemed to have asked really before, in a general equilibrium model. So that was my first paper on the subject.

I want to get back to Kidder Peabody. Kidder Peabody, as I said when I started, was a sleepy number seven ranked investment bank. Didn't dominate any market. But it came to completely dominate the CMO market. So this, as I said, was a multi trillion-dollar market. And if you look at all the pieces that are being promised here, what Kidder decided to do was, things got more and more complicated, there weren't just two pieces, there weren't just three pieces, there were 90 pieces typically at that time.

And so a typical investment bank that wanted to buy a pool and sell these CMO pieces. Like let's say, Salomon or First Boston, they were the ones who first did these CMOs. They would try to line up 90 buyers and figure out what each of the 90 were willing to pay. And then when they added up the prices each of the 90 willing to pay, they'd go and if they thought the sum of those was bigger than the price in the pool, they'd go to the pool and buy the pool. And then immediately make a profit by selling off the 90 pieces.

Well, at Kidder, the head of mortgages, who by this time by the way, was my young little cousin, who through hard work had worked his way up to running the small CMO operation. He got the idea that what Kidder could do, was to find a buyer for the riskiest piece. The one that I've put in yellow, remember, the residual. Which was maybe called an inverse IO, it had different names, but the riskiest pieces.

Once he found the buyer, a place to put the riskiest piece, he would borrow the money to buy the pass through, and hold an inventory of all the other pieces. So he knew, that eventually, we would be able to sell off all the other pieces. And so we had a tremendous advantage. While everyone else was looking for buyers for 90 pieces, we had to find a buyer for one piece or two pieces.

So we then had an inventory of 88 pieces, say, to sell. And as we did deal after deal, that inventory would get bigger and bigger. So our sales force would have a much easier time selling a piece. We wouldn't have to call someone up and say, do you want piece number four, that's the one we're trying to get you to buy.

We'd call up and say we've got a huge stock of different kinds of pieces, of all kinds, from all kinds of deals. Maybe you're interested in one of them. So it's much easier for a sales force to sell with that situation than it was for the other investment banks to sell. So of course, we lured away some of the best sales people.

So of course, the down side to all of that is we had to be very careful that the pieces we held, we knew how to hedge. We had to be very careful that when we held all these pieces, between the time we held them and finally sold them, we didn't lose a lot of money. And sometimes we were the ones who held the most dangerous piece, too. So we had to figure out what were going to be the cash flows of these pieces and how to model them. And that's what we're going to mathematically talk about the next few days. But I'm going to give you a hint of this now.

So we had to predict, among other things, what the prepayments would be, OK. And so, if you talked to a macro economist, especially in those days, and you say, what do you think is going to happen in the world? They'll usually say, well I think in the next two quarters, unemployment is going to go up half a percent, and then maybe things are going to get better for the next year, and beyond that it's too hard to tell.

So these mortgages are 30-year mortgages. You can't have a prediction that lasts a year and a half, you have to have a 30-year prediction. And the macro economists, by the way, are typically wrong. Even in their two-month prediction or six month predictions. So how can we risk billions of dollars holding an instrument that's 30 years long, which depends on what people are going to do over the next 30 years?

Well the answer is you can't make a prediction and expect it to be right. But you can make a conditional prediction. You can say, if interest rates do such and such for the next four years, and housing prices do such and such for the next four years, then in that fifth year, pre payments will be so and so. That's a conditional prediction. It's much easier to make a conditional prediction than an unconditional prediction.

It's shocking that so many economists are lured into making unconditional predictions. It's not a business we should be in. We should always be making conditional predictions. Well, I learned that lesson at Kidder. So, you know, the idea of possible worlds is going to play a central role in our course from here on out. And the possible worlds are the paths of interest rates or home prices.

And so here's a very short, this is maybe a one year into the future, and we'd have to go 30 years in the future. As you can imagine, and this is just actually a much smaller group of possibilities than we used to consider. But, you see all the possibilities that could unfold over the next 12 months. And you see the blue line is one of those possible paths.

So we have to predict, if we knew which path interest rates and housing prices were going to take, like that blue line, could we then predict, at the end of the blue line, what pre payments would be. Of course, if you followed another path down to here, we'd make a different prediction about pre payments.

So, the cash flows are the pre payments and defaults. That's what we have to predict. And so, you might say, what good is it to predict a different number on each of these trees. Then basically what you're saying is you don't know what's going to happen. But that's far from the truth.

If you know what the payments are on each of the paths and then you can hedge those paths. So for example, suppose the payments are very low here and very high at the top. Well, if you can buy another instrument that pays you something at the bottom, and you pay at the top, you can offset the variation in cash flows and the mortgage by that other instrument and guarantee yourself the same safe payment all the way through.

That's the sort of thing that we did it at Kidder Peabody, which we’re going to study in great detail. We held these complicated pieces whose payments would vary tremendously. So we ran a huge risk of losing all our money. But then we would hedge them with some other instruments. So in the end, we've got a pretty safe return. And because we knew the return was safe, we felt we could pay up a good amount of money for it. Whereas other buyers who couldn't hedge it, wouldn't want to buy it at all.

So, now let's just see how good the predictions were. Here's a typical history of pre payments. So you can see it ranges to 99 from 88, something like that. So it's 10 years, 11 years, of pre payments. So that's the percentage of people who prepay every year. So it's the annualized percentage. So every month, you check how many people paid off their mortgage early, and you assume that rate stayed the same for the whole 12 months, what would the pre payment percentage be? That's what the number measures all the time. You see in some months it's practically zero.

These are pre payments for a mortgage issued in 1986 that had an 8% coupon. So all basically, 1986 mortgages with an 8% coupon issued by Fannie Mae or some big pool of them anyway. So you see sometimes the pre payments are very low, sometimes they're incredibly high.

How standing at the bottom at the back there, in '88 say, could you possibly have predicted that pre payment going forward? You know with the ups and downs and stuff like that. Well you couldn't, if you made it unconditional. But if you made it conditional, it's not so hard. Because what do you think happened in '93 that made pre payments go up so high? It was interest rates went down.

Homeowners had an opportunity to refinance at a lower interest rate with a different mortgage. So of course they did that. And in those other months, when pre payments are very low, the interest rates were higher. So clearly, there's a connection between the mortgage rate you got and your original mortgage and the new rate that you can refinance into. And that is part of your conditional prediction.

But another thing that's interesting is burn out. So if you look at two pools, one issued in '95 and one issued in '92, and with exactly the same coupons, you'll see that the older one is always pre paying less than the new one. At least starting from '95 onward. So it's as if the old one burns out. Once it has an opportunity to pre pay, you see a lot of pre payments and then they slow down.

So that's another thing people had observed. So you might have thought the standard way of modeling things at that time was to just do a regression. You say, well we figured out that pre payments depend on the new interest rate and how much more in the money you are. And it depends on how long you've been in the money because of the burn out. You try to estimate a curve, like an S curve or something, that depends on parameters, that's on the interest rate, and on the burn out. How long you've been in the money. And according to some parameters that described the S and you tried estimate those parameters.

So that would be an old fashioned way of measuring things. But as we're going to see in this course, we look at everything from an agent based approach. So what are the individual agents doing? So, I got the idea of trying to model all you care about is aggregate pre payments, what is the whole pool doing? But I decided, let's try and predict every single homeowner.

So let's try to put ourselves in the mind of every homeowner in the country. Why are they pre paying or not pre paying? Well, they're pre paying if they get an opportunity to save some money. But, we know from that graph that lots of people don't prepay. Even when there's a tremendous opportunity, only 60% percent of the people are pre paying in a whole year. So that means the whole year goes by, and only 60% percent of them have prepaid.

So every month, 8% or 10% are pre paying. So the other 90% haven't seen their opportunity. They've missed it or they've waited. So clearly, not everybody jumps at the opportunity. So it must be that people are different somehow. Even though they're in exactly the same circumstance in terms of refinancing.

So, you have to account for the difference. So, I imagined that different people have a different cost of prepaying. It's a hassle to prepay. Maybe you literally have to pay some money to prepay. Maybe you have to take a day off from your job. Some people have other things to do. They're not that alert because they're paying attention to their kids or they're paying attention to their work.

So not everybody has the same alertness. And not everyone follows financial matters as closely as everybody does. So, also over time, people are getting more rational and beginning to understand the market and pre paying more and more. And of course, people hear from their friends. If their friends are all prepaying, it's more likely that they'll think about it and prepay themselves.

So I built a model, together with the researchers at Kidder Peabody. And then later we built the model of pre payments based on every individual making a different decision. So an individual is characterized by his cost and his alertness. So different individuals have different costs and different alertness. And by watching how they behave, we can come to guess what the cost and alertness is of each of these people.

So you see this model captures all the effects that we talked about already. If you have a pool of people, you can see the vertical is how many people of each types. So there are different costs and different alertnesses. And so at the beginning, when you have a new pool, like on the right, there are a lot of people. And as, they get opportunities to refinance, it's not random people who prepay, it's the people with the highest alertness and the lowest cost who prepay.

So over time, as a pool gets older, the distribution of people is going to shift. It's going to be have less hyper alert and low cost people and more high cost and low alert people. That's why the pool is going to slow down. So burn out is naturally explained by an agent based approach. So anyway, as we go through the course, we're going to emphasize this agent based approach.

So building a model like that, starting in the 1980s, and making a conditional prediction, you can get a fit that looks like this. So notice, you make some gruesome errors, like over here, that was an expensive, '97, '98, that was an important mistake. But you can fit this kind of pre payment surprisingly well.

Chapter 4. The Rise of Ellington Capital Management and the Role of Hedge Funds [00:35:40]

OK, so that was my Kidder Peabody days. And I thought, my gosh, we're doing incredibly well. We're helping the country. We're doing things in a colossal scale that nobody had ever done before. I think I may have not emphasized enough that Kidder Peabody came to dominate this market.

We controlled over 20% of all the issuance. Remember, there were trillions of dollars of things being issued. So this little group of 20 year old kids basically, plus me the old guy in research. So them, the traders, they were issuing something on the scale of half a trillion to a trillion dollars of these CMOs. These kids in their mid 20s or late twenties.

And the world seemed to be a better place for it. And I thought, my gosh this is an untold story that needs telling. And it's an incredible success story. Well, things suddenly changed and in 1994 there was a crash. So this is the first of three crashes I've lived through.

There was a scandal at Kidder Peabody, the Joe Jet scandal, who was a trader, a government bond trader, who was accused of doctoring the books and faking his big profits. He had been Kidder Peabody man of the year in 1993. And then in 1994, they decided that all his profits were fictitious and that he doctored the books.

And so he was fired. But he sued Kidder for discrimination. And it was a tremendous controversy that was in the front pages of the papers for months on end. And finally, General Electric, who owned Kidder Peabody, closed the firm. After hundred 135 five years, or I guess 129 years, they closed the firm.

So I had to go back one day, from Yale, to Kidder Peabody, and I invited those 75 people in the research department into my office, and I said you're fired. And then I got up and I went to the office next door and somebody said to me, you're fired. And so we all fired each other and the entire 130-year-old company came to a close.

So, the head trader of mortgages and bunch of his top lieutenants decided from a hedge fund. So Kidder Peabody got closed, actually sold, it wasn't closed, it was sold to Paine Webber. And Paine Webber dropped the name Kidder Peabody and hired many of the people from Kidder Peabody. And in fact, the second tier, or the more junior group of mortgage traders at Kidder Peabody, which was the leading mortgage company at the time, they took over the desk at Paine Webber.

And then Paine Webber was bought by UBS. And those same guys took over the desk at UBS. So the junior crew at Kidder Peabody, some of whom were Yalies by the way, ended up running the UBS. Or a big part of the UBS mortgage desk.

But in any case, our mortgage traders, the head guys, decided to from a hedge fund. Instead of selling, the CMOs, they would buy the CMOs. And we called it Ellington Capital Management. So I was one of six partners. I was a small partner, because again, I stayed at Yale.

So this introduces the last player, the hedge funds. So what is a hedge fund? You heard the name, I'm sure a thousand times. It has a bad name now. But a hedge fund basically means four things. It means it's someone who hedges. You don't just buy and hold the thing and hope that the cash flows are good, you try to protect yourself against as many risks as possible.

Just like I explained, we were doing at Kidder when we tried to get the same cash flows in every scenario. And we're going to mathematically study that in remaining lectures. But you try to hedge, that's where the name comes from. So you try to offset as many risks as you can. Of course, you still run some risks. But you're offsetting setting as many as you can.

The second most important thing, is most of it's done, or a lot of a buying is done, not most, but a lot of the buying is done with borrowed money. That's called leverage. You don't just take your investor capital and buy something, you take your investor capital, you borrow some extra money, and then you buy some stuff. And the stuff you buy you use as collateral to guarantee that the lenders are going to get their money paid back. So a hedge fund is generally leveraged.

The third definition of a hedge fund, third characteristic, is that they're very lightly regulated. So what does that mean? That means that a broker who sells something has to make sure that the client on behalf of whom he's buying, the broker, the stockbroker, has to make sure that the purchased item is appropriate for his client.

We have no such obligations. We can buy and sell with anyone we choose on behalf of our clients. But our clients have to be sophisticated investors. We have to vouch that they're sophisticated. If they have enough money, like say $5 million to invest in the hedge fund, they are by definition sophisticated. And then they don't need to be protected by having a broker who's necessarily watching out to see whether investments are appropriate.

We tell them what we're investing in and, of course, we're obliged to explain our strategy. But once they understand our strategy, we don't have to meet the same test that a simple broker meets on behalf of, let's say a poor retired individual. So there's less regulation and the fourth characteristic is hedge funds typically charge higher fees.

So that's what a hedge fund is. Now, the first hedge fund was started in the 1940s, I believe, by someone named Jones. And he was a stock picker. And what he did, is instead of picking let's say the best car company, which he might have thought was Ford and leaving it at that, he would try to hedge his risk. So he'd buy Ford, and he'd short all the other car companies.

So effectively, he wasn't just betting on Ford doing well. Because he could lose if the whole economy went badly, Ford and every other car company could go badly. Instead he was betting that Ford would do better than the other car companies. So he was concentrating his bet on something he understood more.

You can't be an expert about everything. Presumably he was an expert about cars. And he knew Ford was better than General Motors. So that was a bet he wanted to take. But he didn't want to take a bet on the whole economy doing well or badly. So that was the beginning of the idea of the hedged fund.

Now there's so many risks that a car company or any company runs. Does their president know what he's doing? Is Detroit going to be a good city? Is there going to change in government regulation? Is some foreign competitor going to appear on the scene? Is the price of oil suddenly going to change?

There's so many things they can go wrong in a company, it really is very hard to hedge. So hedging really makes much more sense when you can make the problem mathematical. Well see, with mortgages it really is a mathematical problem to a much greater extent. So it makes much more sense to hedge. And I think it makes much more sense to be a hedge fund if you're trading in the mortgage market. Anyway, that's what I found out when I visited Wall Street in those days.

So what did the hedge fund do? Instead of creating these CMOs, the hedge fund would buy them. And of course, buy the most complicated one. So in effect, the hedge funds would buy the most complicated, the residual piece we're talking about, and try to hedge that. So effectively, the hedge fund, by hedging it, was really carving out the cash flows of that last piece into complicated ways and selling them off to stabilize its profits.

So really, the hedge fund actually was continuing the work of the investment bank of creating more and more pieces and trying to allocate the risk. And so the hedge fund is part of the entire operation, which was making this mortgage market behind the scenes make home ownership so much easier and reduce people's risk. Or so it seemed.

We had a hard time raising money to begin with. Because after all, Kidder Peabody had just gone out of business. And not only had we gone out of business, but there was a general crisis at the time. It was not only the trading scandal of Joe Jett, but at the same time, there was a crisis in the derivatives market. Because all these complicated pieces that were being created where pieces went up and pieces went down, of course if you didn't know how to hedge your risk, you could end up losing a lot of money.

So Orange County went bankrupt in 1994. And what did Orange County do? They bought a bunch of inverse floaters. So when the interest rates were going down in the early '90s, they were making a huge amount of money. And the fellow who ran Orange County's municipal investments, was twice municipal investor of the year.

But in 1994, when everything turned around and interest rates skyrocketed for the year, his inverse floaters became almost worthless, and he bankrupted Orange County. So, it was some of our inverse floaters that were being bought, which helped bankrupt Orange County. Plus Kidder Peabody had just gone out of business because of Joe Jett. So it was very difficult for our hedge fund to raise money.

Our motto was we created the mess, let us clean it up. Nonetheless, we did raise some money. We had an important starting investment from the Ziff Brothers and we had an incredible boom time. We made fantastic returns our first few years. 50% returns. And we grew into the biggest mortgage hedge fund in the country. And things were booming along. And then suddenly, there was another crash in 1998. So this was the second crash that I was exposed to.

So, what happened in this crash? Well, as I told you, we were buying mortgages as the hedge fund, buying that residual piece, and buying it with borrowed money. So we would buy the piece, let's say for $100, by borrowing $80 dollars. So we'd use $20 dollars of investor capital, we'd borrow the other $80, we'd buy the piece for $100, and then we'd leave the piece with the lender as collateral. So if we didn't pay them back the $80, the guy could keep our piece. So his $100 piece was protecting his $80 dollar loan.

So things were swimming along. We're going to talk a lot about leverage later. Well, in 1998, one of the big competitors, Long Term Capital, which was founded, as I think I mentioned earlier in the course, by Meriwether, who was the most important, famous, fixed income trader on Wall Street from Salomon. And two Nobel Prize winners, who I've mentioned many times, and you're going to hear about again, Merton and Scholes, creators, especially Scholes, of the Black-Scholes model, the most important tool on Wall Street for managing risk, as we'll discover in a few lectures.

So these guys, the three of them, and nine other partners, 12 of them, created this wildly successful hedge fund. And in 1998, it suddenly went out of business. And in fact, the government had to step in with all the big investment banks, coordinating the big investment banks, to take it over so the whole market wouldn't crash down around it.

We celebrated the crash of Long Term Capital, figuring that was one of our big competitors out of business and now we would have an easier time. But, I remember thinking at the time that was a mistake. But anyway, a few months later, we got a margin call. So suddenly, on a Friday morning in October of '98, one of our lenders called us up and said, we think the prices have gone down, it's not $100 anymore, it's less than that. We need X million dollars of extra margin to put for us to contribute.

Because the $100 piece protecting the $80 loan, it was no longer $100 piece, it was a lower piece, so there was less cushion. We need extra cash to have a bigger cushion. So we said, oh this is crazy, the prices haven’t gone down that far, that's not fair. And they said, you have until 4 o'clock Friday afternoon to pay us back. And if you don't pay us back, then on next business day, which happened to be Tuesday morning, because it was Columbus Day holiday on Monday, we'll just sell off all your pieces and pay ourselves back out of the proceeds the $80 and give you whatever is left over.

Well, we figured what was left over wasn't going to be very much. Because they only needed their $80. So they didn't have an incentive to sell for the best possible price, although of course, they would tell us they'd sell for the best possible price. They wouldn't and it would be a fire sale. So we didn't know what to do. And we couldn't raise the money by 4 o'clock on Friday.

So we called up Warren Buffett. We said, well this is so unfair, they're forcing us to sell, there's no reason why we should have to sell. This margin call is not proper. It's not right. There's no reason we have to sell. What's going to happen is on Tuesday we're going to be forced to sell all of our bonds at the same time. And they're going to all sell for like $80 or something. It's going to be a terrible blood bath, a fire sale.

Prices will go for nothing. We'll be totally wiped out so unfairly. Why don't you buy our firm. We'll give you half our firm, some big percentage of our firm. You just make the margin call for us. And then you could have this firm with these great bonds. It's just a travesty. And it's going to be such a good investment for you. And you can save us and save the bonds and stop this travesty.

And he said excuse me? And we said, well it's unfair, they're going to force us to sell on Tuesday. And the bonds will go for nothing when they are perfectly valuable bonds. And they are going to wipe us out. You can prevent that from happening and own our great firm. And he said, hell I think, I'll wait for Tuesday and buy the bonds myself. And so Warren Buffett didn't rescue us. And so that Tuesday it looked like we would go out of business.

But over the weekend, we managed to hold an auction and sell the bonds ourselves. So in other classes in economics, you find out how to conduct auctions. I don't have time to describe this in great detail. But let me just say, that in a typical auction you have to worry about the winners curse. Everybody's thinking, I'd better not be the highest bidder, because that means I probably overpaid. Because those smart guys are bidding against me.

So what we did over the weekend was, we called everybody up and we said, we've been forced by a margin call to sell these things. There's going to be a great deal for you. So why don't you come back from your vacation, one guy came back from Budapest, one big trader on Wall Street. We got a huge collection of traders there over the weekend. We showed them our bonds so they could think about it.

And on Monday we held our auction. We didn't sell everything at the same time. At 12 o'clock, we sold the first third of our bonds. At 2 o'clock, the second third. And at 4 o'clock, the third third. Actually, we got a little behind schedule, but that was our plan.

And so what happened was that at 12 o'clock, everybody basically bid $80. So we said, we're one of the bidders ourselves, of course. We told everybody, they all did it by email, we emailed everybody back, you've been outbid. Either by someone else or by us. You've been outbid.

And all these guys, one from Budapest, from all over the world back there, thinking we're going out of business, bidding $80, they don't get to buy anything. So what would they do? What would they be thinking? So between 12:00 and 2:00, we're on the edge of our seats. Are they going to bid $80 again? Which means, of course, we get nothing. Because we just pay back the loan for $80 and we go out of business and our investors go out of business.

Or are they going to realize that they have to bid more in order to beat other people. So we had no idea what was going to happen. But the price was basically $95 a 2 o'clock and $99 at 4 o'clock. So we celebrated saving our firm. And it turned out there was another complication after that, which I won't get into. Which is that the thing had happened over the holidays. And so the next morning, the buyers didn't get their bonds.

Remember, we don't have the bonds to sell them. They're sitting with the person who lent us the money. So it was over a holiday, and the bonds didn't get transferred to the buyers. And so the prime broker who was supposed to vouch for all this, it was a holiday for him too.

And so the whole thing was a mess. And it's actually quite an amazing story, which I don't have time to tell. But anyway, when the buyers didn't get their bonds, the next morning on Tuesday, they all got alarmed that maybe we didn't even have the bonds. And so everybody made a margin call against us. And we went from celebrating to suddenly thinking we're out of business again.

But sometimes, when you're so badly off, people realize that they have to stop. And so there was a big conference call on Wednesday in which all of lenders, the sort of number two people at all the big investment banks got together in the conference call, and they agreed that if they all tried to get their money back at the same time, none of them would get it. And so they waited for us to gradually work ourselves out of the predicament.

Chapter 5. The Leverage Cycle and the Subprime Mortgage Market [00:52:52]

So we survived the crash of 1998. And then after that, we had another boom year. An incredible year in 1999. And things kept booming again for eight years or nine years until 2007. Before, I get to that, that experience was so seared in my mind, the second crisis, the 1998 crisis was so seared in my mind, that I wrote a paper called The Leverage Cycle.

In which I said, what basically happens is people borrow a lot of money and they're very leveraged. Then something bad happens in the economy. And the lenders suddenly reduce the leverage. And so the big leveraged buyers, they go out of business. The leverage goes way down. Lenders will ask for more collateral. And there's also the bad news. And those three things together completely crush the market.

And so lots of people are out of business and the whole thing's a mess. But then after things settle down, there's another boom. And then it's going to repeat itself over and over again. So that's The Leverage Cycle story I told in 2003. I wrote it right after the crisis of '98 and I said that this crisis, which in '98 seems to be so small, could be repeated on a much grander scale for the whole economy.

So we had these boom years, right after the crisis of '98. And then in 2007, 2008, there was another crash. This time on a grand scale. But it was exactly the same kind. And the last two lectures of this course are going to be about the most recent crisis. And so we're going to mathematically reexamine this entire story.

And again, after the 2008 crash, when we almost went out of business again, our hedge fund. In 2009, we had the best year we've ever had. Better than all those other years, even in '95. So three times in a row. There's a crash, a boom, a crash, a boom, a crash, a boom. We made back more money in 2009 than we lost in the crisis of 2007, 2008.

So the course has to end with an explanation of why these cycles happened over and over again. It can't be an accident. It can't be all my fault that I've been in three of them. There's got to be something systematic going on. And that what I've been writing about it. And how I'll end course. And by the way, as you'll see, I spent a lot of time talking to the federal reserve, and Bernanke, and Summers, and also with the ECB, the European Central Bank about the leverage cycle. Which I think is generally becoming recognized as a central problem.

But I want to end this lecture in the next 15 or so minutes, 20 minutes, by pointing out one change in the market. The next big change in the market. The next wave of securitization.

So, this is the sub prime market, which we haven't mentioned yet. And which I will now describe. OK, so what is the sub prime mortgage market. So I already pointed out that securitization was a great idea. Now it was applied, remember, to prime mortgages. You had to be a very reliable homeowner to get one of those Fannie or Freddie loans. And because those people were so reliable, although they didn't behave perfectly rationally, you had cost and alertness issues, they still were very predictable.

Well, a combination of things happened in the late '90s. Investors began to get the idea, this has worked out so well, and there's so many people who don't own homes in America because they're poor. Maybe we can extend home ownership to still more people. Of course, they're going to be riskier, because they're poorer, they have less resources. But we'll charge them a higher interest rate and that will compensate us for the extra risk.

The government also saw this as a great opportunity to help the poor. So this combination of investors getting the idea and the government wanting to sponsor it, led to the creation of a new market, the sub prime market, which is what I'm going to describe now. And I'm going to talk about what went wrong, or begin to talk about what went wrong, but I'm going to leave most of it for the last two classes.

So as I said, securitization is so important because, let me remind you, typically when a bank makes a loan, the bank finds out a lot about who it is lending to. But the bank is a bunch of managers and stuff. The shareholders of the bank are the people whose money is at risk. They aren't the ones looking at the loans.

So if you're a shareholder, you own stock in a bank, two years later the bank is going to make a loan to somebody, risking your money. And you've got no idea who the guy is that the bank is lending to. Or what the characteristics are of the person who is getting a loan. Think of the securitization when you're buying a pool. When you get this pool, you could be told what are the characteristics of all the homeowners in the pool. It's something very concrete. A bank lends to homeowners, it lends to businesses. It's doing 6,000 things.

Of course, it's risky to have bank capital. You know, equity, own shares of the bank. Because you don't know what they're doing. And they are doing so many different things. You should feel a lot more comfortable if where your money is going is very well delineated. That's the idea of a securitization. It's very well defined where the money is going.

So it makes people more willing to lend. Corporate debt, that's another way of lending money. But if you own bonds in a firm, you're not the first person to get paid back. The people lending with collateral get their money first. And if the firm goes bankrupt, there is a big court case and it takes a long time to get your money back. Not in a securitization.

So there are many ways that a securitization is more attractive than lending money through a bank. And that's the reason for the securitizations. Now here are the current numbers, the 2007 numbers. And there haven't been much securitizations since then. So the agencies are now up to $4 trillion, Fannie and Freddie.

Now those jumbo loans I told you about, the big loans that the rich people take out, remember, which was $0.5 trillion in 2003, is now $0.8 trillion. But there are two more markets that have been created. There's the Alt-A. Those are people who don't have credit ratings good enough to get the Fannie Freddie loans. That's $0.8 trillion. But they are pretty good credit ratings.

And then the sub prime population, of which there were $1 trillion loans issued by 2007. That's $1.8 trillion of lesser credit loans that didn't exist at all in 2002. So you add these numbers up, 4.8, 0.8, and 1, that's $6.6 and $4 trillion of unsecuritized the banks are just holding loans. So over $10 trillion of loans as I said.

OK, and then you add the commercials on top of that and you're getting very close to the value of the stock market. So all right so we've talked about this. The advantages of securitized loans.

All right, so how did the sub prime market began? Well, you had to have a few legal hurdles. The first one, which some countries haven't managed to achieve and some people now are doubting we should've ever done, is it legal to charge a high interest rate to someone just because they're riskier. Is that usury? Or is that a reasonable return, because you're making a bigger risk?

So, there are anti usury laws on the books. And so in order to make these loans, they had to become legalized. You had to get congress to pass a law to say that higher interest rate loans for riskier people is not usury. So then, you also had to figure out the tax treatment, which happened in 1986. And so the first pools were created in the late '80s, the early '90s, and Kidder Peabody, by the way, the firm I worked for at exactly that time, was one of the first creators of these sub prime like loans.

So it was very small in the early days. And it worked pretty well in the early days. And, as I said, by 2007 it had grown to $1 trillion, 5 million people at an average loan of $200,000. Now, these are people who, as I said, have bad credit ratings. And so they pay a higher interest rate. Instead of say 6% at the time, or 5% at the time, they might pay 5% or 6% of the mortgage rate at that time. They might pay 8% for the first two years. And then at the end of the first two years, the rate jumps up by another 3%. Maybe to LIBOR plus 6%. LIBOR, remember, is the inter bank interest rate.

So, it's already high, and then it jumps up after a couple of years. Now, many people have said this is the kind of predatory lending. That the homeowner is being lured into taking out a loan, not realizing that in a couple of years, the rate is going to go way up. But actually, there's some rationale to this. It turns out, that if you've made your payments for two or three consecutive years, the market thinks you're not such a subprime person anymore.

After all, many subprime borrower are people who are young. They didn't pay their credit cards in college. Or they defaulted on something. And once they're married, have a house, have kids, and they pay for three consecutive years, the market figures, well these people are much better. So we'll give them another loan, they can refinance into a loan with a lower interest rate. Because they are no longer considered subprime. Maybe they moved to Alt-A or maybe even into prime.

So you could count on 70%, by the end of the third year, of people refinancing their loan. During all those years from the 90s and the early 2000s. So you'd be left with 30% of the people who didn't refinance. Why didn't they refinance? Probably because there's something wrong with them, they were missing their loans [correction: payments]. They were much riskier than the original pool to begin with. So naturally, you're going to charge them a higher interest rate. So it's not such a crazy thing that the interest rate went up after two years. It isn't necessarily a sign of predatory lending.

OK, now how do these loans get started? There's somebody called an originator. A broker would find a homeowner, a subprime homeowner, and then go to originator who would create the deal. Now what would the deal be? The pool would be a bunch of subprime homeowners and you'd have to get some data on the people. You'd have to figure out do they have a job? What's their debt? What's their income? Things like that.

And all that would have to be reported to the potential buyers. So a buyer in the shares of these subprime pools would get a list. Not by name. You can't reveal the name of the homeowners. But what the zip code is of every house. And what the basic qualifications were of the homeowner. Do they have a job? What's the loan to value on the loan? What's their debt to income ratio? Stuff like that.

Now, the servicer, who's often the originator. Let's say the originator is an investment bank. The servicer might be the same investment bank or a specialty servicer. What they do is just what the banks did in the Fannie Freddie story. They're the ones sending the letters, telling people they have to pay, collecting the money, and dividing it up to the bondholders, as we'll describe in a second.

And they have one extra job. Because they're subprime loans, you can expect a bigger percentage of them to default. And when people default, sometimes they can't make the payments, they lose their job, something like that. The servicer is given the right to modify the loan. The servicer can say, OK I understand you can't pay, of course you can't pay, you're out of a job, you don't have any money. There's no point of us throwing you out of the house right away, maybe you'll get the job back.

So we'll work out a deal where we delay your payments, or maybe reduce the payments, maybe even we reduce the principle that you owe, because in the end that's going to make our bondholders more money. So the all the bondholders have no right to talk to the homeowner. They don't know even though the homeowner's name. But the servicer who is sending the homeowner letters all the time and getting the money, of course the servicer knows the name and has the right to modify the loan.

And if the homeowner doesn't pay and the loan is not modified, because the servicer doesn't think it's worth it, then the servicer can kick the homeowner out of his house. And take the house and sell it and pay the proceeds to the fund. But there's one extremely interesting provision that if the homeowner is not paying, during the time the homeowner doesn't pay, the servicer has to make the payments for the homeowner into the trust for the bond holders.

That's supposed to give the servicer an incentive to hurry up and figure out what to do to change the loan or throw the homeowner out. And of course, when the house is finally sold, the servicer can recoup his advance payments out of the proceeds of the sale. And then later, the rest of the money goes to the bondholders.

Now, the rating agencies played an important role in determining the ratings that I'm about to describe. So here's the thing would look. You have all the homeowners stuck into a pool, all those loans. I'm going to go for ten more minutes, by the way. You have all these loans stuck in a pool.

So the money would be coming in, the homeowners, remember, are paying this high interest rate because they are subprime borrowers. So there's $100 of loans that went out. Let's say each loans is for $1.00. Where does the originator get the money to lend to the homeowners?

Well, the originator is at the same time creating the bonds. So there'd be AA bonds, AA, A, BBB, with a bunch of minuses too, the over collateralization and the residual. So you'd create $81 of AAA bonds. These guys would pay that LIBOR, the inter bank rate, which let's say was 5% percent. Plus a tiny bit, 20 basis points, so 5.2%.

The AA would pay a little bit more than that. And the single a little bit more than. That's probably misprint, that should be LIBOR plus 10. LIBOR plus 20, LIBOR plus 30, and then the BBBs, LIBOR plus 130. So you'd issue $81 worth of triple AAAs, $7 of AAs, $5 of As, $5 of BBBs. That adds up to $88, $93, $98. You're only getting $98.

So you've got $98 worth of bonds that you used. So the originator has sold these bonds for $98. He's got to lend $100, so he's $2.00 short so far. But you notice that the interest rate all these bondholders are getting is 5% percent, or a little bit more. The homeowners are paying 8%. So there's extra money coming in that doesn't have to go to the bondholders.

So there's this residual piece that gets the right to get the extra interest payments. So the originator either holds residual piece or finds some hedge fund and says, OK, you buy the residual piece. And maybe the residual piece is worth $5.00. So $98 plus $5.00 is $103. So the originator has gotten $103. The broker's fees, remember the broker had to find the homeowners, gets $2.00. So now he's only got $101. So he lends the $100 and pockets $1.00. So the originator would get 1% bear no risk for his work in creating the deal.

Now, how can these AAA pieces be rated AAA? Now, to be rated AAA mean there's a one in a hundred chance, or something, that you're going to default. So how could that possibly be, when the loans are so risky? Well, what happens if somebody defaults? So this is the calculation. What happens if somebody defaults?

Well, if you haven't paid for the first 30 days, that happens quite often. But if you go, 60 days without paying, nothing bad happens to you. After 60 days, you get a letter saying you're delinquent and a note is being made that's going to have an impact on your credit score. So being 60 days delinquent, seriously delinquent, is bad for the homeowner.

After 90 days, you're considered likely to default. And so you get these very threatening letters from the servicer. And after 120 days, the servicer can start to try and throw you out of the house. But to throw you out of the house, maybe they have to go to court, they have to do a whole bunch of things. And so in those days, it took 18 months, 14 more months, after the four already, to throw somebody out. It now, has taken on average, a couple of years, or three years. It's getting more and more complicated to throw somebody out. Which we're going to get to in a minute.

So you could be thrown out after say 18 months. Now what happens if you're thrown out of the house. Let's say, the house is sold for $0.80. When you originally had a $1.00 loan. Well, during the time the guy hasn't paid for that year and a half, if it was an 8% coupon, that means a year and a half is 12%. So $0.12 the guy hasn't paid. So the servicers had to pay up the $0.12.

And then you have to hire a broker to resell the house. That costs six more cents. And the guy probably didn't pays taxes for the whole year and a half. So that's another three cents or so, that you've lost, depending on what the tax rate is. So you've lost $0.12 of servicer advances, $0.06 to the broker, that's $0.18 $0.03 cents for the taxes, that's $0.21. And the house only sold for $0.80 instead of 100. That's $0.41 cents you've already lost.

That's a sort good scenario. You've already lost 40%. So the bondholders know that with this scenario they're going to lose 40% about of their loan. Where does that 40% is $0.40 cents, because remember there's only $1.00 loan to that guy, where does that come out of?

Well, there was this over collateralization of $2.00. This is extra money pouring into the deal because everyone else is paying a higher coupon. So you take the money out of this extra cash flow that's coming in. If the default, the lost $0.40 is more than the extra cash flow coming in that year, then you reduce the over collateralization. There was $2.00 dollars. These bondholders were only owed $98 and there's $100 of loans outstanding. So you could lose $2.00 of loans and still have as much loans backing these bondholders.

But once you get to the next dollar, you're going to take the $0.40 first out of the BBB piece. So as more and more houses go under, you go through the over collateralization, the extra interest. Then you start taking things out of the BBB piece. After you wipe that out, then you go to the A piece, and then the AA and finally go to the AAA.

So how could you possibly think those pieces at the top were AAAs? Well because let's do a simple calculation. Actually, they look incredibly safe when you start to think about it. All this extra interest and the over collateralization and stuff like that it’s sort of 8% protection. And on top of that, you've got the lower pieces bearing the original losses.

OK, so even if you expected 40% percent of the homeowners to default, which is an astronomical figure. It was typically, in the past, less than a few percent. Even if you thought 40$ of the homeowners would be thrown out of their houses, and 40% of each homeowner was going to be lost in the scenario we just described of the 40% loss, 40% times 40% is only 16%.

But the AAA pieces are protected by 8% and then are protected by another 19%, because they're only the top 81%. So 16% doesn't come anywhere close to the AAA guys. So that, in 2007, was a horrible scenario to imagine. That 40% of the homeowners, they are 5 million homeowners. That means 2 million people tossed on to the streets, losing 40% of each of the homes and you don't come close to touching the AAAs. That's why it seemed like they should be AAAs and so many people were willing to buy them.

Chapter 6. The Credit Default Swap [01:13:51]

I’ve got one more thing to add to this background piece. The credit default swap. By the way, things are even rosier than that, at least looked at from the point of view of 2007 Because remember, 70% of the people were always prepaying. That means you got 70% of your money back for sure. So you only had 30% of the people left who could possibly default. So instead of having 40% default, you'd probably have 40% of the 30%. So you see, you could easily imagine that your AAAs were completely safe.

And so I think the credit rating agencies didn't do such a horrible thing rating these things as people say. But I haven't finished the story. So watch what happens. So a new invention that happened in the end of 2005 was the credit default swap. CDS they are called, which you've heard a lot. So what is a credit default swap? It's just insurance on each of these bonds.

So a credit default swap on the BBB would say, if there's a dollar of principal lost, because the homeowner's default, you take the loss out of the triple BBB, you can get insurance on the triple BBB. A CDS just is a promise to pay a dollar for every dollar that's lost in the triple BBB. And the CDS on the A is a promise to pay a dollar for every dollar that's lost on the A.

So that was a huge market. It suddenly took off in 2005. And then there was an index that was created in 2006. So these are insurance on particular deals. But if you put all the insurance together, you can make an index of what the insurance is like. And so you can tell how valuable the bonds are. Because if you see that the insurance premium is changing you know that people realize there's a bigger chance of default of the BBB.

So the insurance market, the index of it, is going to tell you a lot of information about what what's going on in the subprime world. OK, now, I don't have time to get to these legal issues. But I want to add the last fact. Things seemed to be going so well in the subprime world from the late 80s, early 90s, all through 2000, all through the middle 2000s that people got more ambitious still.

So what did they do? They created CDOs. So what are CDOs now? Remember, we had these loans that got cut up into bonds. There was insurance in the bonds. So people said, let's take the BBB bonds. They're sort of at the bottom, they're protected a little bit, but near the bottom. Let's cut those up into different pieces. So we take BBB bonds from all different deals, we put those into another pool, and now we cut those into different pieces.

And so you might have a pool that's from California, and a pool that's from Detroit, and a pool that's from Florida, that would just be a horrible combination, or a pool that's from Illinois and Ohio and stuff like that. You put them all together. And now, somehow the market decided that we could cut these into AAAs, AAs, and As with the same logic as before.

But, this turned out to be catastrophic. In fact, the market went one step further, and not only took the BBBs that we said before and made them collateral for more AAA bonds and CDOs. But then took the As from this, which came from the BBBs and cut those up into more AAAs, CDO squareds.

So that was the mortgage market, the subprime mortgage market. So you see, that there's a tremendous amount of synthetic stuff. By the way, these insurance pieces, if you're writing insurance, you're promising to pay basically what the BBBs pay, so it's like an artificial BBB. So they created the synthetic BBBs and used those to cut up too in the CDO market.

OK, so I'm going to stop now with one final word. So the subprime market of $1 trillion, plus the Alt-A market, became the new frontier of mortgages. And as I said, everything went swimmingly in the 1990s and the early 2000s. And then in 2007, there was a tremendous crash. So in January of 2007, February of 2007, the BBB insurance index plummeted from 100 to 75.

At that point, everyone declared, oh this is just the subprime market. It's small thing, don't worry about it. Bernanke said there's no problem. The stock market didn't blip at all. It continued on until October of 2007 when it hit its high. The world took notice of the subprime collapse, but everybody said it wouldn't amount to anything, because everybody underestimated the importance of the mortgage market.

Chapter 7. Conclusion [01:18:36]

So we're going to see in the last two lectures, how this unraveling of the subprime market led to the unraveling of the entire economy. And we're going to show that everything that I've described, although it sounds very much more complicated with subprimes than it did with the prime mortgages earlier, is really when you get down to it, the same story of leverage and crashes and then booms. And we're going to begin the next class by talking about the mathematics of the prime mortgage market and prepayments and how to value those. And gradually we're going to get to the crisis and what caused it and what we should do to prevent future crises. Thanks.

[end of transcript]