A Root of the Financial Crisis
Posted on: September 18, 2008
(This article is a transcript of an interview with Yale School of Management Finance Faculty, which took place on July 16, 2008. It will appear in Q4 magazine, to be published in October 2008, which tackles the question, ďWhatís the new capital up to?Ē Learn how to request a copy of the print edition of Q4, and visit the Q3 website.)
A series of legislative moves in the 1980s made subprime lending possible. By 2006, the subprime market had grown to 20% of the total U.S. mortgage market, and 75% of these loans were securitized and sold off to investors around the world, facilitating an influx of capital. With credit easily available, more people than ever before were able to buy homes ó but then the market seized up. Three scholars who have studied securitization analyze what went wrong and what it means for the future.
Frank J. Fabozzi, Professor in the Practice of Finance and Becton Fellow, Yale School of Management
William N. Goetzmann í86, Edwin J. Beinecke Professor of Finance and Management Studies and Director of the International Center for Finance, Yale School of Management
Gary B. Gorton, Professor of Finance, Yale School of Management
Q: Mortgage securitization has gotten a lot of attention lately, but how long has it been around? Frank Fabozzi:
Mortgage-backed securitization goes back about 40 years. The first Ginnie Mae securitization was 1970. When the first one was finally paid off, they had a big party down at Ginnie Mae. I was invited to that; thatís how I know itís been a long time. Will Goetzmann:
There was an earlier period of securitization of real estate loans in the 1920s, and they were commercial loans. A lot of buildings in Manhattan were financed by bonds issued to the public that were backed by mortgages for these buildings.
The history of securitization more broadly goes back much further. The early securitizations, where portfolios of bonds were put together into a trust and then securities were issued against them, appear to have started in the 1700s in the Netherlands, as a way to bundle up securities that individually might be risky but collectively created a stable, reliable payoff. Some of those risky bonds were obligations of the early United States of America. Fabozzi:
What weíre talking about today is whatís called asset securitization. If you look at securitization in general, itís really the use of public markets and avoiding the use of intermediaries. For example, common stock is an example of securitization, where instead of having private equity holders, youíre using public markets.
In asset securitization, youíre using assets as collateral and youíre not looking at the credit risk of the issuer; youíre looking directly at the credit risk of the underlying pool of loans.
Q: What role has this kind of securitization played in the development of the current credit crisis? Fabozzi:
Here, I think I have a better perspective because I think Iím older than Gary and Will. I went to get my first mortgage in about 1974. And I could not get that mortgage until my local S&L had enough money to be able to provide me with funds.
Can you imagine telling people who want to buy homes today that, even though they have enough for the down payment and they also have good credit, they have to only rely on their local S&L to get funds? They have to wait until funds were available? Do you think that would be an acceptable market to America?
The question then becomes who was holding the mortgages? They were primarily held by S&Ls. And what did that lead to? Gary Gorton:
The S&L crisis. Fabozzi:
The S&L crisis. S&Ls, as well as banks, were basically lending long-term on a fixed-rate basis ó because the government insisted that we have only fixed-rate mortgages at the time ó and then borrowing short-term. Great market ó until you started hitting double-digit interest rates.
Who in the world was going to hold all the debt of the U.S. mortgage market ó the largest debt market in the world? If weíd wanted S&Ls to continue holding it, weíd have had a bigger crisis much sooner. So, we created this mortgage product.
What were the benefits of the mortgage securities market, holding aside what happened recently? You could say we had the best housing finance market in the world. Securitization, generically, has been a very successful product, both in the mortgage market and in non-mortgage assets. I mean, General Motors would have much more difficult problems if it wasnít for the fact that GM can securitize its auto loans. We have governments that are able to raise money that way.
What caused the problems with this market? It moved from a market in which the underwriting standards were very good to one where, in my opinion, credit standards dropped greatly. As a result, we have a lot of people who failed to meet their obligations.
It was also overloaded. Remember, if I buy a $100,000 house and I put down $20,000, Iím leveraged five-to-one on the down payment. Now, a hedge fund goes out and buys that product on a leveraged basis, by borrowing money so they can get another five- or ten-to-one leverage on that. Look at the extreme leverage that creates.
The model that was used previously was, basically, create a loan, hold it in the institution, meaning a bank or an S&Lís portfolio. For example, my local S&L knew everything about the property I wanted to purchase. It was in their area. The manager responsible for providing the loan, the loan officer, would know about the underlying property. The manager would know trends in the community, and if the S&L put that loan in its portfolio and there were problems, the loan officer was accountable.
Now the concern with securitization in general is that issuers will reduce credit standards and, as a result, push the poor credit risk onto the market. Gorton:
I donít think ó I donít agree with this, at all. Fabozzi:
You mean you donít agree that that was the view? Gorton:
I think thatís the dominant view out in the world; I just think itís wrong.
If we ask, was the systemic financial crisis caused by a lowering of credit standards? I think the answer is no, thatís not the problem. Not only is there no econometric evidence that supports that view, but it has a couple of other problems.
One problem is that you have a whole variety of asset classes that have been securitized, and those markets are 20, 25 years old: credit card receivables, automobile loans, tobacco settlements ó there are about 30 asset classes. And somehow, according to this view, in one of these asset classes, for some unknown reason, the credit standards went down. That just doesnít intuitively make any sense, in addition to not having any particular evidence.
My view of what is happening is that this particular asset class, which is subprime mortgages, is a new product. So, 20 years ago, the four of us would be having a discussion like this, and the subject would be why banks donít lend to poor people, the riskier borrowers.
The problem for banks and mortgage originators was and is that poor people are, in fact, riskier. And theyíre riskier partly because their income is just more volatile but also because they have other attributes that are riskier. They donít necessarily have income they can document, for example. They may be illegal aliens. So the question was how can you possibly lend to them?
And the short answer is you donít lend to them for 30 years because theyíre just way too risky. But what you do is you take advantage of one thing that we know pretty well, which is that for most people, including probably the four of us, a large part of our net worth is tied up in our home equity. So, if you can create home equity for these people over a short horizon, then they may change their incentives and work harder and be more likely to repay.
So, a bank is willing to say, ďIíll take a risk that for two years, if you can build up some home equity over that period, then weíll give you another loan for two years. And maybe eventually weíll give you a long-term loan.Ē Thatís what some of the loans that reset after two years essentially did. Now building up equity requires house prices to rise, so this product is extremely sensitive to house prices. And not only is the mortgage extremely sensitive to house prices but so is the actual structure of the securitization.
Securitized products are very complicated. Frankís written a lot about this. For subprime securitization, the structure is very different than other securitized products, and it inherits the sensitivity to house prices. And so what happened was that there was a very long, complicated set of structures ó not only securitization products but other structured products, CDOs [collateralized debt obligations], SIVs [structured investment vehicles], all sorts of products ó and this sensitivity to house price risk was spread through all these products in a way which became very opaque. Everybody saw that it was a problem when we finally had a market open that revealed it, which was a synthetic index related to subprime, the ABX index. And then the whole house of cards came tumbling down because we saw that house prices werenít going up, and that was affecting the value of these products.
This is a very special case of securitization. But the wider world, including all the bank regulators and central bankers, have this view that the problem is all of securitization. Hopefully they wonít destroy all of securitization. But I think itís very important to stress that this is a rather unique set of circumstances. Fabozzi:
Let me go back and address two points. First is what I was saying about the issue of holding versus transmitting risk to the marketplace. That was intended to describe what the general view has been as a criticism of securitization. It wasnít my position, obviously, because Iím positive toward securitization.
The second point is that itís true that you have a situation where youíre basically making a bet on real estate prices when you attempt to build up equity for someone who has a blemished credit record or insufficient funds to put down a suitable down payment. But certainly there were other issues as follows: If you look at the worst mortgage-backed deal that was done, they packaged together mortgages that were second-lien mortgages on properties that basically had loan-to-value [LTV] ratios of 100%. How could something like that pass muster ó Gorton:
Wait a minute, wait a minute. That in itself doesnít make it a bad deal. If you look at 100% cumulative LTVs with second liens, and youíre only securitizing the second lien, youíll see that the structure of the securitized bonds are a lot different than if youíre securitizing first lien. Fabozzi:
But it doesnít make any difference. Itís still a bad deal. The structure was rated AAA at the time, but that bond went under probably within one year. Gorton:
But, Frank, I donít see what thatís evidence of. Thereís a lot of stuff around the fringes here. Did some guys just neglect their lending standards? Probably. Were some brokers getting kickbacks? Yeah, probably. Were some people put into mortgages they shouldnít have gotten into? Yeah, probably. But if youíre going to redesign regulation for financial institutions, whatís the central issue?
At the end of the day, if you ask what caused this crisis, I donít think the evidence is there that there was a decline in lending standards. I donít think the evidence is there that the incentives were not aligned all along this chain. I think you had a product, which was designed to address this particular clientele. It happened to be, unfortunately, linked to house prices ó and it almost had to be.
The problem is that the strength of that link was not common knowledge in the market until the ABX index crashed four times in a row. Everything else, like beating up on the rating agencies, is just not the central issue. Goetzmann:
This observation that the housing price is the fundamental variable that underlies a lot of the turmoil is an interesting one. I think that we have in the United States a fairly short history of home price data. Itís only been in the last, say, 15 years, through the efforts of people like Bob Shiller and Karl Case, who created these repeat-sales housing indices, that we really felt like we could put the forecast of housing prices into models so that we could have some estimate of recovery in the event of a default on a mortgage.
In some sense, I think that the statistical foundations for including the home prices into valuation models are a driver of the current set of events. What we got was fabulous information about how housing has trended, since about 1980. But it really only shows a couple of dips and rises historically. You can take these housing indices ó and Iíve done this myself ó and use them to make forecasts. And the econometrician feels pretty comfortable about using these forecasts because they have a lot of inertia in them. So if you go back to, letís say, 2004, you put in the various housing indices for the various cities, what you see is that they were all forecasting increasing rises in home prices. And the confidence levels of those forecasts were pretty high.
I donít find it surprising that lenders would be willing to make a mortgage to people with low credit scores, because, if they were going to get the house back in three years, the models would be telling them that the house would be worth something. Fabozzi:
But when you securitize, in the securitization, you shouldnít be looking for appreciation in the collateralís value. Gorton:
Thatís why this is different. Fabozzi:
Lending on an asset which youíre relying on appreciating is a mistake. Even when you look at CDO deals, for example, you see very few market-value CDO deals ó Gorton:
They all died. Fabozzi:
Shiller called the housing crash two years before it happened. But we didnít have data, and the rating agencies didnít have sufficient data, to be able to properly assess what the risks were. And a point that Iíve made many times is before you decide you want to securitize a new asset class, and this was a new asset class, you should have a good history. Gorton:
But you never do when itís an innovation.
One of the things that happened was the house prices went up in zip codes that had a lot of subprime mortgages, right? You can get the zip codes of the mortgages and go look at the house price changes in those zip codes. And it looks like those zip codes had latent demand for mortgages, which appeared to cause the house prices to go up.
But, more generally, with a lot of innovation, you just donít have the data that we would like as academics, but that doesnít stop the market from creating products. I mean, you canít wait. You try to make money. Fabozzi:
Thatís true. But this was extreme. I was involved in the first mortgage securitizations in Denmark. This was an innovation there. Before the first deal came out there was a tremendous amount of work that was done looking at historical databases.
In subprime, I think there was insufficient data to be able to generate a market of this size. And whether you attribute it to just housing prices alone, or housing prices combined with inadequate data and lower credit standardsÖ Goetzmann:
Well, I think itís a little subtler than that. What happened was, for the first time, we had these historical indices, city by city and in many cases zip code by zip code, that provided some comfort to an underwriter that they might not have had before. And that led to a willingness to extend credit into areas and to people who very well might not be able to pay off their loans.
The big problem in all of this is that 20, 30 years worth of housing data, under lending conditions that were unlike the ones weíve now created, just doesnít give you any information about what will happen in a crisis. So the crash that weíre seeing in housing prices, we donít see that in all those historical indexes.
And, as Gary pointed out, the availability of credit itself affected housing prices. So, as economists weíd like to say ceteris paribus, hold everything else equal, but the ability for the first time of poorer people to buy houses in their neighborhood actually pushed prices up. And then lenders cheerfully forecasted a continued increase in prices.
And so you could say it was a virtuous or an evil spiral, whichever way you look at it. It meant a lot of poor people owned houses they couldnít own before, but it also led to conditions under which eventually the price level would drop. Fabozzi:
Now remember, there were also new mortgage designs that were permitting this. The fact that you allowed borrowers options as to whether to pay ó that was something we had no information about.
I think we have views that are similar on why the crisis occurred. I put a lot more emphasis on credit, and Gary puts more emphasis on housing prices. Is that fair? Gorton:
Q: Youíve focused on the subprime market. Why have the problems spread outside of that area? Gorton:
One thing to keep in mind is that subprimes were not just securitized once. So, you take the mortgages, you put them in a portfolio, you create securities out of that portfolio, and you sell the securities. But when you sell them, you donít necessarily just sell them all directly to investors. Many of them were sold into other structured products. And then those structured products were, in turn, tied to derivatives. And/or those other structured products ended up in the portfolios of still other structured vehicles.
You have this very, very complicated chain of the movement of the risk, which made it very opaque about where the risk finally resided. And it ended up residing in many places. So the whole infrastructure of the financial market became kind of infected, because nobody knew exactly where the risk was. So the monoline insurers had some of it, and big banks had some of it. And monoline insurers happened to back municipal bonds, and so it spread there.
Then because banks were forced to write down billions of dollars, there were fears about liquidity in the market, so everybody stopped lending and nobody would do repo. The effects of this were magnified, and it was already a pretty big number to start with, because the amount of subprime in 2006 and 2007 was $1.2 trillion. You take $1.2 trillion and you donít know where that is, itís easy to have a whole financial system be at risk. Fabozzi:
If you recall newspaper articles coming out at that time, we just got pieces of information. At one time, the report was the losses to financial institutions would be x. Then all of a sudden it increased by a multiple of that. Did you see the latest numbers, Gary? Gorton:
You canít even keep up with it. Fabozzi:
The critical thing is that the press has only recently started to distinguish between losses and write-downs. So most of the stuff is not really losses, itís write-downs. A write-down is an accounting event. The accounting profession has bought into the finance professionís idea of efficient markets. So if the price of something goes down, the idea is that the value of it actually went down. And thatís a problem when youíre in a panic or a crisis like weíre in today where there are no prices. And to the extent that there are prices, theyíre fire-sale prices.
So youíre writing things down under GAAP accounting using prices that we know are not exactly what they should be. And the problem is that GAAP accounting has real effects because the measured capital ratios for regulated financial institutions go down, and then those institutions have to raise capital at very punitive rates.
At first they got sovereign wealth funds to invest, but now theyíre having to sell things. And the obvious thing to sell is all these bonds that were marked down. So, now you sell them and youíve crystallized the loss, which happened because of the write-down. And so the whole thing is a sort of vicious circle where you canít distinguish between whatís really a reflection of a future realized loss that weíre predicting with these market prices, and to what extent the market prices are just a reflection of the fact that thereís no liquidity.
The way the accounting works made everything much, much worse. Fabozzi:
Itíd be interesting to get a response from the accountants who read this.
Q: Given the problems with data around the subprime market and how unclear the prospects were for some of these products, who was actually buying them? Fabozzi:
The driver was highly leveraged investors who saw attractive spreads available on the product. Weíre not talking about the AAA-rated pieces but some of the real risky pieces offering very wide spreads. When Wall Street saw that there were entities such as hedge funds and CDOs that were willing to buy this product, they found a great desire to create it.
The hedge funds and CDO managers were able to buy them on a leveraged basis, and thatís why who got stuck holding them at the end, to our surprise, was the banks. Banks were taking them in as collateral in repo transactions, and when collateral calls were requested they didnít have the ability to dispose of them.
If the end result was that the investor group holding the product was hedge funds, I see no problem with that. Thatís what hedge funds are in business to do: take on great risk, and with great risk come potentially greater losses. Goetzmann:
I would like to go back to the theory of securitization for a moment, because it helps you understand what was envisioned by this new market. The notion of securitization is that you take a lot of assets that may not have liquidity, and you create liquidity by bundling them and parsing them. By parsing them, you create securities that have characteristics that some people would like to own. For example, you take a pool of mortgages, and you split them up so that one part is a very safe set of payments and another part is very risky. In a world where everybodyís different, and some people have a greater taste for risk, there is a price at which somebody, a speculator or just somebody with a tolerance for risk, is willing to hold the risky part. You kind of match the risk characteristics of the assets to the risk attitudes of the purchasers. Thereby you have an efficient allocation of risk across the economy. The notion is that once youíve done that, that risk is broadly diffused throughout the economy. Not that it disappears, but itís residing with those entities that feel comfortable in holding that risk.
So, what went wrong, or how come that vision didnít play itself out? How come we didnít simply have a case where a bunch of hedge funds that wanted to hold risky product got hurt?
What we didnít quite envision in the theory is the extent to which holding all of these different pieces of paper had to be done through the medium of financial institutions of some sort. Intermediaries like, letís say, Bear Stearns were fundamental to the process of holding and trading these securities. And in some sense, the institutional framework was the weakest link in this chain.
You might have had hedge funds that said, ďWe can accept this risk.Ē What you didnít have was a signoff by the financial system that said, ďIf something goes wrong in one part of this, we can wall it off from the other part of the system.Ē
So, thatís what we had. We had a propagation of the consequences of a relatively small part of the housing market, that is, the extreme tranches of these subprime mortgages, impact the financial system because of the leverage extended by intermediaries, like the investment banks. And once you start shaking the foundations of the institutions, everything rattles. Thatís something that may not have been envisioned by the pure theory of securitization. Fabozzi:
Will makes a good point. The theory of securitization would say that the risk should be out of the banking system, into the hands of a broader base. But the exact opposite happened.
Q: What steps might decrease the panic in the market now? Gorton:
I think Bernankeís been doing a fabulous job. I think heís really the only one whoís done anything. But the problem is that I donít think weíre anywhere near the end of this. One thing you have to remember is that you have millions and millions of mortgages now that have nowhere to go to be refinanced. They probably canít be refinanced. Many of them are going to fail.
The reality of the situation is pretty bleak because things that couldíve been done werenít done. Now with a lame duck president, things are not really going to be done. Some new presidentís going to come in and wonít do anything for six months. By then itíll really be too late. So weíre going to muddle our way through this, and itís going to be extremely painful.
Q: Will, do you see historical parallels between current events and any previous financial crises? Goetzmann:
The most extreme example might be the Bubble Act of 1720. In England there was a proliferation of corporations that followed the financial revolution of 1688, which actually was a ďglorious revolution.Ē And there was a period of wonderful financial activity that could have started the Industrial Revolution a century earlier. But Parliament decided that there was too much speculation, and they basically suspended the ability of companies to register as limited liability firms in 1720. And that set the financial world back a century.
I think weíve seen hints of this with things like Sarbanes-Oxley where politicians feel the need to do something. And what results is a regulatory framework that is not developed organically by the markets but sort of a heavy-handed response that doesnít set things right. Weíre seeing this loom on the horizon. Actually, itís not on the horizon; itís on the tables of Congress. Every congressman who wants to brag that he did something about the crisis has some new constraint or new regulatory entity ó limits to short selling or God only knows what.
Iím concerned that there is going to be a permanent, or at least a decades-long, setback in the ability of the financial markets to create wealth and to finance innovation and to help people at lower income levels. Thatís the scary part about this crisis. We can weather things like bailouts of Fannie and Freddie, and hedge funds losing a lot of money, and even large banks going under. But what we really have to be concerned with is well-intentioned but poorly designed solutions.
Q: Frank, do you see any potential solutions arising from the market, as Will said heíd like to see? Fabozzi:
It may be too early. We see more and more talk about using a different structure called covered bonds. Itís been around in Europe. Gorton:
I think thatís probably a good idea, Frank. Fabozzi:
I agree. But theyíre still working it out. Gorton:
I agree with everything both of you just said. Thereís a danger that government will overreact, but there are going to be millions of Americans who get chucked out of their houses. And thereís a huge social cost associated with that. Itís just massive.
Youíre going to have boarded-up neighborhoods and people arenít necessarily going to have places to go. That is something that ought to be addressed. Itís very difficult for the government to do anything about that because the natural entities to do that would be Freddie and Fannie. But you canít in the current political environment, given their difficulties, expand their criteria. It would be perceived as bailing out Wall Street. Goetzmann:
The role of government at this point is to figure out how to extend a helping hand to reduce the grave social effects of what weíre confronting. That would be a more productive thing for Congress to be focusing on, as opposed to creating new oversight and new rules. Interview conducted and edited by Jonathan T.F. Weisberg.