Credit Crisis Interview: Richard Herring on Mortgage-backed SecuritiesPublished: June 20, 2008 in Knowledge@Wharton
The home price run-up that preceded the credit crisis "is the sort of thing that happens time and again, bubble after bubble," says Wharton finance professor Richard Herring. He is one of seven Wharton professors interviewed by Knowledge@Wharton for this special report.
An edited transcript of the conversation follows.
Knowledge@Wharton: The subprime crisis seems to have been aggravated by the use of mortgage-backed securities that many people didn't understand and now are hard to value. What was different about these securities from the "Plain Jane" mortgage-backed securities that Fannie Mae and Freddie Mac have been writing for years?
Herring: You are absolutely right. The origination of the securitization market, which is probably one of the most important financial innovations in the post World War II era, was based on securities that were guaranteed by Fannie and Freddie, which had the implicit guarantee of the U.S. Government. So, the private sector didn't have to worry about the creditworthiness of these instruments, although they did have other tricky bits such as prepayment options and so forth.
Private securitization evolved as people found ways to substitute for the credit guarantees from Freddie and Fannie. And there were essentially three ways that people did this. One was reliance on credit rating agencies, which would examine each of the securities and issue them a letter grade with several notches around it indicating how safe they felt they would be.
The second approach was to model them very carefully using statistical models based on past repayment history to try to indicate the amount that you would need to set aside to make securities perfectly safe. And the third technique was credit risk tranching. You designated certain claims on a pool of securities as bearing the brunt of the risk. And if you had a sufficient amount of those securities, then you could make most of the rest of the securities very, very safe. This was an extremely popular innovation. It led to a profusion of perhaps $6 trillion or more in securitizations, many of these mortgage-backed, and it began to become more and more complex.
The residential mortgage-backed securities are fairly straightforward. It is relatively easy to see what the underlying [issues] are. There are some pretty close relationships between your claim on the pool and the pool itself. But there was a much bigger demand for investment grade securities than there are investment grade issuers in the world. So, the financial engineers began to innovate to find ways to synthesize investment grade securities by making combinations of these underlying mortgage pools and other sorts of securities.
One famous innovation was called "collateralized debt obligations" (CDO) and collateralized debt obligations would take pieces of the securitized pools, oftentimes lower pieces but sometimes upper pieces. They would add to them pieces from other pools and they would also tranche the CDO so that they designated with the assistance.
And it is a little bit ambiguous to say the role of the ratings agencies. It was something less than being consultants on how you do it, but it had the same effect because a potential issuer would essentially show a securitization to a ratings agency and ask what kind of rating do I get with this? If they didn't like the answer, they would go back and try another set. So, you were able to get pieces that would bear the first tranche and these would be below investment grade, but that meant that the rest of the pool had a very, very high quality and were investment grade ratings.
This could be complicated still more by doing CDOs of CDOs or CDO-squared or cubed in principal. And sometimes they were fed into structured investment vehicles or sometimes in asset-backed commercial paper conduits. And all three of these, and indeed the whole CDO market, have completely collapsed. And essentially the reason is that the three props that many people feel confident in making investments in these things have turned out to be not very reliable. The ratings have shown themselves to be completely unreliable for this whole category of structured finance.
We have seen multi-notch downgrades of many of these securities that are huge multiples of what one would expect, even in the worst times for corporate debt. So the implicit assumption -- and I would stress that it is an assumption because there were lots of reasons that people should have believed otherwise -- that an AA corporate bond was the same as an AA tranche of a securitization was simply false.
People also lost faith in the underlying statistical models because the most sophisticated players in the business, the people who were really making markets in this -- Bear Stearns, Citibank, UBS, Merrill Lynch -- had major losses and indeed had trouble figuring out even what their exposures were, so people completely lost faith in the valuation models.
And then finally, monoline insurance, the last safeguard became unreliable as people started to worry about the mounting losses ... that they had guaranteed relative to their equity base and their ratings came under pressure. Once they have lost the ratings, they have no value and that actually led to contagion to other kinds of products that are guaranteed a monoline insurer, such as muni (municipal) bonds.
Knowledge@Wharton: One of the problems, I gather, is that if you go back to the kinds of securities that Fannie and Freddie have been issuing for a long time, they have track records. They know how many homeowners default or fall behind on their payments. And so, they can project pretty accurately, based on history, how a new batch of these securities will behave. But some of these newer securities had no such track records and some of the models that were used to guess what the track records were, were just faulty. Is that right?
Herring: That is one of the problems behind the statistical model. There simply wasn't enough data over enough cycles to be able to predict. Another problem was that underwriting standards had deteriorated in an amazing way, which undermined the assumptions about diversification. Because one of the ways you can perform the financial alchemy of turning less than investment grade securities into investment grade securities is by diversifying out all of the idiosyncratic risk. Well it turned out that underwriting standards had deteriorated across the board almost all over the United States.
Knowledge@Wharton: Now, just to focus on the alchemy that you talk about and to try to put it into layman's terms, as much as possible. As I understand it, you have, say, a large pool of mortgages and altogether it has a certain level of risk, a kind of average risk for the whole thing. But then you could slice it up and say, well we are going to give certain investors first dibs on the payments made by homeowners. And they will be very low risk and they will get relatively low interest rates. And then other investors will have sort of last dibs. They will be the first ones to suffer if people fall behind in payments, but in return they will be rewarded with higher interest rates assuming they get what they are expecting. In layman's terms is that roughly it?
Herring: That's pretty much it, but there [were] some other bells and whistles attached to make it safer still. If you were not utterly persuaded that simply having a buffer of creditors who would suffer loss before you were, you might insist that you have a trigger clause in your contract, such that if there were cash flow problems, you could demand that the entire pool be liquidated and you'd be paid off.
Or if you were very, very conservative, you might say, "Gee, this is all good, but I would feel even better if I had monoline insurance guaranteeing my part of it." And most of the monoline insurance went to the AAA, which became known as super AAA tranches.
Knowledge@Wharton: Now some people have described the collapse of all this and the subprime mortgage market as a kind of perfect storm because a lot of these securities were predicated on the assumption that home prices would continue rising and that interest rates would stay low and that people's incomes would rise enough to cover future resets as their mortgage payments went up.
And, on the other hand, it looks like those three things, although they don't always happen at the same time, a reversal in those three things is not all that unpredictable. If you say, well interest rates were at extraordinarily low levels, you assume they have got to go up, and the same with home prices that were extraordinarily high. Was this an unpredictable perfect storm or were people just ignoring the risks that were apparent there?
Herring: It is the sort of thing that happens time and again, bubble after bubble. There was in no sense a perfect storm in this case. It was simply a return to more normal sorts of economic relationships. If you look at the growth in house prices in the United States, it has been relatively modest.
Until about 1998, you were making about as much on your house as on your treasury bill, as if you have invested in treasury bills. The appreciation started just about as interest rates went down, which was very fortuitous because it helped soften the collapse of aggregate spending in the recession. But it led to a circumstance where most of the people doing the business had not in their lifetime experienced a downturn in house prices.
And there are a number of psychological mechanisms that suggest that when people don't actually know the probabilities of things, they will tend to extrapolate recent experience. They will use something called the availability doctrine to see how easy it is to imagine something happening. Well, even if you do start with the notion that these could decline, if over time you have had more and more favorable experience, that tends to fall to a very, very low number. And if you believe that house prices are going to continue to go up, then you don't care so much about the ability of the borrower to repay because you figure you can always get the value out of the house.
Now, obviously, these are all, especially from the perspective of hindsight, utterly foolish assumptions, yet they are characteristic of virtually all the bubbles that we have experienced and there seemed to be bubbles at least once every 10 years or so.
Knowledge@Wharton: Well in this bubble, as in the dot-com bubble before and others, many of the participants have been punished pretty badly. Shareholders have lost a lot of money. Top executives at Wall Street firms have lost their jobs, although they don't seem to have wandered away impoverished. Is the marketplace punishing people enough to deter this kind of behavior in the future, or do you think something needs to be done on the regulatory area?
Herring: Well, that's a very complicated question. The market is to a very large extent self-regulating. The ratings agencies, [all three of them], are already very busy trying to figure out ways to improve their methodology and to make it more transparent. As you mentioned, several executives have lost their jobs, but not necessarily their bonuses and there are some real questions about compensation structure that I think probably will need to be reviewed.
I think, more fundamentally, we need to rethink the way in which we have redesigned bank regulation. Bank regulation under Basel II, the new view that has been worked on for over a decade, relies very heavily on credit ratings on the one hand, or for the more advanced banks, the internal models. And both of them have turned out to be not very reliable.
Beyond that, liquidity issues have turned out to be enormously difficult. Those are to be handled under Pillar II, but in a way yet to be determined. We are promised that the Basel committee is going to tell us something about that later on. Another issue that has surfaced that we thought we had put to bed with Enron is the use of special purpose vehicles or off-balance sheet entities. These special purpose vehicles did in fact comply with the post-Enron rules; they just became more complicated off-balance sheet vehicles.
But at the end of the day, reputation risk caused most of the sponsors, particularly the large-bank sponsors to bail them out, yet neither the regulators nor the shareholders in general even knew they existed. So, we have in fact created an off-balance sheet banking system that is huge, very nearly as large as the on-balance sheet banking system.
Knowledge@Wharton: One of the proposals for dealing with that is still forming, I think, but the general idea is to require that the ... institutions that create the off-balance sheet entities keep some skin in the game, as people have said. They continue to own the kinds of securities that they are trying to sell to other people so that they share some of the risk. Do you think that is a good idea?
Herring: I think that it would improve the incentive structure for sure, although we have to remember that a number of the institutions that have suffered the biggest losses maybe unknowingly had a lot of skin in the game as well: Merrill Lynch, Citibank, UBS and Bear Stearns were all big producers of these securities yet themselves got caught up in it.
Knowledge@Wharton: What is your general assessment of the way the Federal Reserve has responded to this?
Herring: They have been very innovative. They have taken the view that the first thing to do is ensure that there is sufficient liquidity to enable markets to function well. There have been real problems in doing that because it is more than a liquidity problem. It is potentially a very serious capital problem.
And it is a capital problem not only because of the losses that key institutions have suffered -- because of the new business they will not do, but [still] have an inventory and ... loans they claim to have been securitized -- but also because of the potential that they may have to bring a lot of this off-balance sheet business back on to their balance sheets. And to do that will require raising enormous new amounts of capital. So, they have been very reluctant, I think, to lend to each other long term because they think they may well have very much better uses for that capital in their own book of business.
One thing relative to an earlier crisis that I applaud and think has worked very well is the prompt corrective action solutions that we have put in place in the United States [that] have encouraged banks to recapitalize very, very quickly. Sovereign wealth funds -- to some extent private equity [and] to some extent shareholders -- have been recapitalizing these banks in massive amounts, which will enable them to continue and to give them some time to restructure and find new and more profitable business models.
But contrast that with what happened in the 1980s when basically the hypothesis the Fed was using is that the best way to recapitalize the system is to ignore the losses and let banks earn enough in other lines of business, so that they can retain earnings to replace the capital that we all know is already lost.
Knowledge@Wharton: That gets us to the final question, which is do you think the recovery will be faster than it was in the 1980s and are we closer to the end now than the beginning?
Herring: I think [it] ... depends. ... What's different about this of course is that it is in the housing sector and much depends on how long it takes to work down the huge inventory of houses. That is really very substantial. And there isn't a lot that the Fed can do about that. There is a genuine excess supply of houses out there that will have to somehow be worked through. Now this isn't entirely bad news. If you are a young household trying to find a home, there may well be good bargains ahead. And by lots of measures, the house prices we were seeing generally over the last three or four years were simply unaffordable to middle Americans.
So, there is a sense in which this is a transfer cost problem, but there will be winners as well as the losers we all read about in the newspaper.
Knowledge@Wharton: I'm sure that it will be unfolding for quite some time. Thank you very much.
Herring: My pleasure.