The role that derivatives and financial innovation played in destabilizing the banking sector, or the financial sector as you would put it, what were you seeing? As this was going on, what were you thinking?
Well, there's nothing wrong with financial innovation and derivatives and so on if they are properly done, properly understood, and, I guess most importantly, properly rated by outside parties who are supposedly objective and you can rely on them.
And I think one of the most important reasons that this crisis got as big as it did is that these, the exotic forms of these particular instruments, incomprehensibly some of which were rated AAA by the rating agencies -- I cannot imagine how they could come up with subprime mortgages AAA-rated.
And therefore, particularly when it's AAA-rated, people don't do due diligence. I mean, they say that they're off by a little bit, so then they're only AA. (Laughs.) It's still -- they just bought this stuff.
And I'll make this statement, that if the rating agencies had not rated so much of this, what I would call toxic subprime mortgages, AAA, this crisis would have been small enough to be manageable and would never have gotten to the size it did, if just the rating agencies had done their job.
The rating agencies were paid by the banks, would have lost business had they not awarded those --
Not really, because there's only three of them. So the instruments either would have been issued at such a small volume it would never have made a difference, or now they made more revenue by rating these OK, but they wouldn't have lost anything, because there was nothing to lose, because there was nothing there.
That's the incentive to rate them highly.
Well, supposedly not. They're supposed to be independent.
Supposed to be.
But they weren't.
That's correct. Now, again, I'm not trying to make excuses for the management who originated this stuff, but we have in this country checks and balances. And if people want to know why this crisis got as big as it did, one of the major culprits, and there are some more, were the rating agencies rating subprime mortgages AAA, which to me was totally incomprehensible.
So what I was saying was bubbles occurred. In fact, we said it in our annual reports, that the spreads on all risk classes were way too narrow, as if there was no risk in the system. And we saw risk all over the place. And there was hardly no [sic] distinction between what we would say were high-risk activities and low-risk; they were all way overpriced.
In fact, we didn't even participate in the exotic subprime side of the mortgage, because we knew that this was absolutely wrong for our customers if we would have done it, and it would have been wrong for us because we think this thing is going to blow up.
How does the business work? I mean, why do a CDO of a CDO of a CDO?
What everyone is trying to create is something that has a high rating, and a high yield. That's the holy grail. That's the goal, is to mix together assets in some way so that you come out with an AAA, and a big return.
Now, the way that you do that is by looking out to the universe of financial products that you could buy -- bonds, mortgages, lots of different products -- and then using math to tell you how much of that deal could get the AAA rating. How much of it will be safe?
You're putting in stuff that's not safe, but the question is, of your product, of what you end up with, what percentage of that will be deemed safe, will be deemed AAA. …
And part of that dance has to do with putting on your little tie, and your little briefcase, walking over to Moody's or to Standard and Poor's, shaking a hand, and say, "Buddy, help me here," right?
Part of the dance certainly has to do with going to Moody's and Standard and Poor's and persuading them that they should give a high rating to as much of your deal as possible.
But you have to remember that when you're walking over to Standard and Poor's and Moody's, you're not walking over there by yourself. You're walking over there with a big check in your hand, because you're going to pay them for their rating. And their business thrived on structured finance ratings, on ratings of CDOs. These were very high-profit margin deals for S&P and Moody's. For a while, S&P and Moody's had the highest operating margins of any companies, period, in the United States, any large public companies. They were cash cows. And so when you walked over to them, it wasn't like you were having to persuade them to do something they didn't want to do. They were very interested in doing as many structured finance deals as they possibly could. …
But I don't understand that. If I go to a restaurant, and I think it's got three stars, I think it go three stars because someone decided that without bias. If I knew that the person that gave it three stars received a big check--
Standard and Poor's and Moody's ratings are not like the ratings that you would get from a restaurant guide, or a movie guide. They are ratings that have been embedded in the system since 1975. And because of a whole web of regulations, they're ratings that you have to get. They are a key that unlocks the financial markets.
I call it a regulatory license, sort of like you have to have a drivers license to drive, or a fishing license to fish. In order to sell bonds in the capital markets, you have to have a regulatory license from the credit ratings agencies. The AAA ratings is what unlocks the door to being able to sell to a pension fund or an insurance company or a bank, because there are so many rules that require you getting a rating. …
And the result of that is that the rating agencies don't have to be reliable or accurate in their jobs. … Think about what S&P and Moody's have done over the last several decades. They rated Orange County very high investment grade right before its bankruptcy. They rated Enron investment grade days before its bankruptcy. They gave Lehman Brothers, Citigroup, all of these financial institutions, AIG their very highest ratings right before they collapsed in the financial crisis.
They've been as wrong as we can possibly imagine a credit rater could be. If they were film raters, we would never go see films they rated highly anymore. And yet, we continue to use credit ratings. And the reason is because investors have to, because there are rules out there that require that they use ratings.
I should say here just as a footnote, the one great thing that the Dodd-Frank legislation did was to strip out -- or at least try to strip out -- that regulatory reliance on ratings. It told regulators, "These rules were a bad idea, and you've got to get rid of them."
And our whole regulatory apparatus is trying to do that right now. And my hope is that in the future, people won't rely on ratings as much as they have in the past. Although, with what you're seeing in Europe right now, people continue to rely on ratings, even though they're not accurate. …
Underlying the whole crisis in 2008 was the number of ... subprime mortgages. How did innovative financial instruments or whatever you want to call them -- credit default swaps, collateralized debt obligations -- what did they contribute to the problems that we faced?
... It used to be that when you wanted to get a mortgage you would go to your bank. The bank would lend you the money. It would make a judgment about whether you could repay, because it would know that if you couldn't repay it would bear the losses.
But then there was this idea called securitization that arose that said they would originate the mortgage but then sell it to someone else, and that other person would have to bear the losses. But the idea was you put a lot of mortgages together and the probability that a very large fraction of them would have a problem at the same time was very low.
Except the reasoning behind this was flawed, because if there was a bubble, prices went up, then they would all go down. They would all have a problem. If the economy went into recession, many people would have a hard time repaying their mortgages.
So the underlying assumption that large numbers would [not] simultaneously be affected was just wrong. ...
In fact it was insured by the fact that they were forever putting more money into the housing market.
[The] securitization process itself is what fed the bubble, which in fact made it inevitable almost that there would be this problem of a large fraction of them collapsing, going into default at the same time. So they created the problem that actually brought them down. …
You needed to have the investment banks that would put these together, ... the CDOs and complex products. Now if you had thousands of mortgages in a product, no one could inspect to see whether each mortgage was a good one. It was all based on trust. ... So you created a system in which incentives were such as to make sure that the system failed.
Then you had the rating agencies being part of ... I would almost say a conspiracy. The rating agencies would look at these bundles -- they obviously couldn't look at each of the mortgages -- and they would say if you put together large numbers of mortgages that ought to have been graded each F, by putting them together they blessed them as if it was financial alchemy that converted lead into gold. In this case, it converted F-rated subprime mortgages into an A-rated security.Why was that important? Because then you could sell this to a pension fund or to lots of other people who could only buy A-rated securities.
Then you had more money to send back to the originator to send back out and make more mortgages.
Exactly. And then make the price go high, and they'd look at this and say aren't we brilliant?
You thought you could do business effectively by holding less -- you know, economically, you could do business with less capital than what the regulators were requiring.
I think the idea was that we looked at the credit risk of the portfolio of loans we had, and it was very, very high quality compared to what the regulators considered an average portfolio. And we felt that the risk capital that we should have held was less than what the regulators might have said.
Now, again, in the '90s, the regulators had a very, very simple regulatory capital model. So there were lots of banks who looked at their own portfolio and said, "It's very different from what the regulators see as an average portfolio."
And these are not conversations that we considered secret. We certainly had direct conversations with the regulators: "Look at the credit quality of what we have in this portfolio. Look at the types of borrowers that we have." You know, we are holding a huge amount of regulatory capital against what we consider to be very low credit risk on average.
And you felt, therefore -- correct me if I'm wrong -- that you could effectively put more money to work for the bank.
Yeah, I think the concept was we wanted to be more efficient with our capital use. Now, just to put it into context, we as an institution were rated AAA in '94, and in '95 one of the decisions we made as a bank -- and I say "we" as if I were involved in these decisions. I was a new hire; I was very junior. But one of the decisions we made in -- that JPMorgan made in '95 was that they didn't need to be a AAA institution; that, from a competitive landscape, most of the other banks with whom they were competing weren't AAA institutions. And what that meant is that they held less capital against the risk that they had. They were slightly more risky institutions. And we, JPMorgan, must have needed -- you know, must have looked at the competitive landscape and said: "OK, we need to be more competitive. We need to have a better return on capital for our shareholders, and to do that we need to be more efficient with our usage of capital...."
How does that idea come forward? Are you excited? Is it a eureka kind of a moment, or is it incremental in its bits and pieces; you already had thought about it and it isn't a moment of invention at all?
People are always disappointed to hear it, but really this wasn't a eureka moment. It was, as you've suggested, very much an iterative process. It was over time, a coalition of people working together whose ideas were coalescing and they built upon each other.
The essence of the idea of a credit derivative in and of itself is intellectually compelling, and one can imagine instantly if you say, "Well, if I could buy credit insurance, then I could use it to free up risk, I could use it to improve returns on capital deployed, I could use it to create investment vehicles for investors." But unless you actually have a market in which all of these things can occur with reasonable liquidity and transparency, the idea in and of itself is not that significant.
What was significant was the work that began in the early '90s and continued throughout that decade to create a market for the ready transfer of credit risk in derivative form. And that took many years to evolve, and it was a function of many different things occurring cumulatively.
It was the initial ideas, of course, but it was then the discussion around those ideas, the introduction of those ideas to clients and customers and other competitors. It was creating documentation so that whenever you thought about doing a transaction, you didn't have to make up the terms of that transaction from scratch and imagine them and write them down. Rather, you could refer continuously to a template that other people were familiar with.
That took many years of negotiation and work and industry forums, working with ISDA, the International Swap Derivative Association, in order to standardize documentation. It meant working with regulators to establish, is this a permissible activity? And if so, how would that activity be accounted for from a regulatory point of view, from a capital requirement point of view?
It meant working with ratings agencies. If a rating agency is able to assign a rating to a company and say, "This is an A-rated company," then how would they create a credit derivative referencing that company? And all of these dialogues, many, many steps along the way, together cumulatively created eventually a market for the free transfer of credit risk.
Initially [it was a] transfer of credit risk referencing individual companies, so Ford Motor Company, or XYZ Corp., each individual transaction referencing one company. Over time, that too evolved so that just as in the equity markets you see index transactions, so you can buy a stock or you can buy an index, the S&P 500, for example, which is a basket of stocks.
In just the same way, in credit markets we saw indices referencing multiple companies at once evolve, so that now transactions could occur where, for example, you defined the investment grade corporate credit universe and transacted on several hundred names in that universe at once.
So the same tools that you saw in other markets began to be deployed in credit markets as part of a natural progression, and that really was the process of innovation. It was all of those small steps incrementally adding up to the creation of an entire marketplace, so it wasn't any longer just an idea in a room in Florida, it was transactions that had a reasonable likelihood of being able to be transacted in a transparent fashion on an arm's length basis between sophisticated counterparties.
I should emphasize, there was never any retail dimension to this whole market. It was always sophisticated counterparties, so banks and financial institutions and major corporations transacting amongst themselves rather than individuals on the street.
And indeed, that is one of the reasons why when subsequently, in later years, credit derivatives became entangled in the recent financial crisis, your man on the street was so shocked and surprised. There was this large market that he had no familiarity with.
Well, the reason for that is because individuals at a retail level, unlike in the equity markets, were never really involved in transacting in these transactions because by their very nature they were designed for large, sophisticated institutional activity rather than retail investment. …
What is a synthetic CDO, and how did it make this whole mess worse?
A CDO is a collateralized debt obligation. It's essentially a mixture of lots of different assets -- that's the collateral -- that are put into a trust, or a company, and then that company issues securities. So it's basically a way of mixing together some kind of an investment to create a new investment.
And the way that it's typically done with subprime mortgages is that there's an arranger who goes out and buys up a bunch of subprime mortgages, and then will go to a bank, go to a credit rating agency, and say: How much of this can we say is safe? How much of it can we say is pretty safe? What kinds of ratings can we get for this group? That's a collateralized debt obligation. It's basically a mixture of subprime mortgages.
The key to a collateralized debt obligation is creating different layers of risk, sort of like the layers in a building, like the floors in a building where the top floors will be the safest, and then as you move down, the floors will be increasingly risky. …
The difference between a CDO and a synthetic CDO is that what you actually put into the building or the trust or the corporation is not real. What you put into the vehicle is synthetic. What do I mean by synthetic? By synthetic, people mean that you're putting side bets based on whether someone will default into the mix, instead of putting the actual bonds into the mix.
So for example, if I wanted to create a CDO based on my mortgage, I would put the actual mortgage, the physical claim on the mortgage, into the investment. And you would look to my payments themselves. If I wanted to create a synthetic CDO, I would have a bank enter into a side bet with another institution based on whether they thought I would keep making my mortgage payments. … And then we would take that side bet, that synthetic investment, and we would have the CDO be based on that side bet. …
The genius of the synthetic CDO was that if you found a CDO that worked, if you found a bunch of subprime mortgages in Riverside County, Calif., that could be bundled and resold in a way that would be attractive to investors, the fact that someone had done it one time in a cash CDO wouldn't stop you from doing it again, and again, and again. In synthetic CDOs, all you had to do was make a side bet based on what would happen to this group of people and their mortgages, and then take that contract -- which would be a side bet -- and have that be the basis of the CDO.
Explain how that caused the contagion, that sort of web that you were describing earlier that made everybody late.
One of the things that happened with synthetic CDOs was that there were certain pools of subprime mortgages which were regarded as the most attractive to use in CDOs. These were mortgages that were really cheap. They were very risky. But because of the credit ratings agency's models, they were going to get very high ratings.
Wall Street went nuts over those kinds of subprime mortgages. They were hungry for them. They wanted to use them as much as possible to create AAA-rated investments that looked like they were safe, but that still had a high return and a high yield.
And so one of the things that happened was the CDOs that were created increasingly were based on these very risky assets. These were the kinds of subprime mortgages that if there were a 30 percent decline in housing prices, they were all going to default. It wasn't the natural mix of mortgages throughout the country. They were highly focused, isolated in areas like California, Nevada, Florida. They were focused and isolated on people with low credit risk. They were focused and isolated on people who hadn't put any money down to buy their house. And these mortgages -- which were again, it's the synthetic idea -- they were being bet on over, and over, and over again. These were the ones which increasingly were populating the synthetic CDOs.
So what you were doing was creating a bunch of synthetic CDOs, all of which were doomed to fail if there was a 30 percent decline in housing prices. They didn't have the kind of protection that would necessarily be there if you were limited by reality, if you were limited by the fact that you have to put a real mortgage in here, and then it's gone. They used the kind of magic of being able to bet multiple times on these riskiest mortgages to massively increase the amount of risk in the entire system. …
So we get to a point where there's a big party going on before it all goes down. Who's bought into this idea that these synthetic CDOs are going to be the best thing since sliced bread?
I think certainly by 2006, almost everyone on Wall Street had bought into the idea that CDOs, synthetic CDOs, and in particular what were called super senior tranches of synthetic CDOs, meaning the safest parts of synthetic CDOs, were basically risk free.
This was the way the most senior people at the major banks were thinking about them. This is the way they were being disclosed, or not disclosed. Everyone was assuming that these kinds of instruments basically had zero risk.
There was a group at Goldman Sachs who looked at the risk exposure of the bank to these kinds of instruments in December of 2006, and they saw that there was quite a lot of risk there. And they more than other banks understood the risk exposure and reduced their risk. But really, Wall Street was still smoking the CDO dope into 2007 when the crisis began.
There were, on the other side of the market, a group of people who had taken the opposite bet. And those people understood the risk very well, and they made a fortune. They bet that the subprime mortgage related instruments were going to collapse. They were the other side of this two-sided bet.
But one of the great ironies, I think, of the financial crisis was that the senior people at the Wall Street banks apparently didn't understand or capture the magnitude of their own financial institutions' exposure to these risks, which is really stunning, if you think about it, that the people who are in charge of these banks don't know what will happen when there's a 30 percent decline in housing prices.
If you think about it, if you're the director or the CEO of a bank, isn't that the one thing you should understand? What will happen to my institution if the following financial variable changes by 30 percent? That kind of worst-case scenario analysis is why you're being paid millions of dollars. That's precisely what these people should have been doing.
And yet, if you believe them in the most charitable version of the story, they had no idea that their banks would be destroyed if housing prices declined 30 percent. They simply weren't able to figure it out. It was too complicated. That's what they say.
What do you think the real reason is?
I think that it's very plausible that the senior people did not understand these instruments, and didn't understand the risk. If you take Citigroup for example, Citigroup did not disclose the extent of its exposure to these super senior positions until relatively late in 2007. There were a whole series of meetings within Citigroup, a kind of fire drill of sorts within the company. At the highest levels, people were trying to figure out how much money they had lost or might lose, what their exposure was. They couldn't even figure this out.
So I actually find it very plausible that the senior people simply did not grasp the extent of the exposure. … I think people would like to believe that the senior executives were criminals. And I certainly think that there was a fair amount of activity that could be labeled "criminal" on Wall Street during this period of time. But I think in some ways, even the more troubling aspect of all of this is that people at the top of financial institutions genuinely believed that these incredibly complicated, Frankenstein instruments did not carry any risk. …
I do think there were several instances of absolutely reprehensible conduct that should be punished criminally. And I'm very hopeful that some of that will happen ultimately. But I think one of the most puzzling aspects of all this is that finance may have gotten too complicated for anyone to understand. That the managers of these large financial institutions in some ways have been given an impossible task that they won't be able to comprehend what it is their institutions are doing. And that is really, really scary. …
When you were looking at steel companies and putting together deals, there was a transformation going on in finance. How aware of it were you, and what was your take on what you were witnessing?
We were aware of it in the sense not that we were investors, but it was clear this subprime thing was getting a little bit nutsy towards around 2004, 2005.
I gave a talk at one of these securitization conferences in Florida. It's the first time I've never had any questions after, because I advanced my theory that the subprime thing was going to blow up, and here were people who were pretty much buyers or packagers of the paper.
At the end of the speech I asked were there any questions. There was not a single question. They couldn't wait to get me out of the room.
... All the way back in the '80s and '90s there is the advent of a lot of technological, financial engineering. People were able to do things they couldn't do in the past. Was that on your radar? ...
I don't think financial engineering as such is the problem. I think it's become a misused term. What it's really become at the peak of the frenzy was a polite way of describing the packaging of securities that should never have been bought into something that was saleable. So I think it was an abuse of the engineering.
The idea of securitization itself is a perfectly sensible idea, taking a lot of little scraps of paper that don't have much liquidity individually, making them into a larger instrument that could be traded -- to me that makes sense. And there should be, logically, a rate arbitrage.
Where I think it went wrong was in overreliance on black boxes and on little mathematical models, because the problem with models are they inherently assume that tomorrow will look a lot like yesterday. The fact is that at turning points, particularly crisis turning points, what really happens is tomorrow turned out not to look at all like yesterday or like today.
So there's an inherent flaw in models that they are inherently based on what had been. They don't really protect you against the so-called black swan events.
How is it that we came to rely so heavily on models?
Again, I don't think there's anything necessarily wrong with it. I think that what is wrong with it is thinking the model is the only reality and not allowing for real stress testing and real contingencies.
Rating agencies, for example, never went and did original field research in subprime. If they had walked into a Countrywide loan production office, they would've understood that this had much more resemblance to a Wall Street boiler shop than it did to somebody who was trying to figure out whether or not they should make a mortgage loan. ...
Why were there no incentives for the rating agencies to do that? How did we get things wrong to allow things to develop as they did?
I think that the models and the model salesmen were much more sophisticated than the people in the rating agencies. ... The people putting these things together are much higher-powered people than the people trying to analyze them at the rating agency.
And who are these people that are putting these things together and who are much more sophisticated?
It's Wall Street people, quants [quantitative analysts]. …
"The FRONTLINE Interviews" tell the story of history in the making. Produced in collaboration with Duke University’s Rutherfurd Living History Program. Learn more...