<< Back
The World According to Steve Ross

Steve Ross is a leading academic actively involved in putting his ideas into practice. He currently serves as Franco Modigliani Professor of Finance and Economics at MIT and has published more than 75 articles in economics and finance. He is best known as the inventor of arbitrage pricing theory and as the cocreator of risk-neutral pricing and the binomial model for pricing derivatives.

Ross is also cochairman of Roll and Ross Asset Management, a $3 billion quantititive financial management firm that specializes in applying arbitrage pricing theory. He has served as a consultant to a number of investment banks, including JP Morgan, Salomon Brothers and Goldman Sachs, as well as to leading corporations and government agencies. He serves as a member of the boards of directors of TIAA/CREF and General Re.

Ross spoke with editor Joe Kolman in August.

Derivatives Strategy: You've spent a lot of time thinking about the relationship between the different ways the derivatives market and insurance market approach risk. What limitations do people in the derivatives industry have?

Steve Ross: People in the derivatives business tend to be too concerned with questions of hedging. They're always asking themselves, “Is there some market in which I can hedge this out, and perhaps eventually offset it with a countertrade?” But not enough attention is paid to combining that with a kind of insurance function.

Risk can be dealt with in two sorts of ways: If you have a certain exposure to a movement in the Deutsche mark, you hedge it out with some foreign currency position. But what do you do with positions where you can't do that? Suppose you have some exposure to California earthquakes in some of these new catastrophe bonds and suppose there's no market for really hedging that out. The other way to get rid of a risk is to diversify and, in effect, take it as an insurance bet, like an insurance company does.

People in the derivatives business tend to be too concerned with questions of hedging. Not enough attention is paid to combining that with a kind of insurance function.

Not enough of the derivatives shops understand that both of these risk-mitigation techniques are going on simultaneously. In the future, the opportunities for shifting risk and lowering risk on the customer side will require that the derivatives business become sophisticated about hedging what it can and retaining on its books what it can't. Or perhaps trying to lay off portfolios of independent insurance-type risks to a large holder.

When an insurance company writes an auto policy, it doesn't go out and try to hedge that. It's too small and idiosyncratic. The company thinks that if it has a million of those policies, the results will be statistically determinate, even though any one of them isn't.

DS: I guess the problem is that people still think about risk fairly narrowly. They categorize it as currency exposure or fixed-income exposure.

SR: Yes, and I don't think those are insensible ways to do it. The problem is that you have two sets of sophisticated people—underwriters and financial engineers—thinking about risk differently. The financial engineer is saying, “Yes, I have some reserving to do, but the bulk of what I'm trying to do is hedge out the risk.” Meanwhile, the underwriter is saying, “The bulk of what I do is reserve for risks, but I have some hedging to do.” The insurer may do some hedging, but mostly he's talking about diversification, about underwriting the risk, which is to say putting aside reserves for the unlikely event that, statistically, 10,000 independent risks will come out on the bad side.

DS: It seems that although both groups may use the same word—diversification—they come from two different intellectual capitals, as it were, and they might as well be like the Americans and the British, separated by a common language.

SR: I think that overstates the case. I think they really are starting to grow to understand each other. The training of people on the financial engineering side is in partial differential equations and binomial trees. And the training of people on the actuarial side is in probability distributions and loss probabilities

DS: Where do you think the next synthesis might occur?

SR: The markets are pushing a synthesis. Many derivatives transactions are becoming increasingly transparent, increasingly commoditized. As that trend continues, I think the shops will find themselves pushed more and more to accept basis risk. And the ones that find that most easy to do are going to be the ones with the highest credit ratings and the biggest pools of capital behind them.

Derivatives shops in general tend to be rather narrowly focused on the collection of activities they do, in part because they have such core intellectual strengths in certain areas. If I'm correct in my assessment, then the future belongs to people who can assess different kinds of risk and take on basis risk.

DS: What kinds of products are we talking about here?

SR: I can imagine a company saying, “We'd like to hedge out some of our income-stream risk.” The income stream may be weakly correlated at the company level with major stock market indices, but there's nothing you can really do about hedging it. If you put a portfolio of these types of risks together, however, you would have something that would look a lot more determinate and lot more “hedgeable.”

There is a view on the part of some folks that somehow or other the insurance people are unsophisticated. That view really is misguided. Insurance is a subtle business.

Or maybe a company would come to you and say, “We want a hedge against changes in our average wage or our cost of doing business.” Then you would look at the company's cost of doing business and discover that it's somewhat correlated with the Consumer Price Index. So maybe you could do something with the new index bonds as a way of hedging against it, but it may not be perfectly correlated. All this is a simple way of saying that as you seek out more interesting opportunities, you'll find your basis risk rising.

DS: Are you talking about moving derivatives shops more into the mergers and acquisitions arena?

SR: Absolutely. You often hear of people at the corporate level justifying the demand for an acquisition in terms of smoothing out an income stream: “We have an earnings stream that has a big seasonal component in it, and it would be nice to acquire a company that is counter-seasonal.”

But the price of doing those M&A transactions is often the trouble associated with trying to blend two rather dissimilar organizations, and it's not clear that you get much value out of that. If one of the major motivating forces is to smooth out an income stream, why not go over to the derivatives shop and ask it to do those kind of things for you?

Here's another example. I always thought it was remarkable that a company would go out and say, “I think we ought to be in an electronics business.” Then it goes and actually searches for an electronics company to acquire. It has some strategic reasons, but ultimately the real reason for the acquisition is financial.

At the same time, 40 percent of the company's assets are sitting in a pension plan. If that's what a company really wants to do, why not just deploy some of those assets into the computer sector? It's a heck of a lot cheaper then getting up the learning curve on the computer business. That kind of integration within a company and those alternatives aren't really well-thought-out. And derivatives shops have wonderful opportunities to provide investment banking functions.

DS: So it's really a failure of vision, more than anything else.

SR: I'm always trying to figure out where the business is going to be in the future. In the short run, you say, “Who's knocking on our door for what products? What do they want? What are our sales people bringing in? What kind of customer flow do we see?”

In the medium term, you ask yourself, “Where is the industry going?” The trend is clearly toward the commoditization of what used to be profitable products. So you have to look at the nontraditional products to make money. That might mean going further out in time. It might mean pushing the limits of your analytics. Or it might mean looking for more esoteric kinds of deals.

Then there's the long-term view, which not enough shops have the luxury of taking. You ask yourself, “What are the real risks in the world that need to be insured or hedged against that we're not really taking care of now?”

There is an enormous movement in this country toward individually defined contribution plans—401ks. I think there's going to be an awful lot of retail-type products developed for that market. We have a puffy market today and everyone wants the best of all worlds. People want a piece of the upside and they want insurance on the downside. That means all kinds of option-constructed contracts. In France, for example, protected index products are quite common. It seems unusual that they're not popular here. I don't understand why that's the case.

I'll give you another example. A lot of people have a substantial percentage of their net worth tied up in the value of their house. People buy insurance against fire but not against the value of their house going down. There really ought to be a way to hedge the decline of the value of a home. Now you won't be able to buy insurance against the decline of the value of your individual house, because you could go out and torch it. But there's no reason you can't buy insurance against the decline in the value of a zip code area or something like that.

That would be an interesting product for a lot of reasons. And it may be exactly the kind of product that some of these megamergered institutions such as Salomon/Travelers would be interested in. They'll have an interesting advantage with their access to the retail market on the one hand, and their ability to perform sophisticated services on the other.

DS: A lot of investment banks are eager to get into the insurance market.

SR: Right now everyone is excited about the insurance market. They see huge pools of money and there is a view on the part of some folks that somehow or other the insurance people are unsophisticated. That view really is misguided. Insurance is a subtle business. It's possible to write a lot of bad business that seems quite profitable, because the losses come in after you write the business.

DS: Derivatives people understand that concept in their own way.

SR: Correct, but I think they'll have to be sophisticated about how to get into that business.

DS: What do you think are the most likely stumbles they'll make?

SR: I think they'll underestimate the importance of underwriting risk and of the underwriting function. They won't understand well enough what it really means to have a good handle on how insurance companies look at risk. That's a toned-up understanding that comes from centuries of examination.

DS: It doesn't come from the Black-Scholes-Merton model.

SR: Black-Scholes-Merton is not useless for those things, but it's not the be-all and end-all. In the end, someone has to sit down and assess what the actual probability of something happening is, and that may well be influenced by the insurance written on it. That's the more hazardous problem. The probability of something happening may also be influenced by an adverse selection issue. If you define some sort of a policy or business you're going to be in, it's not surprising that the people who show up at your door are those who think they will get the best advantage from doing that business. Those are things that insurance companies understand extremely well—and the things derivatives shops don't quite understand yet.

The solutions I see are clearly not going to be solutions that involve solely derivatives. I think there is going to be some sort of securitization there, for instance. There's no choice, because when you're talking about risk, you have to have some underwriting expertise, and presumably the underwriting expertise would have to sit somewhere between the shop that originates the insurance deal and the ultimate financial buyer. Having derivatives expertise in general won't be enough. You want someone with underwriting expertise as well.

DS: You're considered a leading academic, but you're also very much involved in what's happening on Wall Street. Wall Street owes a lot to academic research. What does it still have to learn?

SR: In a number of areas, the Street is ahead of the academics, which is quite a switch. In numerical analysis, for example, there are a number of super-fast numerical techniques that were first devised to make calculations on the rate of development of nuclear processes. These are now being used on Wall Street to compute the value of a fillet piece cut off from a mortgage IO.

I think Wall Street doesn't quite understand how the proper use of financial analytics involves estimation as well as modeling. Let me give you an example. The majority of interest rate modeling and stock option modeling is focused on estimating the parameters of quite elegant models from the current prices of securities. The parameters of term-structure models are estimated from the current yield curve and are designed to fit the current curve. Option models are estimated from the cross section of options of different strikes and maturities.

This all seems sensible, until one looks a bit closer. In effect, each day or week a new model is estimated from current security prices. But what happened to last week's prices? Are they worthless now? Is the current estimate of volatility so precise that we can forget all the previous estimates?

And what about consistency? Rarely is any effort made to tie together the estimates for the three-month spot rate distribution with the estimates for the two-month rate one month later. Surely past information is more useful than that, and this is where practice on Wall Street could be most improved. The marginal benefit from a slightly improved analytic model is dwarfed by the improvement from improved estimation.

Was this information valuable?
Subscribe to Derivatives Strategy by clicking here!