The 1998 Derivatives, Risk management & Technology Roundtable
|Ed Berko, head of the derivatives systems practice,
financial risk management group, PricewaterhouseCoopers
Leslie Rahl, principal and cofounder, Capital Market Risk Advisors
Till Guldimann, vice chairman, Infinity
Reto Tuffli, former managing director of risk control,
Union Bank of Switzerland
Joe Kolman, editor, Derivatives Strategy
Maureen Callahan, CEO, Callahan Co.
Alan Grody, president, Financial InterGroup
Peter Vinella, president, Peter Vinella & Associates International
David Phillip Quinn, assistant professor of finance,
Illinois Institute of Technology
Wendy Brewer, senior vice president, Bank of Tokyo Mitsubishi
William Margrabe, president, William Margrabe Group
Vincent Cirulli, vice president of market risk review, Citibank
Cris Conde, CEO and chairman, SunGard Trading Systems Group
Gena Ioffe, chief developer, Intermark
Mark Galant, vice chairman, FNX Ltd.
Joseph Rosen, managing director, Enterprise Technology Corp.
On June 22, more than 75 senior derivatives professionals met at a special roundtable discussion on technology and risk management, sponsored by PricewaterhouseCoopers and Derivatives Strategy. Topics included the relative merits of in-house and vendor solutions, the difficulties of integrating spreadsheets into global risk systems and new risk management technologies on the horizon.
Here are excerpts from one session, which analyzed the current state of enterprise-wide risk management systems. It was moderated by Ed Berko, head of derivatives systems practice in the financial risk management group at PricewaterhouseCoopers, and Joe Kolman, editor of Derivatives Strategy.
|“What is the current state of enterprise-wide risk management? Are we going to spend the next 10 years talking about this before it's worked out? Or is it here now?”
Ed Berko: The questions we're addressing today are, What is the current state of enterprise-wide risk management? Where are we? Are we going to spend the next 10 years talking about this before it's worked out? Or is it here now?
Leslie Rahl: I think the talk-to-action ratio is quite high. I think it's clearly on most firms' agendas and some people have made progress, but I don't think people would honestly tell you that they have the complete solution. The obstacles are enormous, not only in technology but in communications as well.
When you have markets that close at different times, you have to make a whole series of assumptions to get a single number or single group of numbers. I don't think anyone has really got his or her arms precisely around that problem. We have a long way to go. I hope it's not 20 years, but it's certainly three to seven.
Till Guldimann: I disagree. I think there's a huge misconception about what firm-wide risk management is all about. Firm-wide risk management is not about somebody at the top understanding precisely what the aggregate position is and therefore becoming a super-trader. Firm-wide risk management should allow you to make better judgments about the managers who run the risk positions. It should allow you to decide whether particular traders are risk-return efficient or not, and to allocate limits efficiently and properly to traders who have been trading more profitably on a risk-adjusted basis.
It is a totally different perspective from what risk management at the trading level is all about. With this in mind, I think you have to ask yourself, “What kinds of risks can we compare consistently in an enterprise-wide framework?”
Clearly, we cannot aggregate credit and operational and market risk together across the enterprise. But I think we've come a long way toward standardizing our concepts of market trading risks. If an enterprise is heavily involved in market risks from trading, you may be able to say that a significant part of the trading risks have been consistently measured, and as a consequence, the capital resources have been allocated properly. But do not confuse systems that measure risk or help a trader take risks with enterprise risk systems, which have a totally different purpose.
Reto Tuffli: Another way of putting it is that enterprise risk systems provide a mechanism that allows you to allocate finite resources better—whether it's balance sheet or risk capital, such as market value-at-risk, or credit risk capital.
But when you focus on limits, as Till has mentioned, you're really getting into the whole realm of risk control, which is an important objective of enterprise risk. When people wring their hands and say, “We're not there yet,” it's invariably because the whole mechanism and infrastructure of data flow from front-office systems to enterprise-wide systems is not airtight. The complexity of the legacy system infrastructure then presents a number of day-to-day production issues. These are the issues people are still struggling with in many institutions.
Joe Kolman: The difference of opinion seems to fall into two opposing camps. One group of people seems to think that the risk management numbers coming out of systems are either worthless or not worth a hell of a lot. And another group seems to think that the number that comes out doesn't have to be of the same quality as the more refined risk management numbers you'd need on a trading desk.
Rahl: I think that's probably a fair assessment of what I've heard. But I think the issues are a little bit different.
Although I agree with Reto and Till that enterprise-wide risk management needs a different level of precision, I think people forget how sensitive those enterprise-wide numbers are to the assumptions that are being made. You can get a markedly different picture of what your enterprise-wide risk is, because you are making so many simplifying assumptions. I've actually seen one portfolio in which, if you changed the assumptions, you could change the valuation of the portfolio from a net long to a net short position.
|“I think the talk-to-action ratio is quite high. I hope it's not 20 years, but it's certainly three to seven.”
Kolman: Does that mean the data are worthless?
Rahl: No, it means that few people, including those at the management level, really understand what the numbers are based on or the key assumptions that are critical to the answers they're coming up with. Senior management rarely vets those assumptions. There are a number of banks that lost a lot of money in the European currency crisis in 1992 because some of their young quants assumed that the pound always had to stay in the snake.
Maureen Callahan: I think risk management works best on a personal level. At the exceedingly well-run firms on Wall Street, the guy who runs debt markets or the guy who runs equity has seven direct reports and he wants to know exactly whether he's long or short and what his equivalent risk number is. He doesn't worry about fat tails or standard deviations or any of the fancy stuff. What he wants to know is, If something blows up in the middle of the night, what does his trader have to do? I don't know if you can replicate that sort of personal good management at firms that aren't so well-run, because it's a problem of good people and there are only so many of them around.
Alan Grody: I agree. I was once asked to look at risk management at one of the old partnerships on the Street, before the banks started to take them over. Capital was allocated by the guys who said, “This is what's going to happen this week so give me some more capital to trade.” There's nothing that replaces the personal wisdom of a partner who is trading out of his own pocket and is going back to his desk and giving the commands.
|“It doesn't matter what the total number is for the institution. What matters is whether you can judge the individual departments against one another. “
— TILL GULDIMANN
Guldimann: The idea is that it would all be solved if we could go back to our own little club of partners. But the fact is, we have large organizations, they take large risks and somebody's got to manage them. We have to get the benefits out of the large systems we've set up. It is difficult to get some reasonable numbers together to make judgments. But we have to do it.
Peter Vinella: I think one of the problems with the enterprise risk management approach is that it doesn't model the way investment banks really work. Aside from a few exceptions, such as Merrill Lynch, banks have always decentralized the way they manage risk—usually by product—with little specific risk management input from the executive management committee. Investment banks lend funds to a department as long as the department seems to live within certain general rules and expectations, rather than within specific, concrete risk limits. In general, the executive management leaves these details to the department's manager, holding him or her personally responsible for the performance of the department. In most cases, senior executives are not prepared to make decisions based on product-specific risk factors. I believe this is a reasonable approach. I wouldn't want the CEOs of some of these companies coming in and making decisions about specific risk limits or things at that level. I don't think you want to be limited by the unfamiliarity or inexperience of the senior manager with a specific investment product. So the idea of enterprise risk doesn't map to the business realities of today.
Guldimann: But that's precisely the point. It doesn't matter what the total number is for the institution. What matters is whether you can judge the individual departments against one another. When you decentralize management, you need some sort of measure of how well the individual departments are doing.
Tuffli: Clearly, value-at-risk in and of itself has limited utility. It has to be part of a whole risk management practice that includes having people walking around talking to traders and really understanding the subtleties of risks that VAR is not going to capture.
|“Traders will end up saying, ‘The guy in Singapore has a VAR limit that's three times as big as mine and he's getting a lower risk-adjusted return.' This is exactly the kind of healthy competition you want to see in your organization.”
But at the same time, as a complement, you want to have enterprise risk systems that allow you to quantify risks, to be able to compare apples with oranges as best as possible. Traders will end up saying things like, “I'm making a certain amount of money with $1 million of VAR, while the guy in Singapore has a VAR limit that's three times as big as mine and he's getting a lower risk-adjusted return.” This is exactly the kind of healthy competition for finite resources that you want to see in your organization. Also, if there are questions about whether the numbers are right or wrong, then the inspection of that process and how it's calculated is a healthy inspection.
Vinella: Another problem is that most banks are forced to be in a lot of different businesses that don't make money. As a result, risk-adjusted return doesn't make sense at the enterprise level. I've been in situations in which one British bank was losing £4 million a year in a particular business. It had to be in that business in order to keep its corporate customers in the more profitable areas elsewhere in the bank. Most of the people who run businesses know that they are going to lose money here and they're going to make money there. I don't believe people think in terms of risk-adjusted return right now, and it will be a while before they do.
David Phillip Quinn: I think a real value of enterprise-wide risk analysis is the comparison of assumptions across different parts of the business. I know a bank that was going through a process of trying to rationalize the decision-making process on different desks in the fixed-income area. People were using a lot of models and had a lot of economic assumptions, and the bank found they were actually using contradictory assumptions. I assume that they could have improved their profitability by resolving those contradictions.
If you could simply force everybody to itemize all of their assumptions and their models and then cross-check them, you could probably get a lot of the benefit you're trying to get with more complicated measures that are less accurate.
Kolman: That's sort of a lower-tech way to get to the same place.
Quinn: Yes, and you won't get lost in all the technology issues.
Wendy Brewer: Until recently, capital was not a scarce resource if you worked for a Japanese bank. Now it is, so these issues are extremely important to us.
I don't think we've reached any consensus at all on risk-adjusted return on capital, which is the profit and loss divided by economic capital. If you want to start using this methodology to try to allocate capital toward businesses that are making money, and then try to use that number in the pricing of your swaps and the bonuses of your traders, and then calculate the full overhead into your businesses...Well, I tend to agree with Leslie. We are three to seven years away.
William Margrabe: I'm an economist and I've always thought in terms of striving for economic efficiency. But I don't think any risk management system I've seen in operation has aimed at that in any way consistent with financial theory. If you thought in terms of economic efficiency, you would want to know the actual capital of every desk, its systematic Beta risk relative to the market and its expected rate of return. I don't know any theory of economic efficiency that leads to value-at-risk.
Vincent Cirulli: But wouldn't the risk management systems enhance the value of the company if the systems were used to allocate capital more efficiently within the firm?
Margrabe: If they were, yes. But it's difficult to use them that way. In principle, that's what you want to do.
Cris Conde: I think Vinny has hit the nail on the head. I think the essence of all of this is to allocate capital and therefore maximize shareholder return. One of the megatrends of the last 10 years has been the decentralization of risk-taking. It used to be the Ace Greenbergs who made all the calls. Now the main job of the head trader is to allocate trading limits among the various trading departments and subdepartments. And the role of systems people like myself is to come up with consistent frameworks for measuring risk and profits, and measuring risk-adjusted profits consistently across the enterprise.
The opposite megatrend is happening among corporate treasuries, where more and more of the trading power is being consolidated into fewer and fewer hands. There, the systems tasks are extremely different from those faced by banks. In corporate treasuries, you look to consolidate all of the financial exposure to the corporation, looking at integrating future cash flows, contingent liabilities, cash management and so on into the trading activities, in order to end up with a consolidated value-at-risk and net interest income component.
|“It used to be the Ace Greenbergs who made all the calls. Now the main job of the head trader is to allocate trading limits among the various
trading departments and subdepartments.”
Gena Ioffe: I think one problem we face is that enterprise-wide global risk management systems are misleading people. They're frequently oversold, and that's kind of dangerous because they aren't that accurate—at least at this stage. If you have a few swaps, it may be quite possible to get a good sense of your risk by simply purchasing a few hundred dollars worth of spreadsheet add-ins.
I had an interesting conversation with some people in one of the Asian emerging market countries. They had a limited number of trades, some local Treasury bills and foreign exchange deals, and they didn't even have a local equity market. At some point they asked if we could recommend a good global risk management system because a bank next door to them had just purchased a system from a well-known vendor. I personally think that's a bit dangerous, because they'd be getting something that would be the opposite of what they could use.
Mark Galant: The RiskMetrics marketing group really did spin this thing well. Now everyone's thinking, “This is the Holy Grail number that they have to have at 4:15 p.m.” But everyone's chasing this one number without doing the adequate stress testing that you need to come to the realization of what your true risk is. “The bank to the left of me gets it at 8:00 a.m. the next morning, but the bank next door gets it at 4:15 p.m. so I need it at 3:45 p.m.!”
Berko: So let's come back to our original question. How many enterprise-wide systems are actually up and running now? How many of them actually perform true enterprise-wide risk?
|“If you want to start using RAROC to try to allocate capital toward businesses that are making money, and then try to use that number in the pricing of your swaps and the bonuses of your traders...Well, we are three to seven years away.”
Joseph Rosen: I don't think it exists in the way people have been defining it here today, in terms of risk-adjusted return on capital. The exception would be a few smaller hedge funds and other organizations that started eight or 10 years ago. Some of them do have all this information at their fingertips.
Guldimann: I think we are fairly far ahead in terms of market risk for trading books, but we are nowhere near there in terms of credit risk for trading books or credit risk at large. We are trying to take the technology we've developed on the market-risk side and apply it to the credit side, and we've made good progress. With the increased integration of front-to-back offices, we're learning how to capture execution data, which will help us get a grip on operational risks.
The goal is not to try to find the Holy Grail. It's to try to find incremental solutions for improvements that help you allocate your capital in the firm better. When it comes to things like operational risk, we still have no clue. We haven't really figured out how to quantify that at all. We're far away from the Holy Grail of the globally integrated risk management system. But we're making good progress in individual areas.
|Was this information valuable?|
Subscribe to Derivatives Strategy by clicking here!