Insurance industry risk managers who hoped to spend a few leisurely days at a Lake Tahoe client conference got a bit more than they expected in September. Risk Management Solutions, the Menlo Park, Calif.-based risk-modeling company, had astutely picked the dates of September 13–16 because that’s the likeliest period of the year for hurricanes to pound the East Coast—potentially a perfect opportunity to show off the company’s newfound ability to dispense real-time probabilistic loss figures.
On Sunday, September 12, when Floyd, a category 5 storm, began its approach toward Miami, some of the 330 attendees were dragging their cell phones to the golf course to check on the availability of flights home. Floyd could have been the once-in-a-decade event that the insurance industry had been waiting for, and home was where the models were.
RMS, however, was eager to keep attendees hanging around. The firm quickly established what it called the “war room” on the ground floor of the hotel, where computers linked to RMS’s Menlo Park servers performed detailed projections and analyses.
The room was staffed by an RMS team that collected real-time data on the storm’s location and other storm parameters from the National Hurricane Center and the U.S. Navy. The group worked late into the night and reconvened early the following day in order to provide attendees with regular updates. “The things I remember about the war room at 5:30 in the morning were the fatigue, the excitement, the anticipation of seeing what Floyd had been up to overnight,” says Michael Drayton, principal modeler at RMS, “and the cravings for coffee.”
Few clients were allowed to view the inside of the war room. Instead, clients could visit a nearby hurricane-tracking area replete with RMS analysts and scientists, live media coverage, access to web sites and answers to questions about the storm. Hard-copy notifications were handed out to all attendees twice a day, providing information on the probabilities of landfall by region and full probabilistic analyses of expected losses. And throughout the event, clients could learn how their losses changed as RMS issued new stochastic storm information.
As the hurricane approached the Florida coast on Monday, rumors about loss projections of $80 billion were circulating. A hurricane warning was in effect for the entire eastern coast of mainland Florida, and evacuations were underway. A couple of RMS’s Florida clients had even been called to return to help board up their homes and prepare to move out.
By dawn on Tuesday, RMS was generating the first probabilistic loss outcomes: There was a 30 percent probability of a Florida landfall, a 50 percent probability of a North or South Carolina landfall, and less than a 10 percent chance the storm would miss the United States altogether. Expected loss numbers reached $30 billion—almost twice as much as Hurricane Andrew—for a Miami landfall if it continued along its current track and central pressure.
By Wednesday morning, there were signs that a “recurvature” predicted in Tuesday’s simulations would take Floyd away from Florida. Sure enough, Floyd soon began to weaken and turn toward the Carolinas. New RMS strike probabilities were issued, indicating an 85 percent chance of landfall in the Carolinas. Expected loss estimates fell to between $2 billion and $6 billion—still a major insurance event.
On Thursday, Floyd finally made landfall as a weak category 3 storm, with maximum sustained winds of 110 mph—significantly weaker than when it passed through the Bahamas.
There was more than a hint of the macabre in the air. “I was trying to maintain a certain neutrality in simply telling the facts,” says Robert Muir-Wood, international technical director at RMS and leader of the company’s real-time risk modeling program. “But some people said they were surprised that I appeared to be giving the impression that a big loss is not a good thing. It was quite clear that a majority of the audience viewed a major loss, especially if it didn’t actually hit them, as being a good thing, because much of the audience participates in a risk market that is severely underpriced. If someone could have taken some pain and thereby driven the price up, they would all have been delighted.”
In fact, the event’s sporting undertones almost found expression in a bizarre way. “We thought at one point, just as the hurricane was turning up, that we should construct some kind of lottery based on people predicting the most likely landfall location and the most likely landfall pressure,” says Muir-Wood. “But we decided not to do it, because we thought it would introduce an air of frivolity to something that for a lot of people was a very serious matter.”
Floyd: A Blow-By-Blow Account
|Client Notification Date/Time
||Wind Speed (mph)
||Central Barometric Pressure (milibars)
||RMS Insured Loss Estimate (billions)
7 p.m. PDT
|400 miles SE of Palm Beach, Fla. (over San Salvador Island, Bahamas)
8 a.m. PDT
|255 miles E of Miami, Fla. (over Eleuthera Island, Bahamas)
5 p.m. PDT
|235 miles SE of Cape Canaveral, Fla. (over Abaco Islands, Bahamas)
8 a.m. PDT
|190 miles ESE of Jacksonville, Fla.
8 p.m. PDT
|215 miles SSW of Wilmington, N.C.
10 a.m. PDT
ISDA to Members: Chill Out Over Y2K Disputes
The International Swaps and Derivatives Association is recommending a three-day grace period for problems relating to Y2K, says Robert Pickel, general counsel for ISDA.
The Financial Markets Lawyers Group, a body connected with both ISDA and the New York Fed’s FX Committee, has been examining possible settlement problems as a result of Y2K-type events, focusing on possible disruptions in payment systems or clearing bank shutdowns. “ISDA’s view, by and large, is that if an individual counterparty has a Y2K problem, that’s its own problem,” says Pickel. “But in the payment system and clearing scenarios, ISDA’s view is that, as opposed to everybody rushing for the exits on the first day of the new year if one of these problems arises, it might be helpful if the industry’s practice is to wait a few days to see whether the problem can be solved before parties start trying to terminate contracts.”
ISDA’s recommendation: wait-and-see for three days or so, and then follow ISDA’s formal guidelines for unwinding transactions affected by Y2K problems, which the organization planned to release in late October. Pickel says the procedure relies largely on the mechanics of various industry documents such as those proffered by ISDA and the FX Committee.
“At the end of the day,” he says, “parties have rights under their contracts. If parties want to seek to enforce those rights when they want to enforce them, they are perfectly free to do so. It’s possible that an individual party faced with one of those situations may decide, ‘I don’t care what the “best practices” in the industry are—I have a contractual right to terminate once the party doesn’t pay, and I’m going to exercise that right.’ All we can say as an industry group joining with other industry groups is, It might be a good time to hold off on triggering defaults too quickly early in the year.”
To keep up with ISDA’s Y2K position, see www.isda.org.
Research and Trade Hedge Funds On-Line!
Historically, investors interested in plunking down big money on hedge funds have had to spend considerable amounts of time and money researching their prospective investment vehicles. Now, thanks to a new web site called HedgeWorld (www.hedgeworld.com), the hedge fund investment process has been made much simpler—and far cheaper—than ever before.
Besides offering a veritable treasure trove of hedge fund information, the site offers something never before ventured: an on-line hedge fund trading facility. HedgeWorld Ltd. and the Bermuda Stock Exchange Ltd. announced last month a strategic partnership to create an order-matching system to bring together buyers and sellers of non-U.S.-domiciled hedge funds. The facility is expected to be up and running by January 2000.
In addition to the novel electronic-trading functionality, the site has pulled off another major coup: It provides access to the Tremont Advisers TASS+ database, which tracks performance and other information for more than 2,200 hedge funds managed by some 1,500 fund managers. Basic summary performance information is free; detailed hedge fund profiles cost $50 apiece. If you think that’s a bit pricey, consider this: Before the TASS+ database was made available on-line, the minimum subscription fee was $10,000 per user; clients with more than one user over, say, a local area network had to pay more. By slashing prices, the site hopes to attract a whole new audience to the world of hedge funds.
“The concept,” say Gregory Barrett, executive vice president of HedgeWorld, “is that, by providing access to this kind of information to a broader audience, we can attract a critical mass of accredited investors and remove some of the mystery behind the industry, providing further transparency.”
In addition to the TASS+ database, the site offers bona fide editorial content, including hedge fund news and analysis. It also provides portfolio-building tools that allow users to set up and model their portfolios on-line, and a proprietary mail and alert facility that provides secure communications from fund managers.
In addition to the secure communications, the site caters to fund managers in another way: Fund managers can build customized web sites within HedgeWorld for any visitor to see.
“If you look at some of the other web sites that are out there,” boasts Barrett, “some of them simply act as a distribution mechanism to get out performance numbers. Others are covering the industry from a news standpoint. But nobody is doing as comprehensive a job as we are in providing performance information, content and industry coverage, and providing resources that members of the investing community can use to transact business more effectively on a global scale.”
The site launched on September 16 and, according to Barrett, HedgeWorld is “very happy with the response.” Thus far, he says, institutional investors have been the main users, but after preaching to the converted, Barrett expects HedgeWorld to target high-net-worth investors with an elaborate media campaign.
JP Morgan and E&Y Release Op Risk Tools
JP Morgan, the first bank to promote a proprietary methodology to measure market risk, is trying to secure a pole position in the operational risk arena. The firm has recently released an operational risk management system, distributed by Ernst & Young, that runs over a customer’s own corporate intranet.
Many companies that have tried to monitor and control operational risk use a standard internal auditing process known as a control self-assessment—a bulky, paper-based method for determining risk profiles on a global scale. Horizon promises to perform the CSA function quicker and more efficiently than ever before, with real-time updating and instant, global distribution.
Ernst & Young installs the program on the customer’s intranet and helps the company identify its key operational risk variables. Data can be input from anywhere to be collected and modeled over Horizon’s application-server environment. The system links with a number of database applications as well.
There are three components. A self-assessment engine consists of a series of templates that identify and define major business processes, activities, risks, weighted control procedures and action plans. A performance analyzer contains current and historical risk, control and improvement opportunities displayed graphically for trend analysis, root cause analysis, risk-level forecasting and other risk management needs. Finally, the administration and management reporting function lets administrators add risks, change weightings for control procedures, add new activities, business processes and users, and administer security. The management reporting module also provides graphical management reports that highlight trends, project control targets, and track management’s prudent or imprudent acceptance of business risk.
When Ernst & Young sets up Horizon, the firm works with the client to determine the risk weightings and control procedures the system will use. Each risk element is also given a rating—very high, high, medium or low—which helps Horizon calculate a risk score. Once Horizon is installed, users begin filling out assessments of particular business processes in particular locations. The system first asks for a high-level view of the major risk, and then users drill down to specific risks within a basic tree structure. Along the way, red or yellow lights indicate that problems have been detected or more information is needed; users can proceed only with a green light. When a user gets a red or yellow light, he or she must enter the date the problem is going to be fixed, a description of how this will happen and the names of the people responsible for doing so.
Risk assessments are done on a quarterly basis. For each of the various risks, a risk score is computed from zero to 100, and results can be sorted by location to see how a particular risk is handled in a given area. “Let’s say a company’s operation in Chicago gets a score of 50 for business resumption risk,” says Craig Spielmann, global IT controls officer at JP Morgan, which has implemented Horizon on a company-wide basis. “A manager would have to ask, ‘Is it important that I fix this problem?’ The answer may be, ‘I have a few people in a hotel room in Chicago doing some business. Should I spend money on a business-continuity plan, or can I live with the score of 50?’”
Such grand risk management systems don’t come cheaply, of course. Ernst & Young is charging a hefty licensing fee in addition to regular maintenance fees, in return for all manner of product support, including a help desk, further product development and upgrades.
For more information, see www.horizoncsa.com.
|Was this information valuable?|
Subscribe to Derivatives Strategy by clicking here!