|Monte Carlo' Software Is Applied to Risk Management
Banking and casino gambling should have little in common, but a Phoenix-based consulting firm has introduced PC-based forecasting software that uses "Monte Carlo" analysis techniques to better predict bank earnings.
"Most banks have a difficult time accurately forecasting interest rates. This in turn affects their income forecast," said Dennis Bennett, president of Bennett Management Services, a firm that specializes in helping financial institutions with interest-rate planning, also known as asset-liability management.
"The Monte Carlo technique simulates reality enough times to come up with a probability chart of how reality may come to be," Mr. Bennett said.
In Long Use Elsewhere
The Monte Carlo model has long been applied in other industries. Pioneered during World War II, "Monte Carlo" became a code name for a weapons-industry project that dealt with probability and chance - hence the gambling reference.
The early Monte Carlo simulation was so accurate that the British government has used it to predict traffic and toll patterns as a means of coming up with municipal bond prices. Scientists - particularly biochemists and engineers - regularly use the technique in their research.
Mortgage bankers also use Monte Carlo analysis to adjust prices of asset-backed securities and derivatives.
Technological Limitations
Because of software and hardware limitations, however, Monte Carlo analysis is new to asset-liability management in banking.
"Monte Carlo analysis takes a lot of number-crunching," Mr. Bennett said. "Banks which were using [the method] in the mortgage area typically ran the programs on a mainframe because the analysis required so much power. Up until now, it hasn't been feasible for other parts of a bank to use it."
The new software runs on International Business Machines Corp. (or compatible) PCs using Intel 80286, 80386, or 80486 microprocessor chips. The software costs $30,000.
Traditional asset-liability management software, which often runs on personal computers, has tended to rely on analysis techniques that generate earnings and market value for a limited number of interest rate forecasts.
These types of programs don't have the same data processing requirements that a Monte Carlo-based system does, Mr. Bennett said.
Mississippi Banker's View
Blair Bingham, vice president in funds management at Deposit Guaranty Bank in Jackson, Miss., said the Monte Carlo software offers a new kind of analysis - quicker and more efficient than the earlier methods. "It captures the risk of our net interest margin on one snapshot and gives senior management what they need to see," he said.
Bennett Management is something of an upstart in a software niche long dominated by established players such as Sendero Corp., Scottsdale, Ariz.; Banking Decisions Systems, Waltham, Mass.; Financial Technology Inc., Chicago; Darling & Associates, North Andover, Mass.; and Chase Financial Technologies, a unit of Chase Manhattan Corp.
The Bennett product, called Possibilities, takes simulation to another level by forecasting 200 to 300 earnings scenarios on a single graph, often within minutes.
Expanding Analyses' Scope
"Every method starts with a balance sheet," said Dr. Carl Brooking, professor of economics and quantitative management at Millsaps College, Jackson. "But the fact of the matter is, hundreds of simulations summarized in a meaningful way are a whole lot more helpful than 12."
To maximize efficiency, Possibilities has been designed to exchange data with a bank's host computers as well as with other PC-based software, such as Lotus 1-2-3.
Mr. Sullivan is a freelance writer based in New York City.
