One of their targets has been the concept of VAR, or value at risk. "VAR has been the risk measure of choice on the sell side for many years," observes Raj Manghani, director of product strategy for enterprise risk management at BARRA Inc., "but now investment pros are skeptical about using any one measure. No single measure tells you everything you need to know. Each metric needs to be applied in context. So you have 'shortfall risk,' 'conditional VAR,' 'scenario-based VAR.' There are more flavors of regression analysis–'cross-sectional' or 'time series,' for instance. The nature of the business determines which combination you should use." For instance, Black-Scholes still is used to value European-style options, but now Cox-Rubinstein (also known as the binomial model) has emerged for American-style options, Manghani says.

Leading risk scientists like Rene Stulz at Ohio State University have synthesized the collection of tactics for managing individual risks into a new paradigm for managing risks in a way that increases the value of the firm. "Firms are increasingly focused on measuring risk and managing risk across the whole firm. The benefit from doing so is that they can choose a level of risk that they think maximizes shareholder wealth and manage towards that level. For instance, if a firm believes that an A rating is appropriate, it can achieve that rating and keep it by choosing a level of risk that corresponds," Stulz explains. "As it understands the determinants of the risk of the firm, management can decide whether it leaves money on the table by choosing that rating, whether it can achieve that rating through hedging, or whether it needs to change its investment policy or raise more capital."

Among the new leaders in risk science are John Hull at the University of Toronto, Mark Rubinstein at the University of California at Berkeley and Kay Giesecke at Cornell University.

Most of these new risk scientists are mathematicians or physicists who practice in finance, not finance specialists who learn the science. That's because the science is the hard part. It's a lot easier to teach a scientist what financial applications would be useful than it is to teach a savvy finance pro the underlying science, Lalancette explains.

The change is most obvious at large commercial banks, investment banks, hedge funds and pension funds that make their money primarily in the financial markets. And that is where most of Lalancette's students are destined to work. The treasuries of most industrial and service companies are too lean to afford a Ph.D. in math or physics who is dedicated solely to risk management.

Still, new risk theories and models travel from academia to treasury practice faster than ever, notes Tom Koundakjian, vice president for enterprise solutions and services at BARRA, which applies sophisticated risk science to investment portfolios. That's because there are more former academics working in treasuries who understand the new inventions and know how to put them to use than there used to be.

However, dipping into the published work of these risk scientists could give a busy treasurer a migraine. There's a big difference between reading a corporate balance sheet and reading Greek-symbol-laden risk computations. While the marriage may be productive, there is culture shock on both sides when a scholar enters a corporate treasury. Lalancette was free to pursue his research in any direction and manner he chose when he was at the HEC. At Hydro, he was bound to a corporate agenda, supervised by managers and compelled to work on a team. Treasury pros, for their part, are likely to find the scholars arcane, impractical and just plain weird.

But there are some pricey reasons why it's worth it for a company to bring the two worlds together. Many, if not most, treasuries are probably overpaying for their hedges because they aren't using the latest science.

So how can an ordinary treasury avail itself of these new models and metrics? You can't buy magic software and click on the right icon to get the computer to do it for you. Risk varies greatly from corporation to corporation, so generic solutions won't work.

It does take considerable computing power to make the applications work, however. "Ten years ago, the cost of the hardware and the software to run textbook VAR was out of reach for many companies," Koundakjian notes. "What only investment banks could afford 10 years ago is cost-effective today for a great many corporations."

But it takes smart people to direct the computations. And this is serious egghead stuff, the financial equivalent of rocket science. You can't send an AT to a two-week conference and expect him or her to bring back the solution.

You could hire a consultant, from inside or outside academia, of course. But your most economical and possibly most effective solution might be to offer an internship to one of Lalancette's students, or one of their peers at another B-school, who is concentrating on quantitative risk management and encourage that student to use your company for his or her research.