Harvard Business Publishing

There is clearly some evidence that the current financial crisis was created--at least in part--by poor use of analytical approaches and techniques. If we're going to avoid similar crises in the future, we have to learn from our mistakes.

What's the evidence that the problem was somewhat analytical in origin? Let me count the ways:

• Banks and mortgage companies use analytics to make automated or semi-automated decisions about mortgage loans. Various industry experts have told me that many firms continued to make subprime loans even though a close analysis of the data would have suggested that chargeoff rates were climbing for such loans. The companies simply didn't monitor their analytical models closely enough.

• Saul Hansell in a New York Times blog argues that Wall Street quants "lied to their computers" in their analytical models. He states that they included more years of history in their trading models for mortgage-backed securities than was warranted in order to make them look less risky. Quoting Gregg Berman of RiskMetrics, Hansell also states that traders knowingly traded mortgage-backed securities even when the risk of the securities being traded wasn't accurately assessed by available metrics.

• There seems to have been a general problem in financial analytics with the transparency and explicitness of assumptions behind quantitative models. Many mortgage-oriented models were implicitly based on the assumption that housing prices would continue to rise. Credit default models were based on the assumption of continued liquidity in credit markets. Neither, of course, have turned out to be valid assumptions for the current period.

• I've heard several suggestions that some of the hedge funds that are having difficulty now used me-too trading and valuation models. The quant analysts involved either moved from one fund to another--taking their models and assumptions with them--or they reverse-engineered the models based on what they could learn through their social networks.

• It's also clear that risk analytics are not what they should be. AIG almost fell because of its inability to price and predict credit defaults; Moody's, S&P, and Fitch were clearly unable to assess the risk of mortgage-backed securities and attach accurate credit ratings to them. The 1987 stock market crash was caused in part by a similar inability to assess the risk of portfolio insurance, as Richard Bookstaber describes in the book A Demon of Our Own Design.

There are undoubtedly more analytical problems behind the current crisis. Of course, we can't only blame the quants. I've also heard stories of analysts telling executives that they needed to hedge mortgage portfolios, and they didn't act on the advice. We also don't know who was responsible for violating the implicit or explicit assumptions behind quantitative models.

Going forward, however, financial services organizations need to radically change their analytical focus. They need to incorporate "model management"--the systematic capturing and monitoring of analytical models--into their businesses. They need to be much more explicit and transparent about the assumptions behind models. They--and their regulators--need to be skeptical about the ability to model and manage risk in extraordinary circumstances. And financial firm executives need to learn much more about the models that are running their businesses.

Naseem Nicholas Taleb's book The Black Swan argues that there are some events that are impossible to model or predict. It appears lately that there are more black swans in financial markets than white ones. Our analytical approaches and philosophies must be adjusted to accommodate them.