Models & Meltdown

Given the worldwide economic meltdown that weighs so heavily on just about everyone’s mind these days, it should come as no surprise that many of the articles and columns in this spring 2009 issue of Analytics focus on the financial sector. Because the collapse of the housing market in the United States and the shaky subprime mortgages the housing bubble was built on apparently triggered the meltdown – and since behind-the-scenes mathematical models apparently emboldened institutions to make bad loans – should analytics be the fall guy in this financial fiasco?

That’s the $3.6 trillion (total losses on loans and market value declines as of Feb. 13, according to NYU business professor Nauriel Roubini; www. rgemonitor.com/blog/roubini/) question. The answer, like just about everything else involving the economy, is complex and unclear.

Nevertheless, a mix of business school professors, financial modelers/engineers and analysts weigh in on the subject for various forums in this issue, including Viewpoint (page 4), Forum (page 6) and Last Word (page 42). The list of contributors includes Steven Shreve (co-founder of Carnegie Mellon’s program in quantitative finance), Steven Kou (professor of financial engineering at Columbia University), Don Kleinmuntz (co-founder of Strata Decision Technology), Mike Mount (analyst with a large national insurance company) and ManMohan Sodhi (head of the Operations Research group at Cass Business School in London).

While they all approach the topic from different angles and offer different takes on what happened and how best to fix it, a consensus begins to form around the notion that mathematical models and decisions based on those models are not perfect. As a couple of contributors point out, there’s “risk” in every model, because behind every model is a fallible human being. That’s the real lesson learned.

I look forward to your feedback. Analytics is an interactive exercise. Contributed editorial ideas, articles and columns are welcomed.