Good piece by my friend Joe Nocera in the NYT on value-at-risk (VAR) in Sundayâ€™s NYT Magazine. He explains the subject well, pads around after Nassim Taleb for a while, visits with RiskMetrics, and generally does a nice, nuanced and skeptical piece of work.
After the lecture, the professor who invited Taleb to Columbia took a handful of people out for a late lunch at a nearby diner. Somewhat surprisingly, given Talebâ€™s well-known scorn for risk managers, the professor had also invited several risk managers who worked at two big investment banks. We had barely been seated before they tried to engage Taleb in a debate over the value of VaR. But Taleb is impossible to argue with on this subject; every time they raised an objection to his argument, he curtly dismissed them out of hand. â€œVaR can be useful,â€ said one of the risk managers. â€œIt depends on how you use it. It can be useful in identifying trends.â€
â€œThis argument is addressed in â€˜The Black Swan,â€™ â€ Taleb retorted. â€œNot a single person has offered me an argument I havenâ€™t heard.â€
â€œI think VaR is great,â€ said another risk manager. â€œI think it is a fantastic tool. Itâ€™s like an altimeter in aircraft. It has some margin for error, but if youâ€™re a pilot, you know how to deal with it. But very few pilots give up using it.â€
Taleb replied: â€œAltimeters have errors that are Gaussian. You can compensate. In the real world, the magnitude of errors is much less known.â€