Catastrophic market collapses and rapidly inflating bubbles might challenge traditional financial risk models, but the models should consider extreme events, says a Science News article that cites four SFI researchers.
Discounting extreme events as improbable is a long-held tradition in economics, the article says. Many financial risk models assume economic events follow a Gaussian distribution, where extreme events are outliers that can be dismissed as improbable. But ignoring large, rare events doesn't capture reality, and might provide a dangerous false sense of security. Power laws might do a better job of describing the distribution of large events, suggest the researchers.
“There’s a significant chance that over a five-year period we will get hit by a really big event,” says SFI Profesor J. Doyne Farmer, whose work with External Professors John Geanakoplos and Stefan Thurner is cited. External Professor Mark Newman also is quoted, explaining why the Gaussian distribution has long been the paradigm.
Read the Science News article (Nov. 5, 2011)
Read an SFI working paper (August 1, 2009)
Read a 2008 paper by Farmer and Geanakoplos (November 10, 2008)
Read a 2006 paper by Newman (May 29, 2006)