Magazine article The Wilson Quarterly
The Math Beneath
ON THE SURFACE, IT DOESN'T seem that financial modeling has much in common with climate science, ecology, or neuroscience. But in fact these fields are grappling with similar mathematical problems: how to map nonlinear, deeply interconnected systems and anticipate systemwide collapse, notes George Sugihara, a theoretical biologist at the Scripps Institution of Oceanography in San Diego.
For financial modelers, the challenge is to predict crashes. An investment banker looking at one portfolio will not be able to see the systemic factors that could lead to a meltdown. Likewise, a marine scientist trying to protect a species offish will not be able to account for all of the variables in the system that affect the survival of that particular fish without looking beyond that one species. In the field of climate science, linear models cannot predict what we know historically to be true: that climate change can be rapid and extreme. The existing models are all very good for painting a picture of a complex system at a specific point in time, but they do not have the ability to explain "jumps in variability," or what mathematicians call heteroseedasticity. …