In this episode of the Curious Quant podcast series, I sit down with Nick Wade from Northfield Information Services to look at the idea of innovation in risk modelling, areas that have and haven’t seen change and the importance of time horizon in capturing risk. Reflecting back on his career at Grantham, Mayo, Van Otterloo & Co and his experiences across the US, UK and Asia we explore his views on the benefits and challenges associated with machine learning in finance. He makes some excellent observations.

On time and investments:

“Empires rise and fall, and various other horrible things happen over longer horizons, and generally speaking in the asset management industry we don’t look at that stuff, we worry about volatility between now and a horizon that is not long enough for most asset owners.”

On innovating in models:

“The two things that you can do, one is have an adaptive structure so the model can adapt and learn as new things become important and secondly, use some information that you know today about how the future is going to be different from the past.”

On major challenges:

“The more difficult question is how we spot changes in regime, over the last decade or so we’ve seen a change from very laid back, non-interventionist central bank policy to a very active, macro environment that’s obviously very different.”

You can download the podcast from the Apple and Spotify podcast players or listen below: