How do you apply options pricing to the GLI portfolio? Wednesday 24th January 2018
If you look at Black Scholes, where you have volatility as a component, you can “reverse engineer” implied volatility to get today’s market consensus of the future, based on an option price. Three-month volatility is a relatively decent predictor although even it can change at a moment’s notice. We combine that with historical volatility – reality – and we get a blend of forward and backward looking volatility to reach our target risk: the downside deviation of the S&P 500.
Every day, we monitor volatility through options pricing, from real time Bloomberg feeds. Real risk changes in real time too. So we have a good sense of that and use it to adjust certain things, like leverage. But this is not connected to a trading algo: we’re not sleeping while a machine operates alone. We’re looking at a model, to check against other models.
There are lots of people who try to predict the future, and we don’t. We are seeing what the markets feel about volatility, we use this to keep our risk exposure constant and benefit from world growth driving up asset prices over the long term.
I must admit at the very beginning, we weren’t sure if the entire model would work in real life. We weren’t doubting the validity of the model per se, but we had no proof yet whether the 1998 model would be sustainable in real life, because with models, relative importance of parameters shifts, and we did not know then how much and how fast this would happen. We had a somewhat rough start, but we also knew a 1 to 2-year time frame was too short to judge it.