Commodity Risks: Describing the Unobservable

We have observed in studying commodity markets that 100-year floods occur quite often, even multiple times a decade, so we know simple risk models can be inadequate and misleading. The challenge is that expected risk-return probability distributions cannot be directly observed. Some analysts lean heavily on examining the implied volatilities from options prices.  Unfortunately, the implied volatilities all too often underestimate what is happening in the extremes of the distribution where the high impact risks are located. 

Our view is that we should explore much more robust measures of risk. We need to appreciate that volatility is not the same as risk, and that the standard deviation is a very poor risk metric on which to rely so heavily. Our approach is to go beyond futures and options prices and include information from volumes and intra-day activity in our methodology to allow for multiple scenarios which avoid the bias toward the bell-shaped distributions that appear highly flawed.

In this research, we first briefly make our case for why volatility is not the same as risk.  We then tackle the question of why the implied volatilities derived from options prices can also be a dangerous and misleading risk metric. Then, we describe our approach and apply it to an example from the corn market to give readers a flavor of our research approach.

Download now