Conversations on standard risk modeling in today’s markets often emphasize factor construction and the art of selecting factors best suited to describe an investor’s view on systematic market drivers. While relevant in the context of the investment process, these discussions tend to overlook a key, and arguably critical, element of factor modeling and risk model construction. The benefit of choosing meaningful factors for risk modeling is only one part of the story and is largely diminished if these factors are not modeled in a proper fashion.
Take the global financial crisis for example. Here we saw the massive and systematic breakdown of risk models almost across the board in part because investors were relying on risk models that did not properly model the joint distribution across assets. In particular, many standard risk models in the industry have historically used Gaussian assumptions, otherwise known as the normal distribution, to model the dependence structure across assets. However, these methods are misleading in the real world because they show perfect diversification in extreme event scenarios (i.e. the tails of the distribution) when in reality, all diversification disappears.
Taking a step back, there are some key phenomena that are observable characteristics of financial markets:
Volatility clustering: volatility comes in clusters, with periods of high volatility persisting and similarly periods of low volatility persisting. This implies that realized volatility has an impact on future volatility.
Asymmetry of returns: in return space, negative events tend to be much more pronounced, especially in equity markets, than upside events.
Extreme events: extreme events occur outside the realm of “normal” market behavior. This means normal assumptions will assign unrealistically low probabilities to extreme events, especially during times of market duress.
There are modeling techniques available to combat the pitfalls from all three phenomena described above. FactSet has incorporated these techniques into the Multi-Asset Class Risk Model framework through enhancements to our volatility modeling, marginal distribution modeling, and modeling for the dependence structure between risk factors. Not only do we expect these techniques to more accurately estimate risk during crisis periods, but they’ll also allow for better adjustments to volatility in calm markets and avoid overestimation of risk.