Risk management in uncertain times

From the Hub section

Risk management in uncertain times

Risk means different things to us all. In the context of investments, securities and portfolios major macro events such as the UK Brexit process give food for thought and an opportunity to challenge how you go about assessing, measuring and dealing with risk.

Consider the three related concepts of risk, uncertainty and volatility.

Risk is such a wide notion that can mean so many different things, so let us put it to one side for the moment. Uncertainty and volatility sound so similar - yet consider the post Brexit vote reaction. There was not excessive volatility or market movement given the magnitude of what happened. However there is as much uncertainty as we can possibly try and contemplate. Why is that?

Brexit was all about long-term trends, fault lines and dislocations coming to a head. In many ways it was like a political version of the Lehman bankruptcy. Neither of these momentous events happened in a short space of time. Neither were properly anticipated. Both however make total sense in hindsight. Both will cause massive after-shocks for years.

So why should of all this change our view of risk?

It doesn’t really add much to our collective experience and toolkits for what you might consider “short term risk”. The way traders go about hedging, or how risk control functions monitor short dated Value-at-Risk (VaR) limits for example. To them Brexit would have been seen as just another market event. Unexpected and dramatic definitely, unique even - but there to be dealt with.

It is the medium or long term view where a little more creativity of analysis is needed. This therefore impacts decisions like long term asset allocation as well as long term performance analysis. From my company's work in structured products where typical time horizons are six years I know the importance of accurate long term projections and analysis. It is an extremely difficult and complicated topic with no "right answer".

If we take the view that we have significant uncertainty but not yet volatility, this implies that volatility or different scenarios will play out in the future. In order to manage these risks we need to do more than adjust market estimates of growth and volatility and run our usual models.

Maybe the markets will behave “normally” for a while and then drastically start to diverge depending on what happens further down the line. We could be set for falls then long-term market decline, perhaps instead short term dips following by rebounding over the longer term or just about any other scenario.

The longer your horizon, or the more markets and asset classes that affect you the bigger the challenge. And lest the obsession with Brexit is starting to grate on you do not forget we could be in for the same thing in the rest of Europe, or in a different guise but with much more at stake when the US elects its President in November.

Maybe we should consider models that allow for long term changes happening in some random way in the future. Or models which consider hidden variables that will eventually feed through to market prices and other observables. From a technical perspective this includes non-Markovian models and chaotic systems.

I am not advocating suddenly adopting a slew of complicated models because they have their own weaknesses, on the contrary I am a great believer in the right holistic approach. The alternative frameworks that I described are certainly not new but I suspect they are under considered. In our own client work of stress testing and long term portfolio management we are already re-evaluating how we perform our analyses, and assessing whether we have everything covered.

Therefore the real take away I would like to leave you with is whatever your responsibilities around risk, and whatever your current modelling and data set-up, stand back and ask yourself: “Is my approach comprehensive enough or am I missing something?” That is a question everyone should ask themselves regularly and certainly right now.

Tags: Algorithms Private Debt

Related Posts: