Chris Gilchrist: A love-hate relationship with uncertainty


Many of the risk reduction technologies out there worry me. However smart your box of tricks, it is still a box of tricks playing with historical data. You can spin the stochastic wheels all you like but do not pretend the results give you a handle on the future, because they are only projections of the past. I prefer the original term for stochastics: Monte Carlo simulations.

Many of those applying these tools seem to believe they are controlling risk, whereas all they are actually controlling – and only within the limits given by history – are standard deviations.

Just as the persistence of low volatility tells you nothing about the actual level of risk (a sudden spike up in volatility will instantly unleash a tsunami of expert comment explaining the risk that was always there), the output from stochastic engines will always be converging on history.

The more consistent the outcomes look, the worse they will actually be in terms of the disruptive unknowns that are the real source of risk and which they cannot estimate.

This is all familiar territory to students of Knight and Keynes, both of whom emphasised the far greater importance of unquantifiable uncertainty than of quantifiable risk in investment.

What seems to me most important for financial planners is that people do not, as is often said, hate uncertainty. In fact they have a love-hate relationship with it. Their attitudes towards annuities demonstrate this perfectly.

Why would you not be happy with an inflation-proof income guaranteed for the rest of your life? It is about as much certainty as you can get. But many people want the chance to get higher returns, cash in a bit of capital or hand money onto their heirs. So they exchange certainty for hopes and risks of one kind or another.

Uncertainty is an essential feature of investing. Investment advisers attempt to place limits on that uncertainty because part of their job is to limit the possibility of ruin for their clients. But over-quantifying the process of generating suitable solutions is not a way of improving the solution, nor of improving clients’ understanding of solutions.

The trick – simple but not easy – is to know when to rely on historical data and when not to. One way of framing this is to say that financial variables all appear to revert to their mean: equity income yield, long-term bond yields, cyclically-adjusted price to earnings, book value and so on.

The major uncertainty is the time it takes them to do so and the shape of the journey. However, the longer a variable has been well above its historical mean, the greater the likelihood of a snap-back.

Our current issue is that quantitative easing has distorted markets so much it is hard to make the obligatory intelligent, well-informed guess about how reversion will play out this time around. Needless to say, I do not think spinning the Monte Carlo wheels will help.

Chris Gilchrist is director of Fiveways Financial Planning, a contributing editor to Taxbriefs Advantage and edits the IRS Report