I had the very great privilege of taking part in a roundtable lunch with Myron Scholes last week, thanks to Janus Capital (see unashamed photo below) –
hosted in the Fenchurch Brasserie in the well-known “walkie-talkie” building in London the views were pretty good as well (see below).
Despite his fame, Myron still clearly thinks deeply about markets on a day-to-day basis and made some well-argued and thoughtful points, principally around the interaction of risk management and long-term asset management including –
- The incidence of Tail Risks is very damaging to returns, even over long periods of time
- Beware “convexity costs”: the fact that large positive and negative return swings will tend to drag down compound average returns over time *
- As well as diversifying across asset classes investors should logically diversify across time also, by trying to ensure their portfolio is equally exposed to risk in different time periods
- Fixed allocations while appealing do not naturally have this property, dynamic allocations to assets are required to keep risk constant
- The options market, by it’s “Darwinian” nature may contain useful information on tail risks, which most investors do not or are not able to use
And on LTCM…
- Myron suggested that he was highlighting the nature of the large tail risks in the fund to the managers going into 1998
On the Black-Scholes equation …
- Myron suggested that they spent several years in the early 1970’s attempting to arrive at a formulation that allowed variables like volatility and interest rates to vary over different time-periods, but the maths was too difficult to work so Fischer Black made the suggestion to keep them constant, which greatly simplified the mathematics allowing them to publish
Here are some of my previous thoughts around approaches to tail risk hedging in practice. Click here to download a paper examining different practical approaches to tail risk hedging for UK pension funds.
* for a simple illustration of the convexity cost point imagine a gamble that creates an equal chance of a 20% gain or a 20% loss. On a single-run basis the expected profit or loss is zero, however simple compound maths shows that a 20% gain followed by a 20% loss (or vice versa) in fact creates a loss of 4%. Running the same gamble over and over again results in a substantial loss through time on a compound basis.
