Today’s fashion is to measure. It is the curse of the computer – we can measure it, so we will. More and more people do this when there is not a shred of evidence that the measurement will assist decision making.
This is particularly true of investment because the fashion is to use modern portfolio theory and its associated maths as the basis for measurement. This means predictions of extreme events will fall far short of reality because there are more “fat tail” events than theory predicts.
These days, many people believe it is sensible to say things like “our probability estimate for X is 65 per cent with a 75 per cent degree of certainty”. This confuses measurable risk with unmeasurable uncertainty. JM Keynes insisted on the distinction and said it was uncertainty, not risk, that was the greatest enemy of investors, which he argued could only meaningfully be said to exist if it could be probabilistically measured.
The likelihood of nuclear war, melting of polar ice caps, arrival of aliens on earth and many other events cannot be probabilistically measured. Nor can invention: Bell’s telephone, Marconi’s radio, Berners-Lee’s internet were unpredictable. So are the future inventions that will radically change the world.
We are misled by the incremental nature of most change into forgetting the radical nature of uncertainty. Who could have predicted an earthquake in Japan would prompt Germany to shut down its nuclear power industry and thus become – contradicting all its history – a gas-dependent ally of an increasingly totalitarian Russia?
The quasi-science of probabilism enables one to measure all sorts of relationships in financial data series. Most of this is junk and most people who bet the ranch on trends identified in this way end up riding donkeys.
Of course there are valid correlations and relationships but all are conditional and subject to the traders’ rule that “in a crisis, all correlations go to 1”.
Since 1945, investors have become prisoners of the idea that the right way to deal with risk is to mitigate it using probabilistic tools. But if instead you acknowledge that it is not risk but uncertainty that is the problem, you deal with it not by diffusing it but by embracing it.
A new form of investing that does this is already happening. Equity crowdfunding, offering individual investors the chance to place lots of small bets on “outsider” ventures, is what old-fashioned investment looked like. The risks are unquantifiable. Investment outcomes are pretty much binary – gold or dross. The punters know this and make lots of small bets – as private equity investors used to do before they became leveraged, fat and arrogant. Until the 1960s, the perceived risk of equities outside a small circle of blue chips – no more than 20 listed on the London market, one of which went bust in 1971 – was very high, probably as high as the perceived risk of small-caps today. The probabilism of MPT has allowed the risk to be pretended away – except in those “once in a lifetime” crunches, of which I expect my lifetime count to move from its current score of four to five within the next decade. Volatility is an analyst’s toy. It is not a good proxy for risk – especially not when it really counts.
Chris Gilchrist is director of Fiveways Financial Planning, a contributing author to Taxbriefs Advantage and edits The IRS Report