You are here

Regression to the mean could also result in history repeating itself

WHEN 19th Century scientist Francis Galton set out to investigate the inheritance of physical traits in fathers and sons, he discovered a now-widely accepted statistical phenomenon he later called 'regression to the mean' (Galton is also famous for discovering the science of fingerprinting).

Galton found that tall fathers were more likely to produce shorter sons and short fathers tended to produce taller sons. It was as if some mysterious force was causing human heights to move away from the extremes and towards the mean or average of all humans.

When you think about it, it makes perfect sense - after all, if tall fathers tended to sire tall sons and if short fathers produced short sons, then after several generations the human race would comprise very tall people on one end and very short people on the other.

Regression to the mean is therefore needed to prevent this from happening and thus maintain some kind of natural stability. What's interesting is that the phenomenon holds for almost all fields of scientific observation, the stock market included. There is a long-term average towards which prices will tend, though in this case it isn't a genetic imperative that establishes stability but the natural urges of greed and fear. The interesting question therefore is: where does the mean or long-term average lie for stocks?

With the exception of a small minority, I think most observers would agree that the present climate for stocks is not bullish by any stretch of the imagination. China is slowing, US Federal Reserve hawks are chomping at the interest rate bit (though a rate hike in June doesn't look likely), corporate earnings are weak all over the world and a slide back into a global recession cannot be totally discounted. In about 5 weeks the UK will vote on whether to remain in the Eurozone, the outcome of which has the potential to send volatility spiking up as the Greek vote on a similar issue did last year.

However, even though this is by no means a bull market, it isn't a full-fledged bear market either, thanks to continued hopes of more money printing by central banks, especially the ECB, PBoC and BoJ. As long as markets think they can be bailed out by officialdom, there will be a floor under stock prices. Which brings us back to the original question - if markets are prone to regressing to the mean, what might that be for the local market?

In my view, a long-term Straits Times Index average would be somewhere between 2,500 and 3,000. I don't have any statistical backing for thinking this; my guesstimate is based primarily on my having covered the local market for BT since 1992. More important than my gut feeling, does "the mean'' for the local market encompass the conditions now prevalent, ie low retail participation and a lopsided concentration of daily dollar volume in the 30 STI components (yesterday, 76% of dollar value recorded by the whole market came from the 30)? Does "the mean'' also include little interest in penny stocks, periodic delistings/privatisations and fewer IPOs?

Interestingly, 12 years ago after the market had bounced 30 per cent in 2003, I wrote that "unlike previous rallies, the vast majority of the Singapore retail public had not benefited, which is why there had not been any flow-through impact on the real economy as the retail sector and property market were still mired in the doldrums''. If that sounds startlingly familiar to the current situation, then not only do markets tend to regress to their means, but it looks like history has a funny way of repeating itself. 

writeonthemoney.png

The Singapore stock market observed.