Forecasts are the most unreliable guide to market returns, except for all the others. Ron Bewley explains.
I read a blogger the other day commenting on a newspaper article about various high-profile strategists’ market forecasts for 2011. His point was simple: Why write up all of these forecasts when they are nearly all always wrong? And it seemed to be a point of frustration, rather than a serious complaint.
I update my forecasts every day, even though I know they are going to be wrong. I do not feel comfortable without these forecasts. A good forecasting process is essential for determining whether the market is cheap or expensive. And if the fundamentals actually change – so changing my forecasts – I would not otherwise know that I should adjust my investment strategy.
My forecast for the S&P/ASX 200 for 2011 is 14 per cent capital gains with a volatility of 12.5 per cent. The year 2010 ended at 4745 – so the thick smooth black line in Chart 1 represents my forecast for 2011, ending at just under 5500.
If my forecasts are based on sound research and analysis, I will get the trend right, but market volatility can take reality a long way away from that line. The 10 wiggly lines in Chart 1 are scientifically calculated “simulations” based on my forecasts of returns and volatility. Importantly, each of these wiggly lines is equally likely! There are, of course, an infinite number of possibilities, even if I am right. I show 100 of these simulations in Chart 2. Some of the simulations deviate far from the trend line while most are clustered closer to the trend.
Now you can see the problem with market forecasting. An outcome of 5500 has a reasonable range of something like 4000 to 7000. To claim greater accuracy than that is to refute the market volatility that we are all so familiar with. On top of that, fundamentals might change over the course of the year as big macro events come to light, and so the trend should also change.
Forecasting is futile without understanding the principles of forecasting. Because I believe that my forecasts are credible – they are based on broking analysts’ forecasts of dividends and earnings for the top 200 companies – I can use deviations from the thick line to determine buying opportunities and times when profits might reasonably be taken. Without that benchmark, I could not distinguish between new upward trends and temporary little bubbles.
My method of calculating this mispricing is somewhat more complicated because analysts’ views of the world do evolve over time. My article last month describes my so-called exuberance measure and how to use it.
Importantly, different sectors of the market move at different paces and with different sets of “noise” causing the outcomes to deviate from their smooth paths. So, by determining which sectors are better than others, and which are the better times to buy, market outperformance might be achieved – even when the number, such as 5500, is completely wrong! Indeed, if forecasts were always correct, there would be limited buying.
The wiggly lines will often cross the trend line several times over the course of a few years. So the real play is in monitoring changes in trend and volatility over much longer horizons. No one will get every call right but, as they used to say with British Rail, “Without a timetable we would not know how late the train is”!
I know my forecasts will be wrong, but I cannot imagine life without them!
Ron Bewley is executive director of Woodhall Investment Research – www.woodhall.com.au
