Forecasting is pure folly

New year, new forecasts. Where will stocks be in 12 months’ time? Is CRH worth a punt? What about Apple? Where next for gold prices?

Thankfully, benevolent brokers everywhere step forward to offer their words of wisdom at this time of year, slapping precise price targets on every stock under the sun.

Some sceptics see all this prognostication as laughable, almost touching. Others, like Black Swan author Nassim Taleb, decry the “scandal of prediction”, the appalling arrogance of “experts” who “do not know what they do not know”.

Many people, unfortunately, lap it up, having already forgotten the lessons of 2008. That year, said financial commentator and CNBC personality Jim Cramer, would end with Goldman Sachs at $300 a share (“Not a prediction — an inevitability”). Investors had to settle for $80.

“Buy some of the most beaten-down stocks, including those of giant financial institutions such as Lehman Brothers, Bear Stearns, and Merrill Lynch,” advised Elaine Garzarelli, who made her name after predicting the stock market crash of 1987. Irish analyst forecasts were equally embarrassing.

Of course, it’s not fair to cherry-pick. Everyone gets things wrong. However, 2008 was no exception: it was merely a more pronounced example of what market strategist and author James Montier calls the “folly of forecasting”.

Forecast error

Between 2001 and 2006, notes Montier, the average analyst forecast error over a 24-month period was 95 per cent in Europe and 93 per cent in the US.

It’s not a new phenomenon. In 2010, global consulting firm McKinsey examined analyst earnings estimates over the past 25 years and found that growth estimates ranged from 10-12 per cent a year, almost double the growth achieved (6 per cent). Contrarian fund manager David Dreman analysed almost 100,000 forecasts made between 1973 and 1996, and concluded that forecasts typically missed out by between 30 and 60 per cent.

Big earnings surprises are the norm, Dreman found, and there is a mere one in 170 chance that consensus forecasts will be within 5 per cent of actual results for four consecutive quarters.

Economists? “Haven’t a clue”, says Montier. “The three blind mice have more credibility than any macro-forecaster at seeing what is coming.” Stats back him up. A study of six different national recessions in the 1990s found that in 97 per cent of cases, the economists failed to predict the event a year in advance.

How come? In The Little Book of Behavioural Investing, Montier explains just why the task is so difficult. Let’s say you invest by forecasting the economy, the path of interest rates, sectors which will do well in that environment, and finally which stocks will do well in that sector.

Let’s say you are right on each forecast 70 per cent of the time (way above rates actually seen). The odds of getting all four forecasts correct? Just 24 per cent.

“Now think about the number of forecasts an average analyst’s model contains – sales, costs, margins, taxes, and so on. No wonder these guys are never right.”

It’s the same in all fields. Political psychologist Philip Tetlock examined over 27,000 predictions from 284 “experts” over a 16-year period, and found they were only slightly more accurate than chance. Tetlock has also found that political experts who reported 80 per cent or higher confidence in their predictions were right only on 45 per cent of occasions.

Doctors are even worse. In one study doctors were given patient case notes, asked to make a diagnosis and then state their confidence in that diagnosis. When they were 90 per cent sure they were right, they were actually correct on less than 15 per cent of occasions. In short, we’re awful at forecasting, but do so out of both ignorance and overconfidence.

In Thinking, Fast and Slow, psychologist and Nobel economist Daniel Kahneman describes how he was given data from an investment firm summarising investment outcomes from 25 of the firm’s advisers over an eight-year period. After doing the maths, Kahneman told the firm their results were what you would expect from a dice-rolling contest, not a game of skill.

No one contested the findings – they couldn’t – but the response was “bland”, said Kahneman, who opined that the “illusion of skill” is “deeply ingrained in the culture of the industry”.

Accompanying disclosure

Back in 2005, market strategist Barry Ritoltz scoffed that all pundit forecasts should be accompanied by a disclosure that the forecaster “states that he has no idea what’s going to happen in the future, and hereby declares that this prediction is merely a wildly unsupported speculation”.

Unsurprisingly, that hasn’t happened. For one, it would be like turkeys voting for Christmas. Secondly, we instinctively think otherwise.

Even Kahneman describes how, early in his career, he discovered his own predictions to be “little better than random guesses” but “continued to feel and act as if each particular prediction was valid”.

Investors should note that the best money managers have no truck with the daft guessing game that so often passes for analysis. George Soros, for example, has said that “the idea that you can actually predict what’s going to happen contradicts my way of looking at the market”.

James Montier, too, is an excellent strategist who has detailed quantitative value-based strategies that have long delivered market-beating returns. But basing investments around our “seriously flawed ability to divine the future”? That, says Montier, is “sheer madness”.