Tuesday, December 20, 2011

Forecasting Follies

Writing in his book Thinking, Fast and Slow, psychologist and Nobel Prize winner Daniel Kahneman discussed the perils of overconfidence.

He talks about the illusion that all of us to some degree share about our ability to forecast the future.

Although there is overwhelming historic evidence that many events - political, markets, etc. - are merely random, it doesn't prevent us from listening to forecasts from learned experts.

For example, here's one experiment that Dr. Kahneman describes:

For a number of years, professors at Duke University conducted a survey in which the chief financial officers estimated the returns of the Standard & Poor's index over the following year. The Duke scholars collected 11,600 such forecasts and examined their accuracy. The conclusion was straightforward: financial officers of large corporations had no clue about the short-term future of the stock market; the correlation between their estimates and the true value was slightly less than zero!

Kahneman goes on to describe how the professors asked the CFO's to give "confidence ranges" in which they were fairly certain that their forecasts would be accurate.

Here again, unless the CFO's gave a sufficiently wide range of potential outcomes (e.g., that the market would return somewhere between -10% and +30%), the actual results compared to the "highly confident" forecasts were often wildly different.

I bring this all up because year-end tends to be the time of year when you will see all types of forecasts, ranging from the markets, weather, or elections. However interesting these discussions might be, we should recognize that statistically most of them have little chance of actually coming to pass.

Writing on the blog Business Insider, former Wall Street analyst Henry Blodget discusses the fallibility of forecasting.

He cites a number of different areas where analysts and economists from the Street have been consistently wrong, yet continue to make forecasting that somehow continue to gain a wide following.

For example, he notes that most economists will forecast moderate growth for the coming year. The reason is simple: for a mature country like the United States, moderate growth tends to be the norm. Forecasting strong growth, or a severe downturn, may make headlines but can be severely career limiting if the forecasts prove inaccurate.

Here's what Blodget writes:

If economists can't predict the future, why do they always predict that the economy will grow about 4%? Because that's what the economy's long-term growth average is--and, therefore, that's the prediction that gives the economists the best odds of being generally "right" (or at least not too embarrassingly wrong).

Just as no one ever gets fired for buying IBM or hiring someone from Harvard B-school, no one ever gets fired for predicting that the economy will do about as well as it has always done. And, of course, staying close to the average also gives the economists the best chance of being close to right. So that's what economists predict!