Recently, I read an article that talked about forecasting done by analysts and the like, and how inaccurate they are most of the time. What I found to be the most interesting was the very consistent relationship between past performance and future forecasts. Evidence shows that many forecasts are just lags of what happened in the past, and have no accurate forecasting power. Research done by DrKW Macro has shown that often, forecasts are so off that when bond yields were forecasted to rise, they actually fell more than half the time!
In the world of business, we are addicted to forecasting. Whether its companies forecasting their future performance, preparing budgets, estimating future yields, or trying to predict the GDP two years from now, we often create and rely on forecasts. Forecasts are only useful if they can portray reality but how do you ensure you forecast accurately? Is it knowledge, intuition, or something else?
Some interesting research that was done by Torngren and Montgomery suggests that knowledge may actually cause poor forecasting. A study done showed that professionals actually had less accuracy in picking stocks than those who had no professional knowledge. Experts who are overly confident may make narrow and specific forecasts, e.g. "I predict the earnings for the S&P 500 will move up by 3 to 5 percent", when the real answer lies outside their forecasted ranges most of the time. It seems paradoxical in that, because the professionals are confident about their ability to predict, they end up not being able to predict well.
So is forecasting useless? I would disagree. Sometimes, forecasting is useless. But sometimes it has its advantages to some extent, like preparing budgets that can help a company manage its costs over the upcoming period. The challenge lies in being able to forecast accurately enough so that it is useful. To improve the accuracy of forecasting, I think it is important to tie other indicators (including non-quantitative ones) to forecasting, like incorporating management effectiveness in predicting the future earnings growth of a company.