The human factor, a reluctance to scare investors and the absence of tangible rewards for predicting recessions are some of the reasons why professional forecasters sometimes fail to signal economic downturns.
In their study, Smirnov and Avdeeva found evidence indicating that professional forecasters tend to hold on to optimistic scenarios for too long.
According to the authors, excessive optimism is one of the factors which can hinder accurate forecasting. Smirnov and Avdeeva analysed the quarterly consensus forecasts of real GDP growth rates and expert-estimated probabilities of recession in the U.S. between 1968 and 2015, based on the Survey of Professional Forecasters (SPF) conducted by the Philadelphia Federal Reserve Bank (PhilFed) and involving leading experts from major investment and commercial banks, real sector companies and think tanks.
Based on their findings, Smirnov and Avdeeva came up with a list of possible reasons why professional forecasters can sometimes be wrong. First, the study's authors argue, many analysts are more focused on long-term trends and less on predicting imminent recessions. Second, some forecasters use relatively primitive models and rely too heavily on extrapolating past patterns into the future. But even where a forecaster can clearly see an impending worst-case scenario, they may not necessarily report it. Many experts are understandably reluctant to scare investors and thus trigger a self-fulfilling prophecy, since a negative forecast can cause a company's stock to collapse in a matter of minutes. This is the third reason why recessions may not get accurately predicted.
The fourth reason is the human factor: most experts tend to hope that recession risks might not materialise. And finally, the fifth reason is that experts cannot usually expect a reward for predicting a recession correctly. In such circumstances, many forecasters prefer to ‘follow the crowd’ so that even if they fail to signal an upcoming recession, no one could accuse them of incompetence.
In addition to possible causes of errors, the study suggests that many forecasters are over-optimistic as to the rates of economic decline, duration and depth of recessions. For example, never in the SPF history did consensus forecasts predict any recession to last longer than six months, with the only exception being a period after the second oil shock in the late 70s, when recession expectations were very strong.
Similarly, experts tend to overestimate the rates of economic growth during recessions and underestimate them during economic upturns. In most cases, the expert community starts sending out warnings of a recession after a cyclical peak, when economy is already on a downward trend. In addition to this, consensus forecasts tend to clearly signal a recession only after a so-called ‘black swan’ event.*
However, according to the study's authors, optimistic bias in economic forecasts is not a disaster, as long as it is recognised and adjustments are made, based on assessments by the most reliable forecasters, while policy and business decision-making should take into account the usual delay of negative signals.
*An unexpected and exceptional event with an extreme impact.