Forecasts about inflation, growth, and unemployment made by top economists at major banks shape key decisions for businesses, investors, and the Federal Reserve. These predictions are critical, and when they miss the mark, it can have major consequences—either by being too optimistic or too pessimistic.
However, new research from Berkeley Haas points to an even bigger issue: forecasters are often too certain about their predictions. The study looked at the longest-running survey of professional forecasters and found that while forecasters were, on average, 53% confident in their predictions, they were right only 23% of the time.
“These are experts—highly trained, with lots of experience,” the researchers note. “We wanted to see if they suffer from overconfidence like everyone else. And they do.”
Specific errors
The study highlights a specific type of forecasting error: over-certainty and over-precision.
The researchers analyzed the Survey of Professional Forecasters, which has been conducted by the Federal Reserve Bank of Philadelphia since 1968. Senior economists at major banks and corporations predict key economic indicators each quarter, such as GDP growth. The data is used by organizations like the Federal Reserve to guide decisions, including interest rate policies.
The survey asks economists to predict the probability of economic indicators falling within certain ranges—say, GDP growing by 1-2% or 2-3%. The researchers analyzed 16,559 forecasts and found that the economists’ predictions were correct less than a quarter of the time, even though they were 53% sure they’d picked the right range.
The problem wasn’t that the forecasts were consistently too optimistic or pessimistic—some economists predicted higher numbers, while others expected lower ones. The issue was that they were too precise, overconfident in the accuracy of their predictions.
Experience doesn’t help
Interestingly, the more experienced the forecaster, the more likely they were to get the answer right. But they were also more prone to overconfidence, which offset their increased accuracy.
“This shows that no one is immune to error,” the researchers say. “The Fed should keep this in mind and build flexibility into its monetary policies, knowing that predictions are inherently uncertain.”
Fortunately, while individual forecasts may be overly precise, predictions tend to be more accurate when averaged out. Because there’s no consistent bias towards optimism or pessimism, the combined forecasts usually fall closer to the actual outcomes. In other words, disagreement among forecasters improves the accuracy of the collective prediction.
The study’s lessons go beyond economic forecasting. Whether we realize it or not, every decision we make involves some prediction of future outcomes.
“Overconfidence isn’t just about being overly optimistic,” the researchers explain. “It’s about being too sure you have the right answer. This can cause you to miss opportunities or allocate resources poorly. You can be both too pessimistic and too certain at the same time.”
Most people don’t have access to a wide range of expert forecasts in their everyday lives, so the researchers suggest a few strategies for better decision-making:
- Consider where you might be wrong.
- Weigh the costs and benefits of different errors.
- Stay open to feedback.
“Recognizing your vulnerability to error helps you stay better calibrated in your confidence, which improves your odds of making good decisions,” the researchers conclude.
