How Accurate Are Crowdsourced Economic Forecasts?
Why crowdsourcing is a surprisingly accurate alternative to the consensus.
May 1, 2018
People can crowdsource just about any service these days, from creative services to fundraising. But can it be used to create better and more accurate economic forecasts? Platforms like Estimize seek to democratize securities and economic analysis by allowing anyone to post their views and analysis on stock prices and economic indicators.
New research by a group of academics from the University of North Carolina at Chapel Hill finds that sites like Estimize are giving consensus forecasts a run for their money—by exploring the accuracy of the Estimize platform, they find that their forecasters are significantly more accurate than the best Consensus forecasts, at least when it comes to U.S. non-farm payroll forecasts.
On average, consensus forecasts are more accurate than Estimize for the first release—and equally as accurate for the revised. But the source of accuracy, say the researchers, resides with a small group of “all-stars”—and, on that front, Estimize forecasters are significantly more accurate than the best consensus forecasters. While consensus forecasters are more accurate over short horizons, Estimize participants are more accurate when making longer horizon forecasts. And, when consensus forecasters are uncertain and herd together, Estimize forecasters appear to be more accurate.
The bottom line? Estimize could be a good way to break through the herd mentality that can show up in consensus forecasts—and it might be a more accurate way to capture long-term prospects.