(Bloomberg View) -- The field of weather forecasting holds a lesson that economists might do well to heed: Stop being so sure of yourself.
The reputation of the economics profession suffered after the 2008 financial crisis, in large part because its practitioners were so overconfident. They have since been working to forge more realistic models of the economy, and re-examining some of their basic thinking about markets, trade and human behavior. In doing so, they might take some hints from weather scientists, who made a decision thirty years ago to be less, not more, precise. This has paradoxically made their forecasts more trustworthy.
In a new paper, Oxford University physicist Tim Palmer sheds much-needed light on this fascinating chapter in the history of weather prediction. He argues that the key was a philosophical shift towards embracing perpetual doubt -- seeing it not as a threat to knowledge, but as the very essence of it.
The weather is hard to predict in part because the atmosphere is chaotic. It’s subject to the famous Butterfly Effect: A tiny change in the state of the system can very quickly send it down a totally different path. Since the 1960s, scientists have known that such chaos makes prediction impossible beyond about 10 days, and prone to huge errors over much shorter times. By the early 1980s, they were doing the best they could with supercomputers. They would start with the most precise data available on the current state of the weather, then simulate the laws of atmospheric physics as accurately as possible, running a mathematical atmosphere forward to make a best guess of how things would be a few days ahead.
Around 1985, Palmer and other scientists in England had a subversive thought: Maybe spending all their resources on one “best guess” was a mistake. After all, what good was it if they had no idea how accurate it might be? The weather could still end up being completely different. Much better, they reasoned, to run a number of less-precise simulations and get some insight into the range of possible outcomes. This idea of “ensemble forecasting,” Palmer recalls, provoked strong resistance from experts accustomed to the idea that science should always aim for maximum precision. They worried that the new approach might undermine confidence in weather prediction itself.
The opposite has happened. The ensemble approach has helped improve accuracy: Scientists can now predict weather as well over 4 days as they used to over one day. This has increased public confidence in forecasts, which now inherently include a range of possible outcomes rather than pushing false belief in a single outcome. Oddly, better knowledge has come about by emphasizing doubt and uncertainty at every step.
This can work in economics as well. The Bank of England, for example, has adopted the ensemble approach in its forecasts, always laying out a range of possibilities rather than just one prediction. Economist Roger Farmer of the University of Warwick has been working to bring ideas from chaos theory and physics into macroeconomics, to focus better on inherent sources of instability. Harvard economist Dani Rodrik has emphasized the vast uncertainty surrounding topics such as the effect of free trade agreements, which typically involve a lot more than just reducing tariffs.
Such an emphasis on uncertainty could help economists regain the public’s trust. Unfortunately, too many still project unwarranted certainty, taking public positions based on hugely oversimplified models from introductory economics classes. That’s a pity, because less confidence now likely means more credibility later.
This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.
Mark Buchanan, a physicist and science writer, is the author of the book "Forecast: What Physics, Meteorology and the Natural Sciences Can Teach Us About Economics."
For more columns from Bloomberg View, visit http://www.bloomberg.com/view.
©2018 Bloomberg L.P.