Complexity and Forecasting the Future

Predictions of the future are notoriously prone to error. Up to this point, two major accounts have dominated discussions of why that is the case.

The classic explanation, offered by William Ascher in Forecasting: An Appraisal for Policy Makers and Planners provides detailed evidence documenting 1) that the core assumptions incorporated into the forecasting model are far more important than the methodology used, 2) that 'assumption drag' (the continued reliance on old and outdated core assumptions) is responsible for many of the most spectacular failures of prediction and 3) that failure to incorporate expert judgment (e.g. through the use of a method that mechanically applies past trends and structural relations without incorporating 'intuition') typically decreases the accuracy of the forecast.

The second account, presented in Phillip Tetlock's Expert Political Judgment: How Good Is It? How Can We Know?, extends the emphasis on subtlety and complexity implicit in Ascher's recognition of the importance of expert judgment. Tetlock divides experts into two types -- hedgehogs who are poor at making accurate forecasts and foxes who are more successful. Low scorers look like hedgehogs: thinkers who “know one big thing,” aggressively extend the explanatory reach of that one big thing into new domains, display bristly impatience with those who “do not get it,” and express considerable confidence that they are already pretty proficient forecasters, at least in the long term. High scorers look like foxes: thinkers who know many small things (tricks of their trade), are skeptical of grand schemes, see explanation and prediction not as deductive exercises but rather as exercises in flexible “ad hocery” that require stitching together diverse sources of information, and are rather diffident about their own forecasting prowess.

Here is more from Tetlock himself:



Aside from the emphasis on complexity and subtlety, both these accounts share a materialist philosophy. They see forecasting as the act of understanding and predicting the empirical world of human activity and the institutions created, reproduced, or destroyed by that activity. The success of the forecast is wrapped up in the match between the sophistication and 'core assumptions' of the forecasting model and the material operation of the world.

The flip side of materialist philosophy is idealism -- the notion that ideas cause behavior rather than the other way around. This is the fundamental way in which complexity theorist John Casti's Mood Matters distinguishes itself from previous assessments of the practice of forecasting.

Here is some material from the book's website, followed by Casti's summary of the book's main thesis. There is an interesting correspondence between Casti's argument and the emerging recognition within the environmental community that effective climate change policy can't be forged around a doom and gloom philosophy (if you don't enact cap-and-trade, we're doomed) discussed here a few weeks ago.

Most people think that the outcome of elections causes the mood of the country to change. The opposite is true: The mood of the country determines the outcome of elections.

Most people think that a plethora of happy popular music on the charts makes the public happy and that a plethora of depressing popular music on the charts makes the public depressed. The opposite is true: A happy public pushes happy songs up the charts, and a depressed public pushes depressing songs up the charts.

Most people think that a productive economy makes people optimistic and that an unproductive economy makes people pessimistic. The opposite is true: Optimistic people make a productive economy, and pessimistic people make an unproductive one. Most people think that peace makes people content and tolerant while wars make people angry, fearful and patriotic. The opposite is true: Content and tolerant people make peace, while angry, fearful and patriotic people make war.

It sounds so simple, yet no one in the social sciences has made this case-until now. Mood Matters tells the story of why human events happen the way they do and not some other way, showing how it is the collective mood of a population, its social mood, that biases the events that we can expect to see. If you want to understand how information flows from the individual human impulse to herd together in groups to the overall social mood in a population that gives rise to events, read this book!