The Good Judgment Project: Identifying Power Constraints to Improve Accuracy in Geopolitical Forecasting

This paper sets out to explore the relationships between institutional constraints and predictability in geopolitical forecasting. Despite the increasing complexity of our world today, researchers have found that institutional rules and norms still function to influence human behavior, and, therefore, the presence of well-functioning institutions may lead to greater stability and certainty in predicting world events. Using forecasting data from the Good Judgment Project’s recent prediction tournaments, we test the change in predictability—including accuracy, confidence, and difficulty—against the experimental constraints of diplomatic ties, rule of law, effective democracy, trade dependence, and freedom of the press. Our hypothesis is that each of these institutions, together and in conjunction with one another, are effectively able to constraint power player political behavior and reduce uncertainty in the geopolitical realm. We find that all of the constraints, except for diplomatic ties, actually have a negative correlation with prediction accuracy. Democracy is the strongest negative correlation between the level of constraint and prediction accuracy. We propose that one possible explanation for this result is due to a potential quadratic relationship between democracy and predictability, such that countries who are transitioning from autocracy actually become less predictable than those that are autocratic. Further research in a larger sample set is needed to test this new hypothesis.