Catching Green Swans

Abstract

This paper argues that in order to be better prepared for future emergent crises, we need to move beyond the current scenarios based on frequentist statistical evidence, and adopt an approach that:

1. assumes a complex systems perspective to fully understand socio-environmental dynamics;
2. adopts an “ex ante” perspective on the trajectories of a system, looking for the emergence of novelty as the observer follows the arrow of time, rather than the currently used “ex post” perspective that aims to explain the origins of phenomena observed in the present, going against the arrow of time;
3. views unanticipated events as endogenously created, whenever an individual or group’s information processing is insufficient to understand, and deal with, such events.

A Bayesian learning approach is proposed to reduce uncertainty. Uncertainty is viewed as the uncertainty of the observer, rather than the uncertainty of the data. This has the advantage that such learning, as an attempt to increase understanding, can be initiated when insufficient data are available to use the frequentist statistical approach. The paper makes the argument for this in both a general and a mathematical language, and then gives an example how this approach can elicit thus far insufficiently understood dynamics, in the case of temperature statistics. To conclude, it proposes a Financial Investment and Decision Tool that is based on this approach.

Keywords: green swans, Bayesian statistics, complex systems, emergent novelty

JEL Classification: B40, C11, D81, G32, Q54