Uncertainty vs. Learning in Climate Policy: Some Classical Results and New Directions
Andreas Lange¹ and Nicolas Treich²

¹AREC, University of Maryland

²INRA, Toulouse, France

In this paper, we study and contrast the effect of uncertainty and learning in a two-decision model that encompasses most existing microeconomics models of climate change. We first consider the common expected utility framework: While uncertainty has generally no or a negative effect on welfare, learning has always a positive, and thus opposite, effect. The effects of both uncertainty and learning on decisions are less clear. Neither uncertainty nor learning can be used as an argument to increase or reduce emissions today, independently on the degree of risk aversion of the decision-maker and on the nature of irreversibility constraints. We then deviate from the expected utility framework and consider a model with ambiguity aversion. The model accounts well for situations of imprecise or multiple probability distributions, as present in the context of climate change. In general, ambiguity aversion leads to reduced emissions both in the uncertainty and learning cases that we consider. We finally discuss difficulties in applying non-expected utility models to a dynamic framework which requires updating of beliefs: they can generate time-inconsistent decisions and a negative value of learning.

Responsible for this page: Katja Scherbov
Last updated: 13 Feb 2007