Dependence uncertainty: a decision-theoretic approach

Abstract

We propose and axiomatize preferences on a product state space in light of uncertainty regarding the dependency of different payoff-relevant factors. Dependence structures allow to decompose probabilities and allow to pin down behavior towards dependence. The degree of dependence aversion is measured by dependence premia that are compatible with the encompassed MEU and smooth preferences on the dependence uncertainty set. A separation axiom clarifies when uncertainty about dependence breaks down and the decision maker treats factors as independent, making dependence neglect testable. We describe the dependence uncertainty set as a convex polytope and characterize their extreme points as maximizers of divergences. The model and its tools are applied to simple examples on climate change, insurance, and portfolio choice.