Model uncertainty and policy choice: A plea for integrated subjectivism

• Models trade-off scientific values, contributing to prediction uncertainty.
• Scientific models may be transformed into decision-theoretic ones for policy choice.
• I analyze an example of this process from the history of economics.
• This suggests similar transformations for climate models will inform policy debates.

A question at the intersection of scientific modeling and public choice is how to deal with uncertainty about model predictions. This “high-level” uncertainty is necessarily value-laden, and thus must be treated as irreducibly subjective. Nevertheless, formal methods of uncertainty analysis should still be employed for the purpose of clarifying policy debates. I argue that such debates are best informed by models which integrate objective features (which model the world) with subjective ones (modeling the policy-maker). This integrated subjectivism is illustrated with a case study from the literature on monetary policy. The paper concludes with some morals for the use of models in determining climate policy.