Uncertainty is critical to questions about climate change policy. Recently developed recursive integrated assessment models have become the primary tool for studying and quantifying the policy implications of uncertainty. The first wave of recursive models has made valuable, pioneering efforts at analyzing disparate sources of uncertainty. We decompose the channels through which uncertainty affects policy and quantify them in a recursive extension of a benchmark integrated assessment model. We argue that frontier numerical methods will enable the next generation of recursive models to better capture the information structure of climate change and to thereby ask new types of questions about climate change policy