Statistical decisions and policy choice
Harvard University, Cambridge MA
Investigators
Abstract
The research combines econometric theory with the economic theory of optimal policy decisions to develop a new model of policy choice. The key factor in the new model is that policy choice is modeled as a statistical decision, equating the negative of statistical loss and the overall social or business objective. The PI uses this new framework to analyze the structure of optimal policy choices and the optimal design of experiments that will guide future choices. The class of policy problems considered is broad enough to include both optimal taxation and the choice of inputs to some production functions. The PI is also considering decision problems where the loss or utility of a given policy depends on the full data distribution.
View original record on NSF Award Search →