Abstract
This paper introduces a decomposition of welfare assessments for general dynamic stochastic economies with heterogeneous individuals. The decomposition is based on constructing individual, dynamic, and stochastic weights that characterize how welfarist planners make trade-offs across individuals, dates, and histories. Guided by the compensation principle, it initially decomposes a welfare assessment into an efficiency and a redistribution component, while the efficiency component is further decomposed into (1) aggregate efficiency, (2) risk sharing, and (3) intertemporal sharing components. Five minimal examples and three applications illustrate the properties of the decomposition and how it can be used to draw normative conclusions in specific scenarios.