We construct a method to solve models with heterogeneous agents and aggregate uncertainty that is simpler than existing algorithms; the aggregate law of motion is obtained neither by simulation nor by parameterization of the cross-sectional distribution, but by explicitly aggregating the...