We model unemployment allowing workers to differ by comparative advantage in market work. Workers with comparative advantage are identified by who works more hours when employed. This enables us to test the model by grouping workers based on their long-term wages and hours from panel data. The...

Standard heterogeneous agent macro models that highlight idiosyncratic productivity shocks do not generate the near zero cross-sectional correlation between hours and wages found in the data. We ask whether matching this moment matters for business cycle properties of these models. To do this we...