Future work: thinking about high school tracking/testing, modeling the education supply side (cost of capacity), and thinking more about the extensive margin (college or not).

3. Relaxing these constraints following a shock allows workers to better reallocate---especially those with low earnings potential.

2. We examine the GDP consequences of education systems that rely on capacity constraints and rationing. We find small GDP bumps from relaxing capacity constraints (caveat: we don't have costs).

1. Field choice changes both aggregate and distributional predictions of GE models. Aggregate effects are larger (sure, you can reallocate), but distributional consequences also manifest differently.