The paper itself is much more nuanced. Their model predicts the results of Dube, especially in industries where few of the employees receive salaries near min w

The paper itself is much more nuanced. Their model predicts the results of Dube, especially in industries where few of the employees receive salaries near min wage. It also predicts the minimal disemployment effects (or non-existent disemployment effects, as in Dube) in the short run. Where it differs is in the long-run in industries where 1) many employees receive near min wage, and 2) capital can easily replace labor. So, Aaronson finds disemployment in low cost / fast food restaurants in the long run, but not in expensive sit down restaurants (which can absorb additional labor costs).