They say "we find no statistically significant change in departure rates of males or females across firm hierarchies." Have not read it closely, but "no statis

They say "we find no statistically significant change in departure rates of males or females across firm hierarchies." Have not read it closely, but "no statistically significant" need not rule out the scenario that departures drive the whole effect. It's just saying that the error bar from counting departures is large enough to include zero, while the error bars on mean salary are smaller, because everyone has a salary but only a few leave.