Chart of the Day: Length of maternity leave as a signal of productivity
Today’s Chart of the Day comes from Linh Tô’s paper “The Signaling Role of Parental Leave”. She builds a model where women use the length of leave they take after childbirth as a signal of future productivity to employers. Those that are more productive on average signal that by coming back to work sooner. From this model, she has several predictions of how leave length and earnings will change after a parental leave reform changes the length of maximum paid leave.
She splits the sample into three groups, based on predicted behavior.
Those that would take less than the maximum amount of leave both before and after the reform
Those that would take the maximum amount of leave before the reform, but less than the maximum once the maximum was increased
Those that would take the maximum amount of leave in both regimes
If the amount of leave taken is a signal, rather than just the optimal amount of leave, then those in group 1 could increase the amount of leave they take after the reform, but maintain their signal.
Tô uses two maternity leave expansions in Denmark in 1984 and 2002 to evaluate these predictions. Figure 3 shows that some of the distribution below the pre-reform maximum leave cutoff moves up, particularly with the 2002 reform.
In addition, group 2 could not provide a signal of their post-child productivity before the reform, but they can afterwards. So, if leave taking is a signal, the earnings of group 2 should go up relative to the earnings of group 3. Tô shows this is true for mothers predicted to be in these groups in Figure 4 (and 5 for the 2002 reform, but Figure 4 gives the idea).
In my paper with Nichole Szembrot, we compare women who take short periods out of the labor force after childbirth versus those that take longer periods in the U.S. and don’t find long-run earnings differences. But we don’t look in detail at maternity leave length. And the U.S. doesn’t have a maximum paid leave length to “pool” around. With unpaid leave from FMLA, we are probably seeing a different sort of selection into those who return to work early versus those who stay out longer, based on the ability to afford longer leave periods.
My Reading Notes: