16(38)
5
Results
5.1
Descriptive results
In this section we present descriptive results by presenting comparisons of
parental leave days in the control and treatment groups 20 months after
the introduction of each of the three reforms. We restrict our analyses to
the earnings-related parental leave days and exclude the three-month flat
rate days as they are not relevant to any of the reforms.
In Table 1 the average use of parental leave days in the control and
treatment groups is presented 20 months after the introduction of each
reform.
7
The first reserved month had a large impact on both fathers’ and
mothers’ use of parental leave days. Fathers’ use increased from an
average of 23.4 days in the control group to 31.2 days in the treatment
group. Mothers’ use of parental leave instead declined, from an average of
317.5 days in the control group to 290.5 days in the treatment group. The
obvious explanation is that the first reserved month restricted the
maximum days any of the parents could use by 30 days.
Also, the introduction of the second reserved month shows a significant
effect on fathers' use of parental leave. Twenty months after the reform the
treatment group used on average 42.9 days, whereas 37.3 days were used
by fathers whose children were born just before the reform. As the parental
leave was extended by 30 days at the same time as the second reserved
month was introduced, mothers used more days after the reform. Mothers’
used days increased from 268.0 in the control group to 275.4 in the
treatment group.
In contrast to the changes relating to the first and second reserved months
it is not possible to find any statistically significant changes in average
number of used parental leave days after the introduction of the gender
equality bonus. Fathers’ use of parental leave was on average about 48-49
7
We have also conducted OLS regression where the dependent variable is number of
used earnings-related parental leave days, with a dummy variable included in the
model controlling for being in the treatment group. The models are conducted with
and without control variables. The results in these models are similar to the change
in mean values, as may be seen from the table:
Difference in used earnings-related parental leave days between treatment groups and control
groups
Observed
difference
Regression
without control
variables
Regression
with control
variables
First reserved month
Men
7.9***
7.0***
6.8***
Women
-27.0***
-27.2***
-26.6***
Second reserved month
Men
5.6***
4.8**
5.0**
Women
7.4***
9.0***
8.5***
Gender equality bonus
Men
-1.0
-0.5
0.2
Women
-0.2
-2.1
-3.1
*** Significant difference (1 percent level) between control and treatment groups.
** Significant difference (5 percent level) between control and treatment groups.
Huber/White/sandwich estimator of variance.
