10 / 20
10(20)
3
Data and descriptive statistics
The analysis is based on register data from the SSIA. The population
consists of parents with children aged 1 to 11 years who applied for the
temporary parental benefit between October 1 2010 and December 31
2013. The SSIA’s treatment sample is randomized on the applications
and not on the parents. Therefore, parents who use the benefit more often
have a higher probability to be monitored over time. In order to deal with
this problem we select all parents in a given week (starting in October 1
2010) whose parental benefit claim was chosen to be monitored and select
a random sample of parents (two comparison parents are chosen to each
treated)
8
in the same week who were not selected to be monitored.
9
We have matched information on the socio-economic background of
the parents to the data. Furthermore the number of benefit claims and
paid gross and net benefit days for the period October 1 2009 to March
4 2014 is added. The SSIA’s register includes a rich set of socio-economic
background variables, such as gender, age, marital status, education
level, income, and the number and ages of the children. We also have
information on the parent’s sickness and parental benefit spells.
Figure 1.
The number and share of applications selected to monitoring
each week
8
We select only two matches to the treated to make the analyses sample
manageable.
9
Since the SSIA’s treatment assignment is not done on weekly basis a few (<1 per
cent of the claims) selected parents have more than one benefit claim during the
same week. For convenience we removed these observations from the estimation
sample. Sensitivity analyses show that the results are not sensitive to this sample
selection.