15(38)
births between December and January is slightly smaller than usual in
1994/95,
5
and somewhat larger than usual in 2001/02.
Figure A2 in Appendix A presents the number of children born 17-30 of
June and 1-14 of July in 2004-2009. The difference between number of
children born in June and July each year varies between the years. The
difference in number of births between June and July 2008 is slightly higher
than usual, but not much. Thus, there may be tendencies of sorting in the
anticipated direction for all three reforms. However, differences are
marginal. , and similar variation exists in some of the non-reform years.
Investigating whether there are any differences between the samples in
observed characteristics (see Table A1 in Appendix A), comparisons of the
control and treatment groups show that they are similar to each other.
6
Thus, on the basis of these results, indications the reforms seem to be
exogenous events.
In this study we analyze the immediate response to the reforms. If a
reform changes behavior gradually, but nevertheless initiates a change, it
may be missed if we focus on the first parents to meet the reform (see the
similar argument in Ferrarini and Duvander 2010). Thus the long-term
effect of the reforms may be different from the immediate response
covered here.
5
Also see Ekberg et al (2004) who also investigate the number of births at the
introduction first reserved month.
6
There are however some significant differences between control and treatment
groups (see Appendix A). This comes as no surprise as we should expect that
control and treatment groups differ significantly in some respects (0.05 • the
number of covariates at the 5 percent level).
