20(30)
6
Correlations between children and parents
in-law
So far, we have estimated the transmission from parents to their own
children. However, many people live in households with a partner. For
the socioeconomic status of the household, it also matters how much the
partner is absent due to sickness. Thus, the intergenerational transmission
might be even lower at the household level if there is a positive correlation
in sickness absence between a child and his or her parents-in-law.
We know that there are positive correlations between couples with respect
to education, income, intelligence, physical conditions and health, which
are all explanatory factors for the sickness absence level. Furthermore, the
results from intergenerational income mobility studies show that there is a
gender difference in this respect; daughters to a higher degree than sons
tend to live in households with incomes more similar to their parents
(Chadwick and Solon, 2002; Hirvonen, 2008). In order to shed light on the
intergenerational transmission in sickness absence at the household level,
we estimate the correlation between the children and their parents-in-law.
That is, we estimate model (1) but the parent’s sickness absence level is
replaced by the corresponding parent-in-law. The results are presented in
Table 6.
We do find a positive correlation between children and their parents-in-law.
As is the case with respect to income, the correlation is strongest between
sons and their parents-in-law, implying that daughters to a higher degree
than sons tend to live with a partner with a sickness absence level more
similar to that of their own parents. The correlations between daughters
and their parents-in-law are also positive and significant, but weaker than
the corresponding correlations regarding sons.
To conclude, people tend to live in a household with a sickness absence
level reminiscent of the sickness absence level in the household in which
they grew up. This result suggests that the intergenerational sickness
absence transmission at the household level is higher than the
corresponding transmission estimated at the individual level.
