11(30)
representative of the individual. For each parent and child, we calculate the
annual average number of days absent due to sickness over the respective
observation period.
17
In order to describe the importance of transmission depending on the level
of sickness absence among the parents, we estimate the parent–child
correlation in different parts of the sickness absence distribution separately.
We also control for whether the parents have any sickness absence or not
when estimating the transmission from parents with high levels of sickness
absence.
4.4
The study population
The study population consists of 238,599 sons and 224,385 daughters and
their mothers and fathers. Table 1 presents descriptive statistics of this
study population.
The study population includes children between 32 and 36 years old and
their parents, who are somewhat older: on average 44 and 46 years old for
mothers and fathers, respectively.
18
The mothers are on average younger
when entering parenthood than the fathers.
Daughters have lower labour earnings than sons; daughters’ earnings are
around 65 per cent of sons’ (SEK 149,282 and SEK 229,474). This gender
gap is somewhat larger than the average gender gap in earnings among
men and women in Sweden. An explanation is that the daughters in this
study are of child-bearing age.
In addition, in the parent generation, mothers have lower earnings than
fathers; the mothers’ labour income is on average 57 per cent of the
fathers’. This can be explained by the fact that women to a larger extent
than men work part-time or not at all.
Ideally we would only include those eligible for sickness benefits in our
study population. However, in our data, we cannot distinguish between
unemployment and voluntary non-participation in the labour force. In
order to shed light on this issue, we present the share without any labour
income at all (during all six years observed). This share is lowest among
the daughters (3 per cent) and highest among the fathers (5 per cent).
However, the large majority of people in all the groups have some labour
market income during the observation period, suggesting that they are
eligible for sickness benefits.
In our study population, there is both a gender gap and a gap between the
generations in our rate of sick leave; half of the mothers and the daughters
and at least 25 per cent of the fathers and the sons have days with
sickness absence.
In order to gain a better understanding of what our measure of sickness
absence captures, we also present the number of days with sickness benefit
and disability pension separately in Table 1 for all our sub-groups. The
17
To observe sickness absence during several years and calculate a mean value is
analogous to how a representative income is calculated within the literature on
intergenerational income mobility (see, for example, Björklund and Jäntti, 2009).
18
The age is observed in the last year of the respective study period, namely 1991
for the parents and 2008 for the children.
