Occupations and Parental Leave-Taking in Sweden

Helen Eriksson, Stockholm University, Demography Unit

Time allocations to the labor market are often understood through opportunity costs: the higher the value of an individual’s time in the labor force, the more that individual has to lose by spending time in other activities, such as taking care of a child. Literature on child care in general and parental leave in particular however poses two important puzzles to the opportunity cost premise. First, a higher education of the father, one of the most important predictors of labor market value, is across contexts and time connected with a larger investment in child care as well as parental leave. Second, a higher income of the father, one of the most important outcomes of labor market value, is also positively related to parental leave taking.

In this study, we explore occupational conditions of work as a possible explanation to why highly educated, high earning fathers are better able to take leave from work to care for their children. We investigate a number of work conditions that are associated with income levels: facilitating aspects such as job autonomy, capabilities to negotiate leave-taking and job security but also hindering aspects such as career costs and a demanding job. We study Sweden as one of the few contexts in the world in which fathers’ time allocations respond to the birth of a child. The number of parental leave weeks taken by the father and the mother in the first two years of the child’s life is analyzed using administrative register data for 29,559 parental couples having their first child in 2009. Detailed information on occupations of both mothers and fathers allow for the use of multi-level multi-membership models with each couple nested in 112 father and 111 mother occupations. Conditions of work are added as occupational-level predictors.


Administrative register data from the Swedish Social Insurance Agency provide data for all parental couples having a first child in Sweden in 2009. Two outcomes are studied separately; father leave weeks and mother leave weeks. Leave weeks are measured as the number of full-time parental leave weeks taken by the father or the mother only during the first two years of the child’s life.

Income was measured on the individual-level by the declared annual income to the Swedish Tax Agency in the year before the child was born (2008). Because considerable non-linearity was detected, income for both fathers and mothers were entered as deciles in the models. Although wages would be preferred, only annual income is available through the administrative registers. To allow for couple relative effects of income, income dependency was entered as the surplus share of fathers’ income to total income (ranging from -1 to 1) following the approach by Sorensen and McLanahan (1987). Income dependency was entered in quintiles as measures of fit provided strong support for non-linearity.

Career potential was measured as the difference between the 80th and 20th salary percentile by ISCED-97 educational level and field (275 unique combinations) through the 2009 Structure of Earnings Survey (cf. Hällsten 2010, Brandén 2013) (n=977,373). Dispersion sets limits to the possible salaries for individuals with a particular training, including the availability of managerial positions and their accompanying salary increases.

Two measures were drawn from the Swedish Labor Market Survey (n=206,400). For each of the 112 available occupational groups, job demand is measured as the share of workers reporting any overtime. The amount of overtime required is used as an indicator for the degree to which the worker may be replaced. Overtime suggests that the workload falls on the individual worker with limited possibilities to share work with others and so indicates the degree to which the individual worker may be replaced by someone else. Job certainty is measured as the share of workers holding a permanent contract (as opposed to a temporary contract). The share of workers on a permanent contract indicates the ease by which the employer may fill the position with other workers.

Two indicators of occupational facilitators to parental leave were drawn from the Occupational Information network (O*NET) classifications, the successor to the Dictionary of Occupational Titles (DOT). Developed by the United States Department of Labor’s Employment and Training Administration, the O*NET database provides numerical rankings for a number of cognitive, interpersonal and physical skills as well as characteristics of work for 802 occupations (Handel 2016). O*NET is based on the US Standard Occupational Classification (SOC) that correspond to the 1988 International Standard Classification of Occupations (ISCO-88). The Swedish occupational classification system SSYK96 also corresponds to ISCO-88 so conversions are made possible. Job autonomy is measured as an index built of the items ‘In your current job, how much freedom do you have to make decisions without supervision?’ (no, very little, limited, some, a lot of freedom), ‘How much freedom do you have to determine the tasks, priorities, or goals of your current job?’ (no, very little, limited, some, a lot of freedom) and ‘How automated is your current job?’ (reverse-scored: not at all, slightly, moderately, highly and extremely automated) (alpha 0.62). Capabilities are measured as the basic skills required for the occupation. Basic skills are those transversal skills that may be valuable for a larger part of the labor market. Basic skills are measured as an index comprised of items measuring ‘Reading Comprehension’, ‘Active listening’, ‘Writing’, ‘Speaking’, ‘Mathematics’ and ‘Science’ (‘What level of knowledge of *skill* is needed to perform your current job?’ measured on a scale of 1 to 7) (alpha 0.91).

The population under study is different-sex couples registered in Sweden and having their first child in 2009. First births were chosen to reduce the possibility of parents’ adjusting their labor market situation as a result of leave-taking. In order to study couple dynamics of taking leave from the workplace, only couples in which both parents were in the labor market was selected. Out of the 43,165 couples having their first child in 2009, 11,914 included one or both parents not registered as in the labor market in the Swedish Tax Register (in 5,670 couples only the father was in the labor market, in 2,800 only the mothers and in 3,744 neither the mother nor the father). In the couples in which both parents were registered as in the labor market, 1,721 couples (5.1 percent) were excluded because occupational information was missing from the registers for either the father or the mother. Exclusions of outliers were those 106 couples (0.03 percent) in which the father took more than 40 consecutive weeks of leave. The total selection of couples amounts to 29,559 couples.

Presented in Session 1097: Families and Households