I will in this thesis examine the family gap in pay (i.e. differences in hourly pay between mothers and non-mothers) among Norwegian women.While women in general make up a substantial part of the workforce today, a large literature shows that women with children have lower hourly earnings than women without children even with same type of education and experience. Evidence of a family gap in pay has been found for the United States, the United Kingdom, Germany and also for Norway.
Conventionally, five explanations have been given to explain the lower wages of mothers. (1) Mothers may earn less because they lose experience that are relevant for their work effort, (2) they may be less productive at work due to childrearing,(3)they may choose jobs more compatible with childrearing but at the expense of higher wages, (4) employers might discriminate against mothers, (5)or it could be that non-mothers differ from mothers in ways that are relevant for wages and careers. However, which of these explanations to put most emphasis on is still an unresolved matter in the literature of the family gap.
The available data was obtained from Statistics Norway and the Confederation of Norwegian Enterprise (NHO). It covers the private business sector in Norway for the years 1980-1997. The data contains information about wages, establishments, occupation, age, education, family status, number of children (including when they were born) and more. The feature of the data (matched employer –employee data) allows me scrutinize the causes of the family gap in ways that have not been done before.
One important aspect concerns that of employer discrimination. With the available data I can see if mothers are being paid less than non-mothers once they perform the same work for the same employer. Another central aspect concerns the time-frame of the data. A presumption in the exciting literature of the family gap is that the lower wages for women with children can be explained by the lack of well-developed family policies in such countries. The data I have available covers the years 1980-1997, a period with significant expansion in family policies in Norway. This allows me to see if the penalties to motherhood have changed over time, and to see if the family policies implemented in this years have had any of the intended effects.
I will use Ordinary Least Square Models (OLS) to examine the family gap in Norway with the natural log for wage as the dependent variable. In my analyses I will use different specifications of the OLS model, which will take into account the different levels of the data which arises with matched employer-employee data.The main finding from my analysis is that is has occurred a great change in the situation for mothers in the private sector. The critical disparities in wages between mothers and non-mothers evident in the earliest years are severely reduced by the end of study. However, mothers still earn on average less than non-mothers, and these differences also increase with how many children you have. The main explanation for this is that mothers are sorted on different occupations and different establishment where lower wages prevails compared to non-mothers. Another important finding is when mothers and non-mothers actually work for the same employer and in same establishment, they receive same pay. As such within-job-wage discrimination is not what constitutes the family gap in pay in Norway.
Whether this differential sorting is due to some compensating differentials i.e. mothers seek this positions because they are more easily combined with childrearing but at the expense of lower wages, or if it is a result of employer discrimination is difficult to determine. One could imagine that employers in the hiring situation channel mothers in to lower paid occupations, or give mothers less promotions compared to non-mothers. But also mothers may choose to spend less time at work and /or give less effort at work when they have children. In order to settle such a discussion more research on how family life and work life intersect is needed.