Anna Linderoth

Abstract: This paper tests the hypothesis of reference dependence in men's parental leave take up. Using register data, I leverage two different reforms introducing earmarked provision for fathers. A key empirical difficulty in separating reference point behavior is that the earmark provisions also changed men's financial incentives to bunch around the value of the provision. I net out behavioral responses to the financial incentives by calculating and controlling for their size and exploiting the fact that financial incentives did not change for sub-groups of households. I find that the introduction of reference points via earmark provisions increased the average days of parental leave per man (the intensive margin) and the likelihood that the average man takes any parental leave (the extensive margin). In comparison, responses to financial incentives are modest. The reference dependence effect is large immediately after the reforms but diminishes with time as fathers increasingly take more leave than the earmark provision.

Abstract: Women and men tend to work in different workplaces and this gender segregation is an important contributor to the gender wage gap. This paper studies a new explanation for workplace gender segregation, namely that men leave the workplace when the share of women reaches a certain point. I draw on past research on occupation-level tipping to detect potential composition levels—tipping points—where the share of men in a workplace starts dropping discontinuously over time. This analysis uses Swedish register data for all small- and medium-size workplaces from 1986 to 2009 and is carried out separately by men's level of education (high or low). I find strong non-linear patterns for high-skilled men, but not for low-skilled ones. The distribution of candidate tipping points for high-skilled workplaces are centred around 25% to 35% female. The result of this paper highlight the importance of looking at other mechanism than gender differences in firm specific attributes as explanations for the gender segregation in the labor market. Furthermore, the findings emphasizes the potential importance of preferences of the dominant group. Using enlistment data I find a significant decrease of men's cognitive and non-cognitive skills at tipping point. I draw on my empirical findings to show how compositional preferences can be built into a simple model of self-selection. The model explain the observed negative selection of men at tipping by incorporating the prevalence of a team based setting in high-skilled workplaces.