The Effect of Childcare Leave System on Fertility

Abstract

We analyze the effect of childcare leave on fertility both theoretically and empirically. We show theoretically that the childcare leave system has a positive effect on fertility, and examine this hypothesis using micro data. In addition, we discuss how the effect changes with other working conditions, because the decision of working women depends not only on the childcare leave system but also on other working conditions. Existing studies show empirically that a childcare leave system has a positive effect on fertility. However, this effect has not been analyzed theoretically. Moreover, there has been no empirical study on whether women who are entitled to take a childcare leave have a higher probability of childbirth. Our theoretical model offers the following three findings: (1) Both the introduction of childcare leave and the increase in childcare leave benefit increase the probability of giving birth. (2) However, the effect is small if the working hours of full-time workers are very long or very short. When the working hours are very long, women tend to quit jobs because it is difficult to come back to work after taking a childcare leave. When the working hours are very short, women do not need to take childcare leave. (3) When the working hours of full-time workers are shorter than the optimal working hours before the childbirth, but are longer than the optimal working hours after the childbirth, reduction in working hours increases the probability of giving birth. The empirical part of the paper finds the following three points: (1) Women who are working for firms that have a childcare leave system and are entitled to take the leave have a significantly high probability of giving birth than the other women. (2) Women who work less than eight hours on weekdays have a higher probability of giving birth than women who work more than nine hours on weekdays. Furthermore, the effect of childcare leave on the probability of giving birth is larger for women whose working hours are short. (3) The above effect may be overestimated because pregnant women may have selected shorter working hours. In order to solve this problem, we used working hours of the previous year as an explanatory variable. The estimation result show that the direction of the effect of working hours on childbirth is the same as the original model, the significance level slightly declined. The first two findings are consistent with the inference from theoretical model. The effect of childcare leave on the probability of giving birth is smaller for women whose working hours are long. Therefore,in order to increase fertility, working hours of full-time workers should be reduced, or at least working hours of workers who have young children should be reduced.