Unconditional Quantile Regression Approach: Effects of Education on Housework Time in the US and Japan

Abstract

We compare the association between educational attainment and housework participation among single and married women in Japan and the US. Using the cross-sectional time-use diaries from the 2006 American Time Use Survey (ATUS) and the 2006 Japanese Survey on Time Use and Leisure Activities (STULA) and unconditional quantile regressions (UQR), we test whether educational attainment is associated with less time spent on housework in Japan compared to the US. We find that this assumption stands only for American women and non-married Japanese women. However, married Japanese women are unlikely to reduce participation in housework with an increase in their educational level. Married Japanese women are more likely to do more housework proportionately to the level of their education. The findings reveal the presence of a marriage penalty among highly educated Japanese women. In Japan, the institute of marriage places higher expectations regarding women’s housework participation on married women with higher levels of education, thereby penalising Japanese women with higher educational attainments. Our findings illustrate that the tenets of the resource-based and gender-centred frameworks developed based on the empirical findings in Western countries cannot always directly apply to the patterns observed in East Asia.