Are sons and daughters substitutable? A study of intra-household allocation of resources in contemporary Japan

Gender inequality in educational attainment remains a salient feature of contemporary Japanese society. This inequality lies not at the high school level but at the university level. Equal numbers of men and women advance to high school, but a significantly lower proportion of women advance to four-year universities relative to men. Starting from government statistics which report that Japanese parents aspire to university education more for their sons than for their daughters, I argue that the gender gap in the university advancement rate in Japan stems in part from differences in how parents allocate resources within the household depending on the gender of their children. From the individuals' perspective, the gender composition of their siblings should therefore alter the ways in which resources are allocated to them. Using a 1995 cross-sectional sample of men and women between the ages of 20 and 70 in Japan, I test the null hypothesis that the marginal rate of substitution (MRS) between an additional brother versus an additional sister with respect to the probability that an individual advances to university or not is equal to unity, i.e. MRS = 1 implies that brothers and sisters are perfectly substitutable and that parents allocate resources equally amongst their sons and daughters. My results show that the null hypothesis is rejected for women but cannot be rejected for men: It is not sibship size per se that depresses women's likelihood of university advancement, but the number of brothers in the household. My findings lend support to the position that intra-household resources in Japan are likely to be allocated in favor of sons and away from daughters.