In the remainder of this chapter, we briefly describe what is known about educational differences in family behavior and how those differences have changed over time. We focus first on aspects of union formation and dissolution, then on aspects of childbearing, and finally on patterns of maternal labor force participation.
Union Formation and Dissolution
At first glance, research on marriage timing and the probability of ever marrying suggests that the pattern of diverging destinies observed in the U.S. and elsewhere is of little relevance in Japan. Research on marriages taking place in the 1970s and 1980s consistently found that women with higher educational attainment and higher earnings married later and were less likely to ever marry than their lower SES counterparts (Ono 2003; Raymo 2003; Tsuya and Mason 1995). This pattern is less consistent with diverging destinies than it is of a more “traditional” scenario in which women’s economic contributions to marriage were limited and in which men, even those at the lower end of the socioeconomic spectrum, were capable of supporting a family as the primary breadwinner. In this context, higher education enabled women to achieve a degree of economic independence that allowed some to postpone or opt out of the “the onerous status of the Japanese wife and mother” (Tsuya and Mason 1995: 156).
However, more recent research on marriages in the 1990s and 2000s suggests that this pattern is changing in ways that may be consistent with the general picture painted in the diverging destinies literature. For example, Fukuda (2013) finds that, in contrast to Ono’s (2003) results, earnings are positively associated with the transition to marriage among women born in the 1970s. Similarly, Fukuda and Raymo (2015) find that the negative educational gradient in women’s marriage has disappeared among recent cohorts. This reflects both the relatively high marriage rates of college-educated women beyond age 30 in recent years and the declining rates of marriage among women with less than a four-year degree in the late 1990s and beyond. In closely related research on employment status and marriage, several studies have found that men and women in nonstandard employment (a group that includes large proportions of the less educated) are less likely to marry than those in regular employment (Nagase 2002; Piotrowski et al. 2015). These patterns are suggestive of the growing economic difficulties of maintaining a breadwinner- homemaker marriage. Attitudinal data support this interpretation, with both husbands and wives indicating that they would like wives to work more hours than they currently do (Bumpass et al. 2010) and unmarried men indicating an increased preference for marriages in which their wives are employed (NIPSSR 2012: 62). In Sect. 3.3, we provided updated evidence on change over time in the educational gradient in marriage—focusing on the likelihood of early marriage. To what degree do we see an increasing concentration of early marriage among women at the lower end of the educational spectrum, as predicted by research on diverging destinies?