This article evaluates the relevance of a popular emphasis on the benefits of extended coresidence with parents as an explanation for the trend toward later marriage in Japan. Estimating hazard models for the transition to first marriage separately by gender, living arrangements, and birth cohort, I find that the trend toward later marriage is indeed more pronounced among young men and women living with parents. Conditional on coresidence, however, there is little evidence that access to parental provision of financial resources and domestic services is negatively associated with marriage. I suggest alternative scenarios in which parental resources may contribute to later marriage by facilitating children 's independence rather than by keeping them at home.
Key Words: coresidence, Japan, marriage, nest-leaving, parental resources.
Although links between nest-leaving and the transition to adult roles among young Americans have been studied extensively (e.g., Avery, Goldscheider, & Speare, 1992; Buck & Scott, 1993; Goldscheider & DaVanzo, 1989), relatively little is known about these relationships in other countries. There appears, however, to be a growing interest in international research on the nest-leaving process (Goldscheider, 1997). Interestingly, work from an emerging literature on nest-leaving patterns in southern European countries suggests that, in contrast to the United States where the trend toward earlier nest-leaving has contributed to later marriage (Goldscheider & Waite, 1987), delayed nest-leaving is an important component of the trend toward later marriage in Italy and Spain (Billari, Castiglioni, Martin, Michielin, & Ongaro, 2000; Cordon, 1997; Rossi, 1997). Although less visible in the academic literature, a similar emphasis on delayed nest-leaving as an explanation for later marriage currently enjoys considerable popularity in Japan (see Butler, 1998, and Orenstein, 2001, for summaries in English). Proponents of this popular explanation for delayed marriage suggest that extended coresidence with parents has emerged as an advantageous alternative to early marriage as parents have become increasingly willing and able to provide their children with economic support and domestic services (Miyamoto, Iwakami, & Yamada, 1997; Yamada, 1996, 1999).
Despite its wide popularity, there is little evidence either to support or refute the hypothesized link between later marriage and an increase in the advantages of extended coresidence with parents. The goal of this article is to provide the first empirical analysis of the relationship between premarital living arrangements, the characteristics of unmarried men and women and their families, and the transition to first marriage in Japan. Following a brief overview of trends in Japanese marriage, I elaborate the extended coresidence hypothesis as an explanation for these trends. I then examine data from a large, nationally representative survey to evaluate the implications of this popular, yet untested, explanation for the trend toward later marriage.
MARRIAGE TRENDS IN JAPAN
Table 1, which presents both mean age at first marriage and age-specific proportions never married at each census year between 1970 and 2000, illustrates the trend toward later marriage. Over this 30-year period, the mean age at first marriage increased from 24.2 to 27.0 for women (top panel) and from 26.9 to 28.8 for men (bottom panel). Even more striking are changes in the age-specific proportions never married, particularly for women in their 20s and men in their late 20s and 30s. For example, whereas only one in five women aged 25-29 in 1970 had yet to marry, over half of similarly aged women in 2000 had never married. Among men in this age group, the proportion yet to marry increased by 24 percentage points. For men in their early 30s, the proportion never married more than tripled, reaching two fifths by 2000.
Although marriage behavior in most industrialized societies is characterized by similar changes, the implications of later and fewer marriages are particularly important in Japan. …