This study focused on the leaving parental home and first marriage to examine the household formation behaviors in recent Japan. The Third National Survey on Household Changes conducted in October, 1994 contains rich information on individuals' transitions among household positions. Using this data, life tables of leaving parental home by sex and cohort were calculated. The results showed that the age at which 50 percent leaves home decreased from 22.3 and 22.7 for the 1934-39 birth cohort to 19.7 and 21.7 for the 1949-54 cohort, males and females respectively. This trend reversed recently and the age at which 50 percent leaves increased to 22.4 and 23.8 for the 1964-69 cohort. Unlike other countries, males left parental home earlier than females for all the cohorts examined in this study.
Reasons for leaving home were examined to explain the difference by sex and cohort. It was shown that the recent delay was mainly caused by the decrease in the proportion of leaving home before marriage. For males, the delay in the first job taking due to higher education was also an important factor. There was a huge difference by sex on reasons for leaving home. While more than 80 percent of males left home for education or occupation before marriage, approximately a half of females stayed in parental home until marriage.
The hazard function of leaving home was combined with that of the first marriage to produce multi-state life tables. Four states that refer to the combined incidence of the two life course events were distinguished. The recent delay in leaving home has increased the proportion of unmarried children in parental home. The marriage without leaving home, which was the dominant life course pattern of heirs in the traditional stem family system, has been declining. Persons who left parental home before marriage usually live alone, with some exceptions of sibling-only households or institutions. For females, the proportion of this state decreased with the delay in leaving home before age 25 but then increased because of the later marriage in recent cohorts. For males, the proportion decreased before age 25 and increased after age 30. For 25-29 years old males, the opposite effects of staying in parental home and marrying late were balanced and little change were observed in the proportion of leavers without marriage.