The Third National Survey on Household Changes was conducted by the Institute of Population Problems on the 15th of October in 1994 to clarify how each household was modified and what were the factors that caused the changes. The survey was carried out for the nationally representative sample of household heads and individuals aged 18 and over. The valid samples are 8,578 households and 20,788 individuals. The survey clarified the four elements of the household changes in recent Japan as described below.
1. Household formation behavior in Japan is getting more and more nuclearlized, which started since around 1960.
(1) The proportion of the elderly coresiding with their child is smaller for the younger cohorts, and the proportion is slightly above a half for the youngest elderly (65-69 years old).
(2) The proneness both for men and women to choose coresidence with their own and spouse's parents is lower for younger cohorts' aged 30's and 40's.
(3) Of all households surveyed, 12.8 percent were newly constituted during the last five years, and only 3.1 percent were succeeded from older generation to their offspring. This means that the formation of households in Japan has become a predominantly nuclearized one where every household has its starting point and end point and is not succeeded over generations.
(4) During the last five years, 11.1 percent of households lasting for five years or more experienced the leaving of a member(s) for the reason of marriage. On the other hand, only 3.8 percent experienced the inclusion of a new member(s) through marriage. This means that marriages, the main commencement of households, occur predominantly as the formation of new households but far less as the extension of households.
(5) The proportion of those who experienced the leaving parental home has been maintained to be about 80 percent for cohorts born after World War H in spite of the decline in the proportion of eldest sons among those cohorts. This means that the proneness to leave parental home has been increasing for eldest sons.
2. In addition to the changes in household formation behavior, the increase in middle- and old-age household members caused by the population aging has accelerated the reduction of the household size in Japan.
(1) Of those households lasting for five years or more, 10.9 percent experienced the birth(s) of a child during the last five years. On the other hand, 22.0 percent experienced the leaving(s) of a child for such reasons as educational upgrading and getting a job (10.9 percent), marriage (11.1 percent), and so forth. It can be said that more households are at the late stage of family cycle where grown-up children leave parental home than those are at the earlier stage.
(2) Most of the changes in household composition during the last five years were those from and into the "household composed by parent(s) and child(ren) " The most prevalent one was the transition from the "household composed by parent(s) and child(ren)" to the "household of a couple only" namely the transition into "empty nest". This change was experienced by 4.9 percent of households lasting for five years and more.
3. After the World War II , birth rate, as well as death rate, rapidly declined and demographic transition in Japan finished around 1960. Thus, the sibling size is smaller for younger cohorts born before 1960. This decrease in sibling size is supposed to have raised the possibility for these cohorts to coreside with parents and to have restrained the decline in coresidence with parents or the rise in leaving parental home.
4. The rise in age at leaving parental home could be attributed to such changes as the delay of marriage and the higher educational enrollment after the World War II. Concerning the reason of departure, more and more young males and females leave parental home at their educational upgrading. For young males born after 1955, the leaving for educational upgrading has come to be almost abreast with the most prevalent reason of departure - getting the first job. But for the 1960-64-birth cohort, the percentage of departure on the ground of education did not increase because of the stagnated college enrollment rate after mid-1970.