It is apparent that the household size has shrunk down considerably in the long run during 65 years' time span, but some may wonder why the process of shrinkage in the household size has been so slow in comparison with the fertility decline occurred roughly at the same time. For example, total fertility rate had substantially been reduced from 3.65 in 1950 to 2.00 in 1960. On the other hand, however, the average size of household diminished from 5.02 to 4.52 only by 0.50 in term of persons per household between the same years. By the same token, for the Republic of Korea, while total fertility rate was reduced from 4.47 to 3 .33 between 1970 and 1975, the average household size shrinked from 5.2 to 5.1 only by 0.1 person per household. It seems that there always exist some kind of demographic lags, so that the process of reduction in household size does not necessarily follow the trend in decline in fertility in an immediate sequence. Part of explanation may be sought in the mortality reduction taking place at the same time, the mortality reduction which facilitates enlarging the household size in a completely opposite direction to the effect of fertility decline under the circumstances where fertility does not drop. The substantial reduction in child mortality occurring in the early cycle of demogaphic transition precisely means an increase in fertility in Japan. Between 1920 and 1935, the size of household increased appreciably and this phenomenon is principally attributable to a reflection of mortality improvement. In Singapore, by the same token, between 1957 and 1970, the household size in-creased sizably from 4.8 to 5.4 and this enlargement may at least partly be explained by mortality reduction.
Nevertheless, when demographic evolution proceeds, effects of fertility decline have been captured sooner or later in the household size and structure when the mortality decline has lost its momentum and the rise of life expectancy has hit the plateau. The diminution of household size has started out and then it precipitates. The average size of households for Japan is now 2.84 and will soon become 2.6 by 2010 according to the projections (Institute of Population Problems, 1995).
According to a multiple regression analysis performed by the United Nations Population Division some years age (United Nations, 1969), the shrinkage of household size was mainly attributable to the reduction of fertility. This relationships may still be holding at present. However, the increasingly significant process of nuclearization or nucleation of the family due to mortality reduction, general rise in level of living, urbanization, industrialization, etc., would have also been fairly important for contributing to determine the average household size. This is again not a very recent one, but according to a decomposition study for Japan attributing to various factors the difference between average household sizes in different years, the effect of fertility decline explains 70 percent of the contraction in the average household size from 1955 to 1965; 20 percent were attributed to internal migration and the remaining 10 percent attributed to the nuclear fission, which in turn presumably originated from economic and social factors, notably from the rise in per capita income, urbanization, enhancement in education, and the emergence of individualism, etc. (Kono, 1969).
