@article{oai:minpaku.repo.nii.ac.jp:00004167,
 author = {木下, 太志 and Kinoshita, Futoshi},
 issue = {4},
 journal = {国立民族学博物館研究報告, Bulletin of the National Museum of Ethnology},
 month = {Mar},
 note = {Reflecting recent developments in demographic studies, the first part
of this paper emphasizes the importance of the role cultural anthropologists
can play in this field. The results of major demographic
studies of recent years, such as the Princeton Fertility Project, the World
Fertility Survey and those by the Cambridge Group for the History of
Population and Social Structure, throw doubts on the validity of the
demographic transition theory which has been the single dominant
theory in demography for almost half a century. Drawing upon the
modernization theory, the demographic transition theory postulates that
demographic behavior, particularly fertility, is strongly influenced by
socio-economic factors such as industrialization, urbanization and
literacy rates. Recent empirical studies, however, fail to find such a relationship.
Instead, they find that demographic behavior is determined
more by cultural factors such as religion and ethnicity. It is in this context
that I emphasize the role of cultural anthropologists in demographic
studies which have been long dominated by sociologists and economists.
The second part of the paper examines a somewhat specific, but
crucial, issue to Japanese historical demography. This is the issue of
the underregistration of births recorded in shumon aratame-cho (SAC) ,
population registers most frequently utilized by researchers. This has
been a serious problem since it prevents accurate estimates of basic
demographic indices such as fertility and infant mortality. In solving
this, this paper employs a microsimulation approach. This approach
has several advantages over macrosimulation. The most important is
that with microsimulation, we can consider relatively easily in our model
such critical factors as seasonal fluctuation of births and monthly death
rates of infants.
The model for reproductive process used in this paper and its computer
program are based on studies by the Institute of Population Problems
and those by Bongaarts and Potter. For input data for the simulation,
data from the Tokugawa period is used wherever information is
available; otherwise, I use data of the Meiji and the Taisho periods.
The result of the simulation reveals that the degree of underregistration
of births in SAC ranges from 82 to 88 per cent, depending on infant
mortality levels, and thus 12 to 18 per cent of births were never recorded.
This leads us to conclude that birth rates calculated directly from
SAC need to be multiplied by 1.15 to 1.22 in order to obtain accurate fertility
estimates of Tokugawa peasants.
Another important finding from the simulation concerns the range
of random fluctuation in fertility due to a small population size. With a
cohort size of 25 persons, the simulation yields a standard deviation of
0.25 births in total marital fertility rate (TMFR) between five runs. The
comparable figure goes up to 0.7 births with a cohort size of 10 persons.
This will give us a measure of variance when dealing with the fertility of a
small village population in the Tokugawa period.},
 pages = {877--919},
 title = {記録されなかった出生 : 人口人類学におけるシミュレーション研究},
 volume = {21},
 year = {1997},
 yomi = {キノシタ, フトシ}
}