This paper aims to evaluate the results of Taiwan Census 2000, comparing its census and registered populations. A close analysis has shown that when the number of aliens and people who were outside of the country for 3 months or more are excluded respectively, the number of people surveyed in the census was 157,716 less than the registered population. This accounts for a difference of only 0.7% of the registered population.
However, smaller local administrative units such as townships, counties and cities, impacted by constant institutional changes, Industrialization and urbanization, have seen quite substantial differentials between the census and registered populations. The number of males in the census, for example, is 414,665 less than the registered population. At the same time the number of females in the census is 572,181 more than the registered population. The former reflects the results of institutional changes while the latter impacts of urbanization. Nonetheless, when gender and age are controlled simultaneously, the direction of differences between genders remains unchanged. Nevertheless, the differences between genders have increased for some age groups – the young male census population is 5 to 8% greater than the registered population with most young males being single. Among those who are married, the differences between the census and registered populations are smaller. These differences are mainly due to the impacts of urbanization, education, employment, and national service. Finally, in terms of educational composition between the two populations, institutional changes have seen their differences as high as 76% or more at both primary and junior high school levels.

Using data from the Surveys of Family Income and Expenditure between 1991 and 2000, this paper examines the changes in demographic and economic features of single-person households in Taiwan. Specifically, it focuses on households that are vulnerable and in poverty occupied by a single-person aged 65 and above.
According to estimates from samples across Taiwan, the proportion of single-person households in the country has raised from 5.83% in 1991 to 10.82% in 2000. Similarly the total number of single-person households has also increased from 300,000 to 710,000 between 1991 and 2000. Furthermore, 40% of these householders are elderly people. However, at the beginning of the 1990s, about 70% of all single-person householders were male. They have since dropped to about 50% in 2000.
This research employs the official poverty line as well as other relative poverty measures to indicate the poverty situation of singleperson households of all ages. Not surprisingly, the poverty rates among the aged are higher than the younger cohorts, except for those between 55 and 64 in recent years. Among the aged single-person householders, females were more likely to experience economic plights than males. With the use of logit models, the net effects of gender and education are further confirmed.

This paper applies spatial autocorrelation statistics to analyze the spatial differentiation and transition of aging population during the period of 1980 and 2000 based on data collected from 106 administrative units in Central Taiwan.
Moran’s I statistic is employed in this research to measure the spatial association of population aging. Results have shown that the distribution of aging exhibits a spatial concentration and expansion over time. In addition, spatial autocorrelation correlograms based on Z(I) statistic have also shown that the distance among spatial clusters tends to decrease during the study period. Finally, this study has identified the actual clusters according to local index of spatial autocorrelation. It suggests that most clusters of aging in the study area appear mainly in mountain or rural areas where slower development have occured, while clusters of younger population, on the other hand, concentrate mostly in urban areas. The fact that the number of clusters of aging population increases and the value of Z(I) statistic decreases throughout the study period suggests that distribution of aging population becomes less concentrated over space.
It can be concluded that as population aging in Taiwan becomes more common in recent years, the aging population tends to be more evenly distributed over space. This finding is quite different from conclusions drawn in earlier studies. The distribution of the younger population, however, has shown a tendency towards spatial concentration within a few relatively isolated clusters.

Using data from the 1998 Trends of Social Development Survey, this paper examines the reasons and schedules for young people leaving home between 1950 and 1997 in Taiwan. A multiple decrement approach is also employed to establish the life tables of these young people for leaving home. The results have indicated that for men, the main reasons for them to leave home are employment, marriage and limited space at home with their parents. However, less are leaving home for employment while more are doing so for the reason of limited space. For women the most important reason of moving out has always been marriage, with more than 70% doing so each year. Comparing to the past, for both men and women, their schedules for leaving home have been delayed and the age-specific occurrence rates for leaving home are also decreasing. It is most likely that adult children will continue to stay with their parents in the future.

In recent years because of the aging population, there have been great changes in Taiwan’s fertility and mortality rates. As the proportion of elderly people aged 65 and over increase dramatically from 2.6% in 1965 to 8.8% in 2001, the decrease of fertility rate is most significant. In 1961, the total fertility rate was 5.58. By 1981, it was dropped to 1.67 and in 2001, further to 1.4, a reduction of almost 70% within 20 years.
This paper examines the fertility pattern in the Taiwan area. In particular, it aims to review the various fertility models and seek for the model that is most appropriate for describing the situation in Taiwan. The models considered are Gamma function, Gompertz function, Lee-Carter method and individual group estimation. Data from 1951 to 1995 is used as a pilot for verifying the model which has the best fit for data gathered from 1996 to 2000. It is found that individual group estimation and the Lee-Carter method have least errors for predicting total fertility rates; while individual group estimation and Gompertz function are more effective for predicting age-specific fertility rates.