dhruvalb / macss_thesis

Spatial analysis of patterns and factors for distribution of aging population in Japan

M.A. in Computational Social Science: Thesis

Spatial Patterns and Factors in Aging Population Distribution in Japan

Abstract

Japan, a leading site of productivity and economic growth, is facing serious challenges due to a rapidly aging population. To understand and address the uneven distribution and concentration of aging population, demographic transition theory is considered. An exploratory spatial analysis using the data published by Japanese government for each of the 47 prefectures (administrative units), reveals that highest concentration of aging population is in less industrialized, rural regions and regions with prominent industries and urban regions have lowest concentration of aging population. A model with direct factors for population, birth rate, death rate and migration, explains ~92% of ratio of population over 65 but is diagnosed with a spatial error, indicating missing variables or correlation in the error terms. On the other hand, a model for indirect factors of population aging has dependent variables, such as, densely inhabited districts characterizing urban areas, fertility rate showing changing social norms as well as income, education, industry and job-related migration, shows presence of a spatial lag. After accounting for that, the model explains ~63% of ratio of population over 65. This is study shows the challenges of working with a limited number of administrative units in spatial regression analysis and provides a foundation for further research in role of urbanization in the distribution of aging population in Japan.

Quick Links

• Data Raw and processed data from Japanese Government Statistics database
• Code Code files for data prep and preliminary analysis
• Final Project Report