Abstract
This statistical analysis study clarifies the characteristics of housing market areas and its transition on a national scale. It aims to confirm the basics in order to examine the actual and ideal state of supply and demand of housing on a municipality level. The author has clarified through his studies thus far, the formation process of housing market areas during the past 50 years after World War II in Japan. As a result, 550 housing market areas have been identified in 2000.
Based on these areas, this paper classifies the transformative process of the composition of household style in these areas by cluster analysis and examines the characteristics of such categories. The index used in the cluster analysis is the ratio of owned houses, publicly owned rented houses, privately owned rented houses and issued houses within the following three representative points. The three representative points are set based on the year 2000 (just before the major municipal mergers), 1955 (when the statistical data started to become complete), and 1980 (a midway point).
The result of analytical calculation leads to four (I－IV) large classifications in more than 50 percent of the maximum distance and 10 patterns (A-J) in more than 25 percent of housing market areas. These patterns are classified by the ratio of owned houses and its changes, and are characterized by the transforming ratio of issued houses, privately owned rented houses, and publicly owned rented houses.
Large classification I: This classification has the lowest level of owned houses compared to other large classifications. Rented housing (both private and public) is at its highest rate and the ratio of issued houses decreases while the ratio of owned houses increases. There are four patterns (A-D) within this classification.
Large classification II: The ratio of owned houses is higher than Classification I, but with little transition. The ratio of issued houses is low with little transition. The ratio of privately owned rented houses is equal to Classification I and the ratio of publicly owned rented houses is slightly smaller. There is only the pattern E (78 areas) within this classification.
Large classification III: The total ratio of owned houses and issued houses is far higher than the above two classifications, although it has decreased, the ratio of publicly owned rented houses has increased. Here there are two patterns: F (13 areas) and G (62 areas) depending on the level of issued houses and publicly owned rented houses.
Large classification IV: This classification has the highest level of owned houses, although its ratio has a tendency to slightly decrease. This has the largest number of areas in medium to small local areas. There are three patterns within this classification: H (63 areas), I (103 areas), and J (161 areas).
These patterns bear the following characteristics:
1) There are very apparent local characteristics. Especially the majority of areas in Hokkaido, which are largely different from other regions and Okinawa, who has a unique composition. Others are divided by the category of Koshin and Chubu regions, which are next to the Kanto region, and many other regions that show the average composition of household style in Japan.
2) They are related to the difference in scale (number of cities within the area and population) and transformation of housing market areas.
3) The scale of these housing market areas reflects the classification of the expansion of area composition (how the municipalities are connected).
