2003 Volume 76 Issue 4 Pages 187-210
Much attention has been devoted to the changing structure of metropolitan areas due to population dispersal. In the latter half of the 1990s, however, population recovery was reported in the central parts of large Japanese cities such as Tokyo and Osaka, suggesting a reversal of population redistribution in metropolitan areas from dispersal to reconcentration. This trend may indicate a turning point in the structural development of metropolitan areas.
With this in mind, the purpose of this paper is to examine the spatio-temporal structure of population growth in Japan's metropolitan areas in detail. The expansion method, which assumes spatiotemporal continuity of population growth (1965-2000), is employed as the central model, and the spatial structure of population growth in the near future (2000-2005) is forecast based on the application results. The objects studied here are the nine major metropolitan areas in Japan. To delineate metropolitan areas, the Standard Metropolitan Employment Area (SMEA) advocated by H. Yamada and K. Tokuoka was used. The major findings obtained are summarized as follows.
Although a decrease and a leveling off of the population growth rate and population dispersion were common to all of the SMEAs until 1990, a difference was found both in the process of dispersion and in the trend after 1990. First, in Tokyo, Osaka, and Fukuoka, although centrifugal movement of the center of population growth had been under way until 1990, a shift in population redistribution from decentralization to reconcentration was confirmed after the mid-1990s. Second, in other SMEAs, the center of population growth remained in zones 10-15 km away from the metropolitan center until 1990. A noteworthy distinction was detected for the period starting from the latter half of the 1990 s. In Nagoya and Kyoto, although population recovery around the city center was also seen, it was only a change from absolute to relative dispersion. In Sapporo, Hiroshima, and Kitakyushu, however, population recovery around the city center was not seen and an overall decline was forecast for the near future.
With respect to the variation among the SMEAs, a few possible factors are raised as working hypotheses. First, there may be a difference in the type of population pressure on land use, mainly due to different population sizes and topographical conditions. Second, the differences in urban functions might also be important because these factors are assumed to make a difference in the accumulation of professionals, managers, and information industry workers, who are forecast to be the impetus behind population recovery around the city center. Additional reasons for this population recovery may be the increase in large-sized condominiums due to the recession, diversification of households (including a rapid increase in the number of one-person households), and the restructuring of the metropolitan economy.
The above-mentioned results are different from those predicted by Klaassen's hypothesis of metropolitan development. According to that hypothesis, the stage of reurbanization is assumed after depopulation of the metropolitan area, but a population increase has continued in the major metropolitan areas in Japan. Moreover, land prices fell greatly due to the recession of the 1990s, and a large amount of surplus land was released in the major cities in Japan. It seems likely that the population recovery around a city center might be confirmed before the population of the metropolitan area decreases, implying a different result from the one expected in Klaassen's hypothesis.
The analytical framework of the expansion method assuming spatiotemporal continuity yielded the interesting findings described above. However, this method did not incorporate sectoral differences in metropolitan population increases, and goodness-of-fit was not sufficiently high due to the detached (or noncontinuous) increase/decrease in population.
