LEADS AND LAGS OF ECONOMIC FLUCTUATIONS IN TERMS OF NEW JOB OFFERS: THE CASE STUDY OF THE NORTHWESTERN PART OF THE TOKYO METROPOLITAN AREA

Abstract

In the regional system, there are various impacts to which the areas respond. Economic fluctuations are known as one of them. There have been many geographical researches of this theme, by King, Jeffrey, Haggett et al. The geographical researches of this theme consist of three steps: the identification of the spatial patterns of the economic fluctuations, the research of the processes of the inter-areal transmission from the viewpoint of the areal lead-lag relationships of the fluctuations, and the explanation of the relationships. As the statistical methods to detect the areal lead-lag relationships, regression analysis, lag-correlation analysis and cross spectral analysis have been used in the previous studies. Cross spectral analysis is the most useful and appropriate one. And, to interpret the leadlag relationships, the industrial-mix theory or the export base theory is the most appropriate at present. At the level of national urban systems, the spatial lead-lag relationships have been discussed in the previous studies, but there have been few researches of them within the metropolitan areas. The aim of this paper is to find the spatial lead-lag relationships within the metropolitan area. The study area is the northwestern part of the Tokyo Metropolitan Area, which consists of three prefectures, Tokyo, Saitama and Gumma (Fig. 1). This part has relatively little variations in topography and distribution of cities. As the index to measure the economic activity level, unemployment rate, which have often been used in the same kind of studies, is inappropriate for statistical reasons in Japan. The data used in this paper is new job offers, which is known as one of the time series data which coincides with diffusion index, the general indicator of economic fluctuations. The unit area of this data is a labor market area. The study area consists of 38 labor market areas. The time period in this study is from January, 1967 to August, 1977 (128 months). Prior to the analysis of the lead-lag relationships of the economic fluctuations, the re gional fluctuation patterns were analyzed by spectral analysis. The spectrum of the economic fluctuations in each area was obtained directly by fast fourier transformation (FFT) from the original data (Fig. 2). As the result, we could see the similar spectral patterns in each area. High spectra were indicated at 36 month period and more, 12 month period, and 6 month period. It means that the fluctuations of new job offers in this area consist of the cyclical fluctuations of 36 month period and more, and the seasonal fluctuations of 6 month period and 12 month period. The spatial lead-lag relationships of economic fluctuations were obtained by cross spec tral analysis for all pairs of unit areas. The calculated cross spectrum from the correspond ing pair of spectra is practically used as two measures, coherence and phase angle. Now let us consider the spatial lead-lag relationships at 42. 5 month period and 127. 5 month period. The former is the nearest period to the Kitchin cycles and the latter is the nearest period to the Juglar cycles, because the spectra at the restricted periods are obtained by FFT. To simplify the discussion, the leading months evaluated for each pair of areas at the two periods were averaged as follows: _??_ where Li; the averaged time-lead in month in area i, lij; the time-lead in month of area i to area (lij is negative when the fluctuation in area i lags behind the one in area j), N; number of areas. Fig. 4 shows the distribution of the averaged time-lead in month L.. From above results, the fluctuations at 42. 5 month period tend to lead on the fringe ofthe Tokyo Metropolitan Area and to lag at the core of it. On the other hand, the fluc tuations at 127. 5 month period tend to lead at the core and to lag on the fringe. g nge.