REGIONAL GROWTH AND BUSINESS CYCLES IN JAPAN

This study examines the consistency and gaps in national and regional business cycles in Japan from a Bayesian point of view. The Tokyo monopolar system started in the mid†1970s, and recent descriptive statistics, such as migration and per capita income, show that the system continues, despite severe crises such as the burst of the 1990s economic bubble and the Lehman Brothers bankruptcy. We explore the relationship between national and regional business cycles in the system using a spatio†temporal Markov†switching model with the Markov chain Monte Carlo method. Our empirical results show that overall, the regional business cycle in the Kanto region, including Tokyo, is identical to the national business cycle. Moreover, we find that switches in the degree of spatial dependency occur around the turning points of business cycles, and that the degree of spatial dependency tends to be higher during a recession.