Excess Commuting in the Tokyo Metropolitan Area: Measurement and Policy Simulations

In this paper data from the 1985 Japanese Population Census are used to measure 'excess' commuting among 211 origin and destination points in the Tokyo Metropolitan Area. Using an approach developed by Hamilton, it is found that about 90 per cent of commuting time in Tokyo is excess. Using White's method only about 15 per cent of commuting time is excess. Sensitivity analyses reveal that the results are not very sensitive to the major assumptions of the analyses. The findings are compared and contrasted with Small and Song's results for the Los Angeles metropolitan area. Several simulations are used to examine the effect of decentralisation of employment on commuting behaviour and it is found that commuting time could be significantly reduced by decentralising jobs or centralising workers.