Panel Analysis of Car Ownership in Prefectural Level

Numbers of vehicles and average running distance by vehicle type are important indices to estimate future road traffic demand. These indices are different in each prefecture due to their socio-economic characteristics.This paper aims to grasp the spatial and temporal changes in car ownership and the effects of various socio- economic variables on it. Based on these changes, two panel models, ‘Dummy-Variable Model’, and ‘Variable-Coefficient Model’, for car ownership are developed and compared with each other in terms of goodness-of-fit and their predictive power using panel data set of each prefecture from 1965 to 1989.We may, therefore, conclude that panel models can explain well the spatial and temporal changes and is useful in forecasting car ownership level.