no19-24

[23] Dynamic Residential Location Choice Model for the Prediction of Population Movement after a Catastrophic Disaster

○Rena Koseki¹, Eiji Hato¹ (1. The University of Tokyo)

Keywords:Disaster, Population movement, Recursive Logit model, Location choice

After the Great East Japan Earthquake, population flows occurred over the wide area. In order to develop a reconstruction plan for a city with a declining population, it is necessary to develop a reconstruction management method based on an understanding of the mechanism of population movement with various support after a disaster. In this study, we propose a dynamic residential location choice model under uncertain conditions using Discounted Recursive Logit model. The proposed model is empirically analyzed using actual data on residential histories for 10 years after the earthquake, and the applicability of the model to assessing the impact of policy factors and environmental factors are confirmed. In the future study, it is necessary to incorporate variables that explain the effects of reconstruction projects and other policies for estimation of discount rate, and to develop a model that takes into account time-series heterogeneity in the sampling of residential options.