Portal de Periódicos da CAPES

Origin Distribution Visualization of Floating Population and Determinants Analysis: A Case study of Yiwu City

2011; Elsevier BV; Volume: 7; Linguagem: Inglês

10.1016/j.proenv.2011.07.021

1878-0296

Hongsheng Li, Yingjie Wang, Jiafu Han, Zhuoyuan Yu,

Regional Economics and Spatial Analysis

Based on registered individual floating population data from 2005 to 2008 of Yiwu, the phenomena that population floating to Yiwu City from 34 province and 91 counties in Jiangxi provinces is analyzed. The study aims at analyzing the "pull" forces of Yiwu City and developing migration models for understanding determinants factors of population migration/floating into Yiwu City from other areas in China. The spatial layout of Yiwu's pull forces is proved as a V-shaped pattern consisting of the two axes by using explorative spatial data analysis and map visualization method. The migration models with (model 3) or without (model 2) migration stock are presented and estimated using standard linear regression model, spatial error model as well as spatial lag model at the county scale in Jiangxi province. Based on the likelihood statistics, the AIC and the Moran's I statistics of residuals, the model with migration stock provides an improved fit over the model without migration stock. The correlation between migration ratio and man land ratio is significant at the 0.5 level according to estimates of model 3 and spatial version of model 2. All the three estimates of model 2 and the OLS results of model 3 confirm the distance-decay effect while results from the spatial version of model 3 failed to support the distance rule in population floating. Contrary to the previous studies at the provincial level, the correlation between per capital net income of rural labor forces and migration ratio is not significant according to the three versions of the two models due to the small disparities of income within the counties in Jiangxi. Examination of specification tests in spatial version of model 3 indicates that there is less significant spatial error dependence in the spatial lag models than spatial lag dependence in the error models, further suggesting a preference for the lag model. Model 2 does not suggest any preference for choosing spatial error model and spatial lag model.