Comparing inter-migration within the European Union and China: An initial exploration

Abstract

Labour mobility has been extensively studied in China and the European Union (EU). However, there has been very little attempt to compare interstate migration within the EU and inter-provincial migration in China. This paper provides an account of an initial exploratory quantitative comparison of EU and Chinese intermigration. The paper first makes the case for comparing the EU and China in the context of the growing literature on international comparisons of migration. Problems of data and definition are then explored, and a review of the literature undertaken to identify which determinants could be used as the basis of comparative study. Mobility patterns are measured by comparing selected indicators including inequality, dispersion and effectiveness. The migration process is then measured in terms of the elasticities of intermigration responses to various contributing factors, among which spatial adjacency, distance, economic prospects, migrant networks, and migration policy are filtered out and employed for the study. Gross migration flow models are calibrated to produce initial comparative results, using fixed-effect negative binomial regression methods and a variety of sources of data. The conclusion discusses how this initial exploration has helped identify some potential research directions for future work.

Spatial Interaction Modeling of OD Flow Data: Comparing Geographically Weighted Negative Binomial Regression (GWNBR) and OLS (GWOLSR)

“…Spatial interaction modeling has been employed extensively to analyze a variety of geographical mobility factors (e.g., commuting, tourism, and migration). However, differences in methods used, e.g., OLS, Poisson, and NB regression, to calibrate the spatial interaction models have also been highlighted within the literature [1,7,9]. When considering spatial non-stationarity within spatial interaction models [27], calibration becomes more complicated as mobility often deals with count-type dependent variables (e.g., number of people or goods).…”

“…where M ij is the number of vehicles travelling from county i to county j (indicating the interaction intensity between two areas), GDP i and GDP j represent the push and pull forces at the origin county i and destination county j, respectively, and d ij denotes the network distance between the capitals of counties i and j. A negative exponential function was selected for the distance decay function f d ij at a regional scale [1,42], e.g., exp −βd ij , where β means the spatial distance friction coefficient, where a higher β value indicates that the flow is more sensitive to the network distance.…”

“…The Poisson regression method has been employed in numerous studies to calibrate the model in Equation 1 [46][47][48]. In cases where the flow data is shown to demonstrate over-dispersion (i.e., its variance is much larger than its mean), negative binomial regression should be chosen [1] and Equation (2) updated accordingly, as shown by Equation 3:…”

“…Spatial interaction models, extensively used to investigate and analyze spatial movements, have become a well-established method for understanding factors affecting geographical mobility, such as migration [1,2], transport [3], international trade [4], commuting [5], and tourism [6]. A primary concern in spatial interaction modeling is the statistical and spatial distributions of OD (origin-destination) flows between origin and destination locations, which do not tend to follow normal distribution and spatial independence.…”

A Multiscale Flow-Focused Geographically Weighted Regression Modelling Approach and Its Application for Transport Flows on Expressways

“…Economy, population and distance have been proven to be the most important factors in driving geographical mobility (such as migration [28], tourism [29] and commuting [30]). In this study of regional mobility, the total population in the origin county was treated as a pushing force.…”