Journal of Geographical Sciences >
The spatial patterns and determinants of internal migration of older adults in China from 1995 to 2015
Liu Ye (1986), PhD and Professor, specialized in population geography, urban geography and health geography. Email: liuye25@mail.sysu.edu.cn 
Received date: 20211224
Accepted date: 20220825
Online published: 20221225
Supported by
National Natural Science Foundation of China(42001153)
National Natural Science Foundation of China(42001161)
Although China was one of the countries with the fastestgrowing aging population in the world, limited scholarly attention has been paid to migration among older adults in China. The full picture of their migration in the entire country over time remains unknown. This study examines the spatial patterns of older interprovincial migration flows and their drivers in China over the period 1995 to 2015, using four waves of census data and intercensal population sample survey data. Results from eigenvector spatial filtering negative binomial regressions indicate that older adults tend to migrate away from low costofliving rural areas to high costofliving urban and rural areas, moving away from areas with extreme temperature differences. The location of their grandchildren is among the most important attractions. Our findings suggest that familyoriented migration is more common than amenityled migration among retired Chinese older adults, and the costofliving is an indicator of economic opportunities for adult children and the quality of senior care services.
LIU Ye , HUANG Cuiying , WU Rongwei , PAN Zehan , GU Hengyu . The spatial patterns and determinants of internal migration of older adults in China from 1995 to 2015[J]. Journal of Geographical Sciences, 2022 , 32(12) : 2541 2559 . DOI: 10.1007/s114420222060z
Figure 1 Administrative divisions of provinciallevel regions in China 
Table 1 Independent variables for models 
Variable  Description 

POP_{i,j}  Population sizes of older adults in the origin and destination provinces, persons ^{a} 
DIST_{ij}  Minimum travel time by rail between provincial capitals of the origin and destination provinces, hours ^{b} 
RCONS_{i,j}  Per capita consumption expenditure of rural residents in the origin and destination provinces, yuan ^{a} 
UCONS_{i,j}  Per capita consumption expenditure of urban residents in the origin and destination provinces, yuan ^{a} 
TEMP_{i,j}  Average temperatures difference between January and July in the provincial capitals of the origin and destination provinces, ℃^{a} 
BED_{i,j}  The number of hospital beds per ten thousand population in the origin and destination provinces, beds ^{a} 
MIG_{ij}  The number of migrants under the age of 16 between the origin and destination provinces, persons ^{c} 
Source: ^{a} China Statistical Yearbook, 19962016.^{ b }National Railway Passenger Train Timetable. ^{c} 2000 and 2010 population censuses and 2005 and 2015 1% population sample surveys. 
Table 2 Top 5 provinces of inmigration and outmigration of older adults in China, 19952015 
19952000  20002005  

In  Out  In  Out  
Guangdong  60200  Heilongjiang  49346  Guangdong  108899  Heilongjiang  90739 
Beijing  44889  Sichuan  48360  Tianjin  75012  Sichuan  70986 
Shandong  43278  Anhui  30192  Shanghai  70727  Anhui  57507 
Jiangsu  42652  Jiangxi  28313  Jiangsu  66339  Shandong  53416 
Shanghai  39753  Hebei  26564  Beijing  64453  Henan  47508 
20052010  20102015  
In  Out  In  Out  
Guangdong  667354  Sichuan  441041  Guangdong  201564  Sichuan  125536 
Shanghai  558934  Heilongjiang  424201  Beijing  157672  Henan  93296 
Zhejiang  499986  Anhui  358934  Jiangsu  111183  Heilongjiang  89813 
Beijing  486301  Hunan  333673  Shanghai  102896  Hebei  80023 
Jiangsu  322094  Henan  303145  Zhejiang  62677  Anhui  78891 
Figure 2 Migration efficiency of older adults in China by province in 19952000 (a), 20002005 (b), 20052010 (c), and 20102015 (d) 
Figure 3 The 25 largest interprovincial migration flows of older adults in China in 19952000 (a), 20002005 (b), 20052010 (c), and 20102015 (d) 
Table 3 NBRM and ESF NBRM on driving forces for interprovincial migration of older adults in China, 19952015 
Dependent variable  NBRM  ESF NBRM  

Estimate  z value  Estimate  z value  
POP_{i}  0.418***  6.140  0.494***  6.670 
POP_{j}  0.014  0.220  0.072  1.060 
DIST_{ij}  0.665***  12.410  0.686***  11.190 
RCONS_{i}  1.009***  3.490  0.978***  3.170 
RCONS_{j}  0.031  0.120  0.486*  1.740 
UCONS_{i}  0.407  1.000  0.325  0.790 
UCONS_{j}  2.084***  5.750  1.935***  5.280 
TEMP_{i}  0.509**  2.590  0.790***  3.670 
TEMP_{j}  0.025  0.150  0.105  0.570 
BED_{i}  0.573**  2.590  0.513**  2.230 
BED_{j}  0.010  0.050  0.293  1.500 
MIG_{ij}  0.166***  13.930  0.155***  12.790 
year2  0.108  0.660  0.062  0.380 
year3  0.352  1.280  0.533*  1.840 
year4  2.236***  5.790  2.636***  6.170 
Constant  24.509***  7.260  29.450***  8.080 
Eigenvectors  —  10  
MC of residual migration flows  0.215***  23.373  0.197***  21.530 
N  3480  3480  
Log likelihood  8391.829  8346.159  
AIC  16815.660  16744.320  
BIC  16908.480  16895.160 
Notes: ***, **, and * denote statistical signiﬁcance at the 1%, 5%, and 10% levels, respectively. The variables of year2, year3, year4 denote the periods of 20002005, 20052010, and 20102015, respectively. NBRM, negative binomial regression model; ESF, eigenvector spatial filtering; MC, Moran’s I coefficient; AIC, Akaike information criterion; BIC, Bayesian information criterion. 
Table 4 ESF NBRM on the evolution of driving forces for interprovincial migration of older adults in China, 19952015 
 Estimate  *year2 Estimate  *year3 Estimate  *year4 Estimate  

RCONS_{i}  0.796  1.029  0.112  0.001 
RCONS_{j}  0.313  0.103  0.803  1.218 
UCONS_{i}  1.420*  3.809***  1.744*  1.513 
UCONS_{j}  2.204***  0.524  0.221  3.043*** 
TEMP_{i}  0.178  1.357***  0.959*  1.699*** 
TEMP_{j}  1.034**  1.257***  1.840***  1.563*** 
BED_{i}  1.654***  1.898***  0.548  1.305* 
BED_{j}  0.031  0.456  0.904*  1.183 
MIG_{ij}  0.139***  0.023  0.062**  0.020 
Constant  22.361***  
Other control variables  Have been controlled  
Eigenvectors  10  
N  3480  
Log likelihood  8261.423  
AIC  16628.850  
BIC  16936.320 
Notes: ***, **, and * denote statistical signiﬁcance at the 1%, 5%, and 10% levels, respectively. The variables of year2, year3, year4 denote the periods of 20002005, 20052010, and 20102015, respectively. NBRM, negative binomial regression model; ESF, eigenvector spatial filtering; MC, Moran’s I coefficient; AIC, Akaike information criterion; BIC, Bayesian information criterion. 
[1] 

[2] 

[3] 

[4] 

[5] 

[6] 

[7] 

[8] 

[9] 

[10] 

[11] 

[12] 

[13] 

[14] 

[15] 

[16] 

[17] 

[18] 

[19] 

[20] 

[21] 

[22] 

[23] 

[24] 

[25] 

[26] 

[27] 

[28] 

[29] 

[30] 

[31] 

[32] 

[33] 

[34] 

[35] 

[36] 

[37] 

[38] 

[39] 

[40] 

[41] 

[42] 

[43] 

[44] 

[45] 

[46] 

[47] 

[48] 

[49] 

[50] 

[51] 

[52] 

[53] 

[54] 

[55] 

[56] 
United Nations Department of Economic and Social Affairs (UNDESA), 2020. World Population Ageing 2019, United Nations.

[57] 

[58] 

[59] 

[60] 

[61] 

[62] 

[63] 

/
〈  〉 