Internal Migration in the Countries of Asia: A Cross-national Comparison [1st ed.] 9783030440091, 9783030440107

Citation preview

Martin Bell Aude Bernard Elin Charles-Edwards Yu Zhu   Editors

Internal Migration in the Countries of Asia A Cross-national Comparison

Internal Migration in the Countries of Asia

Martin Bell  •  Aude Bernard Elin Charles-­Edwards  •  Yu Zhu Editors

Internal Migration in the Countries of Asia A Cross-national Comparison

Editors Martin Bell Asian Demographic Research Institute Shanghai University Shanghai, China

Aude Bernard Asian Demographic Research Institute Shanghai University Shanghai, China

Queensland Centre for Population Research The University of Queensland Brisbane, QLD, Australia

Queensland Centre for Population Research The University of Queensland Brisbane, QLD, Australia

Elin Charles-Edwards Asian Demographic Research Institute Shanghai University Shanghai, China

Yu Zhu Asian Demographic Research Institute Shanghai University Shanghai, China

Queensland Centre for Population Research The University of Queensland Brisbane, QLD, Australia

School of Geography Fujian Normal University Fuzhou, Fujian Province, China

ISBN 978-3-030-44009-1    ISBN 978-3-030-44010-7 (eBook) https://doi.org/10.1007/978-3-030-44010-7 © Springer Nature Switzerland AG 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Foreword

This edited book examines the spatial patterns of internal migration across 15 countries in Asia. Why is this important? Rather little has been written about internal migration in Asia outside of the countries of East Asia (China, Japan and Korea). It is important to broaden the evidence base for describing internal migration across the world’s most populous continent. Analyses of internal migration to date, although they may be excellent, are very difficult to compare because of varying definitions, measures and spatial and temporal extents. What is needed is a common framework of theory about the processes of internal migration and its measurement using indicators that can be compared. This book supplies those theories and comparable measures of internal migration. As an edited book with chapters by experts on the internal migration history of their countries, the key role of the national context and various shocks to national systems can also be distinguished. The book is the product of the editors’ grand ambition: to ensure that internal migration can be measured and reported in a robust way. Achievement of this aim will enable internal migration to join, alongside mortality, fertility and international migration, the collections of demographic statistics that are so influential in understanding global challenges. Internal migration has long been the “Cinderella”1 among the demographic components of change. This book is a major step in taking internal migration to the demographic “ball”. Why has it taken so long for this kind of rigorous comparison of internal migration to come to fruition? Demographers had long been aware that the spatial systems used to capture internal migration produce measures that are dependent on the number, size and shape of the territorial units. This has been termed the modifiable areal unit problem (MAUP) investigated thoroughly by Stan Openshaw in his 1983 monograph. The lead editor, Martin Bell, saw that methods were needed to correct internal migration for the effect of the MAUP. Collaborating with an international team in a succession of projects since the late 1990s, Martin and his colleagues

1  The Cinderella narrative is a global story with Greek and Chinese origins and versions in many different cultures (Wikipedia 2019: https://en.wikipedia.org/wiki/Cinderella).

v

vi

Foreword

developed the methods to confront the MAUP and develop indicators that treated most of its symptoms. Initial collaborations with Philip Rees, John Stillwell and Paul Boyle (all British researchers) in comparisons of internal migration between Australia and the UK led to the foundation paper (Bell et al. 2002) that developed a framework for measuring internal migration. The opportunity to operationalise this framework came via an Australian Research Council grant to implement the project Internal Migration Around the GlobE (https://imageproject.com.au). This project involved each of the book editors, together with Polish colleagues Marek Kupiszewski and Dorota Kupiszewska, John Stillwell and Kostas Daras (UK-based) and Philipp Ueffing in Brisbane. Kostas Daras, working with John Stillwell, Philip Ueffing and Martin Bell, wrote the sophisticated software (IMAGE Studio) that enabled the team to compute indices systematically at different spatial scales and for any one scale to compute indices for numerous different areal configurations. Kostas Daras studied at Newcastle for his doctorate on zone design with Serafeim Alvanides, whose dissertation had been supervised by Stan Openshaw, thus linking the problem to its solution. Edited books come in all sizes and shapes. Usually, they bring together insights from a team of authors on a topic, but the editors struggle to find the common or contradictory messages in the work because the analyses are not standardised. This book adopts the model of persuading authors to follow an agreed agenda so that results are comparable but to add their own understanding of specific country factors. So, you will find comparable tables, charts and migration flow plots in each of the book’s country studies. In the last chapter of the book, the editors learn from the country authors what special factors have, from time to time, driven their country off the “belt and road” of the editors’ specification of internal migration theory set out in Chap. 2. Of course, the journey to global batteries of internal migration indicators in published international statistical collections continues. The challenge of solving the modifiable temporal unit problem (MTUP) remains. This problem is how to convert data sources which use different time frames (e.g. 1 year, 5 years, lifetime) in census and survey questions to a common metric. The IMAGE-Asia team may find the solution to this MTUP challenge, perhaps working with Daniel Courgeau, as they did in developing a measure of intensity that overcomes the MAUP problem. In the meantime, enjoy reading and learning about internal migration in Asian countries while Cinderella (internal migration) is at the ball. Professor Emeritus, School of Geography University of Leeds Leeds, UK

Philip Rees

Preface

This book explores the way in which internal migration, the propensity to change residence within national borders, varies among the countries of Asia. Such movements are of rising importance in the modern world. They are the primary mechanism shaping patterns of human settlement, an essential process adjusting labour supply to demand, and key to enabling individuals to achieve their goals and aspirations. As such, migration has relevance across a wide range of public policy. Despite this, much less attention has been given to understanding mobility within countries than has been accorded to international migration, especially in a comparative context. While there is a long-standing tradition of scholarship into population mobility in parts of Asia dating back to the 1970s, what has been lacking is a systematic approach that enables robust comparisons to be made between countries using reliable statistical measures. This book aims to achieve that goal by harnessing the repository of migration data, the analytical techniques and the statistical indicators developed as part of the IMAGE (Internal Migration Around the GlobE – https:// imageproject.com.au) project. Key findings from the IMAGE project have been published in a series of methodological, thematic and regional papers. The unique contribution of this book lies in coupling the IMAGE metrics with local contextual knowledge in a systematic, structured approach to better understand the forces that shape mobility in individual country settings across the length and breadth of Asia. The book had its genesis in a proposal outlined at the Asian Demographic Research Institute (ADRI) in mid-2017, which envisaged a collaborative project extending the IMAGE project across the Asian region. The proposal had two key objectives: first, to draw on the knowledge of individual country experts to better interpret the IMAGE measures of migration and, second, to enhance training across the region in the quantitative analysis of internal migration data. Potential participants from some 20 countries were invited to collaborate and provided with a framework paper and relevant IMAGE migration metrics for their country. As a key step in the project, they were then invited to a 2-day workshop in Shanghai in mid-2018, funded by ADRI, to coincide with the 4th Conference of the Asian Population Association. Draft papers presented at the Shanghai workshop were

vii

viii

Preface

subsequently revised and refined, in close consultation with the editors, to deliver the 15 country-specific chapters now included in this book. A project of this magnitude does not come to fruition without significant contributions from a wide range of sources. The editors are particularly grateful to the Asian Demographic Research Institute, which provided sustained encouragement and considerable financial support for contributors to attend the 2018 workshop and for travel to related meetings by the editors themselves. It is equally important to acknowledge funding support from the Australian Research Council under Discovery Project DP 110101363, through which the IMAGE project was undertaken over the period 2011–2015. Data used in the project were drawn from a range of sources, particularly the publications and datasets of national statistical agencies, but the editors are also pleased to acknowledge the contribution of the University of Minnesota IPUMS data repository, an invaluable resource for cross-national comparisons of this type. The IMAGE metrics themselves, which form the foundation for the work reported here, reflect the combined intellectual output of a number of redoubtable scholars. Of more immediate note have been the contributions of Dr. Chen Chen of ADRI who assisted with data preparation and analysis, Rosabella Borsellino who took care of graphic design and Michelle Burgess who copy-edited the manuscript. Finally, the editors would like to thank the authors who contributed the 15 country chapters, many of whom had limited prior experience in the analysis of population mobility. Brisbane, QLD, Australia

Martin Bell

Contents

Part I The Framework 1 IMAGE-Asia: An Introduction ��������������������������������������������������������������    3 Elin Charles-Edwards, Martin Bell, Aude Bernard, and Yu Zhu 2 Understanding Internal Migration: A Conceptual Framework����������   15 Aude Bernard, Martin Bell, Elin Charles-Edwards, and Yu Zhu 3 Comparative Measures of Internal Migration��������������������������������������   31 Martin Bell, Aude Bernard, Elin Charles-Edwards, and Wenqian Ke Part II The Evidence 4 Internal Migration in China ������������������������������������������������������������������   51 Jianfa Shen 5 Internal Migration in Mongolia��������������������������������������������������������������   77 Solongo Algaa 6 Internal Migration in South Korea��������������������������������������������������������   93 Yeonjin Lee and Doo-Sub Kim 7 Internal Migration in Japan��������������������������������������������������������������������  113 Yoshitaka Ishikawa 8 Internal Migration in Cambodia������������������������������������������������������������  137 Jean-Christophe Diepart and Chanrith Ngin 9 Internal Migration in Myanmar������������������������������������������������������������  163 Maxime Boutry 10 Internal Migration in Thailand��������������������������������������������������������������  185 Aree Jampaklay ix

x

Contents

11 Internal Migration in India��������������������������������������������������������������������  207 Ram B. Bhagat and Kunal Keshri 12 Internal Migration in Bhutan ����������������������������������������������������������������  229 Mayur Gosai and Leanne Sulewski 13 Internal Migration in Nepal��������������������������������������������������������������������  249 Samir KC 14 Internal Migration in Sri Lanka������������������������������������������������������������  269 E. L. Sunethra J. Perera 15 Internal Migration in Iran����������������������������������������������������������������������  295 Rasoul Sadeghi, Mohammad Jalal Abbasi-Shavazi, and Saeedeh Shahbazin 16 Internal Migration in Israel��������������������������������������������������������������������  319 Uzi Rebhun 17 Internal Migration in Armenia ��������������������������������������������������������������  349 Karine Kuyumjyan 18 Internal Migration in Kazakhstan ��������������������������������������������������������  365 Aidan Islyami Part III Synthesis 19 Conclusions����������������������������������������������������������������������������������������������  385 Martin Bell, Elin Charles-Edwards, and Aude Bernard

Contributors

Mohammad  Jalal  Abbasi-Shavazi  Department of Demography, University of Tehran and National Institute of Population Research, Tehran, Iran Solongo  Algaa  School of Arts and Science, National University of Mongolia, Ulaanbaatar, Mongolia Martin  Bell  Asian Demographic Research Institute, Shanghai University, Shanghai, China Queensland Centre for Population Research, The University of Queensland, Brisbane, QLD, Australia Aude  Bernard  Asian Demographic Research Institute, Shanghai University, Shanghai, China Queensland Centre for Population Research, The University of Queensland, Brisbane, QLD, Australia Ram  B.  Bhagat  Department of Migration and Urban Studies, International Institute for Population Sciences, Mumbai, India Maxime  Boutry  PALOC  – Patrimoines Locaux et  Gouvernance, Institut de Recherche pour le Développement (IRD), Marseille, France Elin  Charles-Edwards  Asian Demographic Research Institute, Shanghai University, Shanghai, China Queensland Centre for Population Research, The University of Queensland, Brisbane, QLD, Australia Jean-Christophe  Diepart  Gembloux Agro-Bio Tech, University of Liège, Liège, Belgium Mayur Gosai  Department of Geography, University of Cambridge, Cambridge, UK Yoshitaka  Ishikawa  Faculty of Economics, Teikyo University, Hachioji City, Tokyo, Japan

xi

xii

Contributors

Aidan  Islyami  International School of Economics, Kazakh-British Technical University, Almaty, Kazakhstan Aree Jampaklay  Institute for Population and Social Research, Mahidol University, Nakhon Pathom, Thailand Wenqian  Ke  School of Geography, Fujian Normal University, Fuzhou, Fujian Province, China Asian Demographic Research Institute, Shanghai University, Shanghai, China Kunal Keshri  G. B. Pant Social Science Institute, Central University of Allahabad, Prayagraj, India Doo-Sub Kim  Department of Sociology, Hanyang University, Seoul, South Korea Karine  Kuyumjyan  Population Census and Demography Division, Statistical Committee, Yerevan, Republic of Armenia Yeonjin Lee  Department of Social Work and Social Administration, and School of Public Health, University of Hong Kong, Hong Kong, China Chanrith Ngin  Development Studies, School of Social Sciences, Faculty of Arts, The University of Auckland, Auckland, New Zealand E.  L.  Sunethra  J.  Perera  Department of Demography, University of Colombo, Colombo, Sri Lanka Uzi Rebhun  Division of Jewish Demography & Statistics, The Harman Institute of Contemporary Jewry, The Hebrew University of Jerusalem, Jerusalem, Israel Rasoul Sadeghi  Department of Demography, University of Tehran and National Institute of Population Research, Tehran, Iran Samir  KC  Asian Demographic Research Institute, Shanghai University, Shanghai, China World Population Program, International Institute for Applied Systems Analysis, Laxenburg, Austria Saeedeh  Shahbazin  Research Group on Internal Migration and Urbanization, National Institute of Population Research, Tehran, Iran Jianfa Shen  Department of Geography and Resource Management, The Chinese University of Hong Kong, Hong Kong, China Leanne  Sulewski  Department of Geography, University of South Carolina, Columbia, SC, USA Yu  Zhu  Asian Demographic Research Institute, Shanghai University, Shanghai, China School of Geography, Fujian Normal University, Fuzhou, Fujian Province, China

List of Figures

Fig. 2.1 Fig. 2.2 Fig. 2.3 Fig. 2.4 Fig. 2.5 Fig. 3.1 Fig. 3.2 Fig. 3.3 Fig. 3.4 Fig. 4.1 Fig. 4.2 Fig. 4.3 Fig. 4.4 Fig. 4.5 Fig. 4.6 Fig. 4.7 Fig. 5.1 Fig. 5.2 Fig. 5.3

The drivers of migration.............................................................. 23 Cross-national variations in migration age patterns, selected countries......................................................................... 24 Proximate determinants of migration age patterns...................... 24 Impact of past migration on future migration.............................. 25 Linking development to population redistribution through net migration.................................................................. 26 Estimating the ACMI for Iran, 2006–2011.................................. 38 Measuring the age and intensity at peak migration..................... 39 Circular plot of interregional migration flows for a hypothetical country............................................................ 41 Net migration rates by log population density, regions of Japan, 2016–2017.................................................................... 44 Regions and provinces of China.................................................. 55 Age-specific migration intensities by sex and type of move, registration migration, 2010......................................... 62 Bilateral migration flows, provinces of residence, China, 1995–2000................................................................................... 65 Bilateral migration flows, provinces of residence, China, 2005–2010................................................................................... 66 Net migration rates, provinces of China, 1995–2000.................. 68 Net migration rates, provinces of China, 2005–2010.................. 69 Net migration rates by log population density, provinces of China, 2005–2010................................................... 70 Regions and aimags of Mongolia, 2010...................................... 79 Crude migration intensities, movement between aimags, Mongolia, 1990–2018.................................................... 81 Age-specific migration intensities by sex, movement between aimags, 2009–2010....................................................... 83 xiii

xiv

Fig. 5.4 Fig. 5.5 Fig. 5.6 Fig. 5.7 Fig. 5.8 Fig. 6.1 Fig. 6.2 Fig. 6.3 Fig. 6.4 Fig. 6.5 Fig. 6.6 Fig. 6.7 Fig. 6.8 Fig. 6.9 Fig. 6.10 Fig. 7.1 Fig. 7.2 Fig. 7.3 Fig. 7.4 Fig. 7.5 Fig. 7.6 Fig. 7.7 Fig. 7.8 Fig. 7.9 Fig. 8.1 Fig. 8.2 Fig. 8.3 Fig. 8.4 Fig. 8.5 Fig. 8.6

List of Figures

System-wide migration indicators by type of move, Mongolia, 1990 to 2018............................................................... 84 Bilateral migration flows, aimags of Mongolia, 2005–2010....... 85 Net migration rates, aimags of Mongolia, 2005–2010................ 86 Net migration rates, aimags of Mongolia, 2018.......................... 87 Net migration rates by log population density, aimags of Mongolia, 2005–2010............................................................. 87 Provinces and districts of South Korea, 2017.............................. 95 Aggregate crude migration intensities, South Korea, 1970–2018................................................................................... 98 Age-specific migration intensities by sex, all moves, South Korea, 2018....................................................................... 99 Age-specific migration intensities, South Korea, 1995 and 2018............................................................................................. 100 Net migration rates, provinces of South Korea, 2018.................. 102 Bilateral migration flows, provinces of South Korea, 1975......... 103 Bilateral migration flows, provinces of South Korea, 2018......... 104 Net migration rates by log population density, provinces of South Korea, 1975................................................... 105 Net migration rates by log population density, provinces of South Korea, 2000................................................... 106 Net migration rates by log population density, provinces of South Korea, 2018................................................... 107 Metropolitan areas and prefectures of Japan............................... 116 Inter-prefectural migration, major metropolitan areas of Japan, 1954–2017........................................................... 118 Intra- and inter-prefectural migration intensities, Japan, 1954–2017................................................................................... 120 Age-specific migration intensities, all moves, Japan, 2010–2015................................................................................... 122 Sex ratio of inter-prefectural migration, Japan, 1960–2015........ 123 Age-specific migration effectiveness index, Japan, 2010–2015.. 125 Net migration rates, prefectures of Japan, 2017.......................... 126 Bilateral migration flows, prefectures of Japan, 2017................. 127 Net migration rates by log population density, prefectures of Japan, 2017........................................................... 128 Provinces, districts and urbanisation, Cambodia, 2011............... 141 Year of last move, residents of Cambodia, 2008......................... 144 Age-specific migration intensities, Cambodia, 2003–2008......... 147 Inter-village moves to Phnom Penh by age and sex, 2003–2008................................................................................... 148 Inter-village moves to Pailin by age and sex, 2003–2008........... 149 Composition of migration streams, Cambodia, 2003–2008........ 150

List of Figures

Fig. 8.7 Fig. 8.8 Fig. 8.9 Fig. 8.10 Fig. 8.11 Fig. 8.12 Fig. 8.13

Fig. 9.1 Fig. 9.2 Fig. 9.3 Fig. 9.4 Fig. 9.5 Fig. 9.6 Fig. 9.7 Fig. 10.1 Fig. 10.2 Fig. 10.3 Fig. 10.4 Fig. 10.5 Fig. 10.6 Fig. 10.7 Fig. 10.8 Fig. 11.1 Fig. 11.2 Fig. 11.3 Fig. 11.4 Fig. 11.5 Fig. 11.6

xv

Net migration rates, districts of Cambodia, 2003–2008.............. 151 Net migration rates by log population density, districts of Cambodia, 2003–2008............................................................. 152 Bilateral migration flows, rural and urban areas of Cambodia, 2003–2008............................................................. 153 Origin of migrants living in Phnom Penh, 2003–2008................ 154 Origin of migrants living in Pailin, 2003–2008........................... 154 Reason for last move by sex, Cambodia, 2008............................ 155 Agro-industrial concessions, Protected Areas, hydropower dams, mining concessions and Special Economic Zones, Cambodia........................................................ 158 Regions, states and districts of Myanmar.................................... 166 Net migration rates, districts of Myanmar, 2009–2014............... 172 Net lifetime migration rates, districts of Myanmar, 2014............ 174 Bilateral migration flows, districts of Myanmar, 2009–2014................................................................................... 175 Bilateral lifetime migration flows, districts of Myanmar, 2014........................................................................ 176 Net migration rates by log population density, districts of Myanmar, 2009–2014.............................................................. 177 Net lifetime migration rates by log population density, districts of Myanmar, 2014.......................................................... 178 Regions and provinces of Thailand, 2010.................................... 189 Age-specific migration intensities by sex, persons who changed municipality or Tambon, 2005–2010..................... 193 Age distribution of moves to institutional housing, males who changed municipality or Tambon, 2010 Census........ 195 Bilateral migration flows, provinces of Thailand, 2005–2010................................................................................... 196 Bilateral lifetime migration flows, provinces of Thailand, 2010......................................................................... 198 Net migration rates, provinces of Thailand, 2005–2010............. 199 Net lifetime migration rates, provinces of Thailand, 2010.......... 200 Net migration rates by log population density, provinces of Thailand, 2005–2010............................................................... 201 States and Union Territories of India, 2013................................. 210 Five-year crude migration intensities between villages and towns, India, 1971–2011....................................................... 213 Age-specific migration intensities by sex, India, 2000................ 214 Age-specific migration intensities by sex and reason for move, India, 2000................................................................... 215 Net migration rates, states of India, 1996–2001.......................... 218 Net migration rates, states of India, 2006–2011.......................... 219

xvi

List of Figures

Fig. 11.7 Fig. 11.8 Fig. 11.9

Bilateral migration flows, states of India, 1996–2001................. 220 Bilateral migration flows, states of India, 2006–2011................. 221 Net migration rates by log population density, states of India, 2006–2011........................................................... 224

Fig. 12.1 Fig. 12.2

Dzongkhags of Bhutan, 2017...................................................... 232 Percentage of population who reported being ‘very unhappy’, dzongkhags of Bhutan, 2005............................. 235 Net lifetime migration rates, dzongkhags of Bhutan, 2017......... 238 Bilateral lifetime migration flows, dzongkhags of Bhutan, 2017........................................................................... 239 Share of lifetime migration between rural and urban areas, Bhutan, 2017................................................................................ 240 Net lifetime migration rates by log population density, dzongkhags of Bhutan, 2017....................................................... 241 Main reason for last migration by dzongkhag of destination, 2005..................................................................... 243

Fig. 12.3 Fig. 12.4 Fig. 12.5 Fig. 12.6 Fig. 12.7 Fig. 13.1 Fig. 13.2 Fig. 13.3 Fig. 13.4 Fig. 13.5 Fig. 13.6 Fig. 13.7 Fig. 13.8 Fig. 13.9 Fig. 14.1 Fig. 14.2 Fig. 14.3 Fig. 14.4 Fig. 14.5 Fig. 14.6 Fig. 14.7 Fig. 15.1 Fig. 15.2 Fig. 15.3

States, districts, and ecological regions of Nepal........................ 250 Districts of Nepal......................................................................... 252 Age-specific migration intensities, migration between districts, Nepal, 2006–2011......................................................... 253 Lifetime migration by age, sex and cause, Nepal, 2011.............. 259 Net migration rates, districts of Nepal, 2006–2011..................... 261 Bilateral migration flows, districts of Nepal, 2006–2011............ 261 Net lifetime migration rates, districts of Nepal, 2011................. 262 Bilateral lifetime migration flows, districts of Nepal, 2011........ 263 Net migration rates by log population density, districts of Nepal, 2006–2011................................................................... 264 Provinces and districts of Sri Lanka............................................ 272 Age-specific migration intensities by sex, moves between districts, Sri Lanka, 2008–2012..................................... 277 Reasons for migration between districts, Sri Lanka, 2008–2012................................................................................... 278 Net migration rates, districts of Sri Lanka, 2008–2012............... 280 Bilateral migration flows, districts of Sri Lanka, 2008–2012...... 281 Net migration rates by log population density, districts of Sri Lanka, 2012....................................................................... 286 Net migration rates by log population density, districts of Sri Lanka excluding the Northern province 2012................... 287 Regions and provinces of Iran, 2016........................................... 299 Age-specific migration intensities by sex and type of move, Iran, 2006–2011........................................................................... 302 Age-specific migration intensities by sex and reason for move, Iran, 2006–2011.......................................................... 302

List of Figures

Fig. 15.4 Fig. 15.5 Fig. 15.6 Fig. 15.7 Fig. 15.8 Fig. 15.9 Fig. 16.1 Fig. 16.2 Fig. 16.3 Fig. 16.4 Fig. 16.5 Fig. 16.6 Fig. 16.7 Fig. 17.1 Fig. 17.2 Fig. 17.3 Fig. 17.4 Fig. 17.5 Fig. 17.6 Fig. 17.7 Fig. 17.8 Fig. 18.1 Fig. 18.2 Fig. 18.3 Fig. 18.4 Fig. 18.5 Fig. 18.6 Fig. 18.7 Fig. 18.8

xvii

Net migration rates, provinces of Iran, 2011–2016..................... 306 Bilateral migration flows, provinces of Iran, 2011–2016............ 307 Net migration rates, counties of Iran, 2011–2016....................... 308 Composition of migration streams, Iran, 1976–1986 to 2011–2016............................................................................... 309 Net migration rates by log population density, provinces of Iran, 2011–2016....................................................................... 311 Net migration rates by level of development, provinces of Iran, 2011–2016....................................................................... 312 Sub-districts of Israel................................................................... 323 Aggregate crude migration intensities, Jews, non-Jews and total population, Israel, 1985–2015...................................... 328 Crude migration intensity, migration between localities, Jews, non-Jews and total population, 1985–2015........................ 330 Net migration rates, sub-districts of Israel, 2018......................... 336 Bilateral migration flows, sub-districts of Israel, 2018................ 337 Net rural migration, Jews, non-Jews and total population, 1978–2016 (thousands)................................................................ 339 Net migration rates by population density, sub-districts of Israel, 2018.............................................................................. 340 Regions of Armenia..................................................................... 353 Crude migration intensity, regions of Armenia, 2002–2017........ 355 Age-specific migration intensities, regions of Armenia, 2001 and 2011.............................................................................. 356 Net migration rates, regions of Armenia, 1996–2001.................. 358 Net migration rates, regions of Armenia, 2006–2011.................. 359 Bilateral migration flows, regions of Armenia, 2006–2011......... 360 Composition of migration streams, Armenia, 1972–2017........... 361 Net migration rates by population density, regions of Armenia, 2006–2011............................................................... 362 Regions of Kazakhstan................................................................ 367 Annual migration intensities, regions and districts of Kazakhstan, 2000–2017.......................................................... 370 Age-specific migration intensities by sex, movement between districts, 2004–2009...................................................... 371 System-wide migration indicators, regions of Kazakhstan, 2000–2017................................................................................... 373 Net migration rates, regions of Kazakhstan, 1999–2009............. 373 Net migration rates, regions of Kazakhstan, 2017....................... 374 Bilateral migration flows, regions of Kazakhstan, 1999–2009................................................................................... 375 Bilateral migration flows, regions of Kazakhstan, 2017.............. 376

xviii

List of Figures

Fig. 18.9

Net migration rates by log population density, regions of Kazakhstan, 1999–2009.......................................................... 377 Fig. 18.10 Almaty and Astana share of interregional migration inflows, 1999–2017...................................................................... 377 Fig. 19.1 Fig. 19.2 Fig. 19.3 Fig. 19.4 Fig. 19.5 Fig. 19.6 Fig. 19.7 Fig. 19.8 Fig. 19.9

Aggregate crude migration intensities, selected Asian countries............................................................................ 387 Ratio of lifetime to five-year migration intensity, selected Asian countries............................................................... 388 Age at peak migration, selected Asian countries......................... 388 Index of net migration impact, selected Asian countries............. 390 Regression slopes, net migration rates against log population density, selected Asian countries............................... 391 Trends in crude migration intensities, Japan, South Korea and Mongolia.......................................................... 394 Percentage of population living in urban areas by the net migration-population density slope............................. 395 Net migration rate by log population density, provinces of South Korea, 1975, 2000, 2010, 2017..................................... 397 Internally displaced persons, Selected Asian countries, 1991–2018................................................................................... 399

List of Tables

Table 2.1 Table 2.2

Ravenstein’s laws of migration................................................... 17 Zelinsky’s stages of the migration transition.............................. 19

Table 3.1

Internal migration data collection, selected countries and total Asia............................................................................... 34

Table 4.1 Table 4.2 Table 4.3

Migration data available in China............................................... 53 Administrative areas, Mainland China, 2010.............................. 55 Crude migration intensity by type of move and spatial scale, 2010................................................................ 59 System-wide migration indicators by type of move, 2000 and 2010...................................................................................... 60 Crude migration intensity by sex and type of move, registration migration, 2010........................................................ 61 Total population and migrants aged six and over by education level, 2010............................................................. 62 Registration migration by reason and type of move, 2005–2010.................................................................................. 63 Migration indicators, provinces of China, 1995–2000 and 2005–2010............................................................................ 67

Table 4.4 Table 4.5 Table 4.6 Table 4.7 Table 4.8 Table 5.1 Table 5.2 Table 5.3 Table 6.1 Table 6.2

Crude migration intensities by type of move, Mongolia, 1989, 2000 and 2010................................................................... 81 Number of migrants and crude migration intensities by sex and type of move, Mongolia, 2010.................................. 83 System-wide migration indicators by type of move, Mongolia, 1989, 2000 and 2010................................................. 84 System-wide migration indicators, provinces of South Korea, 1975, 2000 and 2018........................................ 100 Average annual population growth, Seoul, Seoul Metropolitan Area, Gyeonggi Province, and Capital Region, 1960–2000.................................................. 105 xix

xx

List of Tables

Table 6.3

Net migration rate and population density, provinces of South Korea, 1975, 2000 and 2018........................................ 106

Table 8.1 Table 8.2

Migration questions, 1998 Census, Cambodia........................... 139 Internal migration data collected in Cambodia, 1998 onwards.............................................................................. 140 Number of migrants and internal migration rate, inter-village moves, Cambodia, 1998 and 2008......................... 145 Migration intensities by type of move, Cambodia, 1998 and 2008............................................................................. 145 Educational attainment, total population and life-time migrants aged 15 and over, 2008................................................ 149

Table 8.3 Table 8.4 Table 8.5 Table 9.1 Table 9.2 Table 9.3 Table 9.4 Table 10.1 Table 10.2 Table 10.3 Table 10.4 Table 10.5 Table 11.1 Table 11.2 Table 11.3 Table 11.4 Table 12.1 Table 12.2 Table 12.3

Migration questions, 2014 census, Myanmar............................. 167 Crude migration intensity by type of migration and spatial scale, Myanmar, 2014............................................... 169 Reason for migration (between townships) by sex, recent migrants, Myanmar, 2014................................................ 170 System-wide migration indicators by type of move, Myanmar, 2014........................................................................... 171 Migration questions, 1990, 2000 and 2010 censuses, Thailand...................................................................................... 187 Crude migration intensity by type of migration and spatial scale, Thailand, 1990, 2000 and 2010...................... 191 Selected characteristics of migrants and total population, Thailand, 2010............................................................................ 194 Reason for migration by sex, Thailand, 2010............................. 194 System-wide migration indicators by type of move, Thailand, 1990, 2000, 2010........................................................ 195 Crude migration intensities by type of move and ratios of lifetime to five-year migration, India, 2001............................ 212 Reasons for migration between villages and towns, India, 1996–2001 and 2006–2011............................................... 215 Inter-state migration indicators, India, 2001 and 2011............... 216 Share of migrants by type of move, migration between  villages and towns of India, 2001 and 2011............................... 222 Internal migration data collected at the census, Bhutan, 2005 and 2017............................................................................. 231 System-wide lifetime migration indicators, dzongkhags of Bhutan, 2005 and 2017, dzongkhags of Bhutan, 2005 and 2017...................................................................................... 236 Population redistribution through lifetime migration, dzongkhags of Bhutan, 2005....................................................... 237

List of Tables

xxi

Table 12.4

Population redistribution through lifetime migration, dzongkhags of Bhutan, 2017....................................................... 238

Table 13.1 Table 13.2

Migration questions, 1954–2011 Censuses, Nepal..................... 252 Crude migration intensities, districts of Nepal, 2001 and 2011...................................................................................... 256 Reasons for migration by sex, people who changed residence 2006–2011, Nepal....................................................... 258 System-wide migration indicators by type of move, districts of Nepal, 2001 and 2011............................................... 260

Table 13.3 Table 13.4 Table 14.1 Table 14.2 Table 14.3 Table 14.4 Table 14.5 Table 14.6 Table 14.7 Table 15.1 Table 15.2 Table 15.3 Table 15.4 Table 15.5 Table 15.6 Table 15.7 Table 16.1 Table 16.2 Table 16.3 Table 16.4 Table 16.5

Internal migration data collected at the census, Sri Lanka, 1946–2012................................................................. 271 Urban population of Sri Lanka by province, 2012..................... 273 Crude migration intensities by type of move, Sri Lanka, 2012........................................................................... 276 System-wide migration indicators, provinces and districts of Sri Lanka, 2012....................................................................... 278 Migration indicators, provinces of Sri Lanka, 2008–2012......... 279 Lifetime migration indicators, provinces of Sri Lanka, 2012..... 282 Lifetime migration indicators, districts of Sri Lanka, 2012........ 284 Internal migration data collected at the census, Iran, 1956–2016.................................................................................. 297 Regions, provinces and counties of Iran, 1996–2016................. 298 Crude migration intensities by type of move, Iran, 1986–2016.......................................................................... 300 Crude migration intensity by educational attainment and type of move, persons aged 20–65, Iran, 2006–2011........... 303 Crude migration intensity between rural and urban areas by educational attainment, persons aged 20–65, Iran, 2006–2011.................................................................................. 303 System-wide migration indicators, provinces and counties of Iran, 1986–1996 to 2011–2016.......................... 305 System-wide migration indicators, flows between urban and rural areas, Iran, 1986–1996 to 2011–2016............... 309 Sources and types of internal migration data, Israel................... 321 Internal migration data collected in Israeli censuses, 1948–2008.................................................................................. 321 Five-year migration status, Jews and non-Jews by type of move, 1978–83 to 2003–2008................................................ 331 Five-year migration status, Jews and non-Jews by rural-urban moves, 1978–1983 and 2003–2008.................... 332 Crude migration intensities by socio-demographic characteristics, moves between localities, Israel, 2003–2008.................................................................................. 333

xxii

Table 16.6 Table 16.7 Table 16.8 Table 17.1 Table 17.2 Table 17.3 Table 17.4 Table 17.5 Table 18.1 Table 18.2 Table 18.3 Table 19.1

List of Tables

System-wide migration indicators, districts and sub-districts, Israel, 2018..................................................... 335 Net inter-district migration, Jews and non-Jews, 2000–2016.................................................................................. 338 Urban-rural and rural-urban migration effectiveness index, Israel, 1978–1983 and 2003–2008................................... 339 Migration data collected at the Armenian census, 2001 and 2011............................................................................. 351 Regions and communities of the Republic of Armenia, 1 January 2016............................................................................ 352 Crude migration intensities, regions of Armenia, 2001 and 2011...................................................................................... 354 System-wide migration indicators, regions of Armenia, 2001 and 2011............................................................................. 356 Inwards, outwards and net migration rates, regions of Armenia, 2001 and 2011......................................................... 357 Crude migration intensity by geographic level and migration interval, Kazakhstan, 1999 and 2009................... 369 Crude migration intensity by destination and population group, Kazakhstan, 1999–2009................................................... 371 System-wide migration indicators, regions of Kazakhstan, 1989–1999 and 1999–2009................................ 372 Internally displaced persons, selected Asian countries, 2018............................................................................ 398

Part I

The Framework

This book is framed around the models and measures of migration devised and developed in the IMAGE Project (Internal Migration Around the Globe). The three chapters that form this part of the book set out the aims, origins and key features of that project. This, in turn, provides the essential background needed to understand the approach adopted in the empirical analyses of migration in individual countries that are presented in Part II.  Key elements encompassed in these three chapters include the dimensions of migration, the nature of migration data, theoretical concepts, and the indicators or metrics we employ to measure migration.

Chapter 1

IMAGE-Asia: An Introduction Elin Charles-Edwards, Martin Bell, Aude Bernard, and Yu Zhu

1.1  Introduction Migration is singular among demographic processes in its ability to transform the size, distribution and composition of national populations. The impact of migration on national settlement systems is likely to grow as more countries complete the demographic transition and migration becomes the principal agent of regional demographic change. Despite its significance, analysis of migration lags behind equivalent scholarship in fertility and mortality, particularly with respect to cross-­ national comparisons. Comparative analysis of demographic processes is important for several reasons: it reveals commonalities and highlights unusual trends; it enhances methodological rigour; it aids theorisation; it also provides a firm foundation for the formulation of urban and regional policy (Bell et al. 2002). Within the field of migration studies, internal migration, that is, the propensity to change residence within national borders, has been accorded less attention than international migration, especially in a comparative context. This dearth of analysis is surprising since movements within countries outnumber international movements by a factor of four to one. This edited book analyses the way in which internal migration varies between the countries of Asia drawing on a common analytic framework developed as part of the IMAGE Project (Comparing Internal Migration Around the E. Charles-Edwards (*) · M. Bell · A. Bernard Asian Demographic Research Institute, Shanghai University, Shanghai, China Queensland Centre for Population Research, The University of Queensland, Brisbane, QLD, Australia e-mail: [email protected]; [email protected]; [email protected] Y. Zhu Asian Demographic Research Institute, Shanghai University, Shanghai, China School of Geography, Fujian Normal University, Fuzhou, Fujian Province, China e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_1

3

4

E. Charles-Edwards et al.

GlobE  – https://imageproject.com.au). By situating individual country analyses within a global setting, this book provides the first thorough understanding of how and why internal migration varies across the major regions of Asia. With a total population of 4.5 billion, Asia is the largest and most populous continent in the world, home to more than three–fifths of the human population, hosting seven out of the 13 countries that have total populations over 100 million. In recent decades Asia has experienced rapid population growth and tremendous social and economic development; it is the largest continental economy by Gross Domestic Product (GDP) in Purchasing Power Parity (PPP) terms. Some of the longest economic booms in the world since the 1950s (notably those in Japan, the four Asian tigers of South Korea, Singapore, Hong Kong and Taiwan, and more recently in Mainland China) have taken place here, leading to profound socioeconomic transformations of societies, with internal migration an integral part of this process. However, economic development and social transformation have been very uneven and have been unfolding differently in individual ethnic, cultural and geographical settings, underpinned by considerable diversity in trends and patterns of internal migration. Clearly, the vast size of Asia in terms of both area and population, the enormous scale of recent socioeconomic changes, and its diversity in levels of development and contextual settings, make Asia an ideal and important setting for advancing research on internal migration. Given that large-scale internal migration is a more recent phenomenon in Asia, and that research on it is much less developed compared to that in other parts of the world, internal migration in the countries of Asia deserves concentrated attention. While there is a long tradition of research on various forms of population mobility in Asia, evidence remains fragmented, reflecting a diversity of traditions in migration research with literature emanating from different disciplinary perspectives, and often focussed on particular issues or spatial settings. Comparative studies of internal migration are relatively few, but there have been a number of important contributions. Pryor (1979) undertook one of the earliest statistical comparisons of internal migration in Asia, examining moves within five countries: Thailand, Malaysia, Singapore, Indonesia and the Philippines. Among other things, the study underscored the primacy of the major metropolitan regions as migrant destinations, but also the role of state-sponsored development projects in channelling migration to the periphery of national settlement systems. Migration propensities were demonstrated to vary by age and other characteristics, with migrants typically younger and better educated than non-movers. While this work underscored the value of cross-national comparisons in identifying common patterns, it also highlighted the contingency of internal migration processes to different national settings, and the impact of data collection practices on findings. Pryor noted that migration patterns are ‘… essentially a reflection of their specific cultural and historical settings, their stage of modernisation and economic development, and the constraints and peculiarities of their census enumeration and processing procedures’ (Pryor 1979 p. 322). This work was followed in the early 1980s by a series of country monographs led by the United Nations Economic and Social Commission for Asia and the Pacific (ESCAP 1982). This programme of work explored the links between urbanisation and internal migration in a number of countries including Indonesia,

1  IMAGE-Asia: An Introduction

5

Malaysia, Pakistan, Philippines, Republic of Korea, Sri Lanka and Thailand. Findings confirmed the dominance of rural-to-urban migration throughout the countries of East and Southeast Asia, and variations in migrant selectivity by sex according to distance and purpose of move. In contrast, Skeldon’s (1985) overview of internal migration in the countries of South Asia (Bangladesh, India, Nepal, Pakistan and Sri Lanka) identified rural-to-rural migration as the most significant migration stream, with short-distance moves most common and marriage a primary motivator. Skeldon also highlighted the importance of circular or temporary migration as a substitute for permanent flows, with ‘bilocality’ an important feature of the South Asian migration system. The 1990s and 2000s were a relatively low point in comparative studies of internal migration in Asia, with academic attention instead directed towards understanding international flows (Hugo 2005). Since 2010, however, a number of volumes have emerged including the seminal work of Amrith (2011), which traced the historical evolution of Asian migration systems and White’s (2016) International Handbook of Migration and Population Distribution, which included a contribution on Internal Migration in Asia (Charles-Edwards et al. 2016). An important recent volume by Fielding (2015) synthesised existing research and empirical evidence to explore contemporary migration systems in Northeast, East and Southeast Asia. Fielding revealed marked regional variations across Asia reflecting varying levels of development. Contemporary internal migration in Northeast Asia was found to be characterised by inter-urban and counter-urban flows, characteristic of later stages of development. In contrast, rural-to-urban flows remained the dominant feature of migration systems in East and Southeast Asia, accompanied by migration to frontier regions facilitated by large scale development programmes. Taken together, this body of work highlights considerable regional variation in internal migration patterns and processes across Asia, but significant gaps remain, with a dearth of quantitative analysis in many countries and little known about internal migration in the countries of central and western Asia. This book aims to address these deficits by drawing on the methods and techniques developed as part of the IMAGE project.

1.2  The IMAGE Project and the Dimensions of Migration The IMAGE Project was a five-year international collaborative programme of research with core funding from the Australian Research Council Discovery Scheme (2010–2015) designed to explore the way internal migration varied between countries around the world. The project was organised around four modules: an Inventory of internal migration data collection practices; a Repository of internal migration data; a suite of robust Migration metrics that could be used to make reliable cross-­ national comparisons; and the IMAGE Studio, a suite of bespoke statistical software designed to implement the comparative measures and address one of the longstanding issues in migration analysis, the modifiable areal unit problem (MAUP).

6

E. Charles-Edwards et al.

The work of the IMAGE Project has been published in a series of thematic, methodological and regional papers. In terms of substantive emphasis, the distinctive feature of the Project was its identification of several discrete dimensions of migration, each of which, it was argued, provides a unique but complementary perspective on population movement. These are: migration intensity, which indicates the overall level of migration or propensity to move; age composition, which denotes the selective nature of migration (and emerges as a key factor moderating intensity); spatial impact, which measures the effect exerted by migration in transforming the pattern of human settlement; migration distance, which identifies the frictional effect of space on the propensity to move; and connectivity, which recognises the way in which migration serves to establish functional linkages between different parts of the settlement system. For this book, attention is focussed on the first three of these dimensions — migration intensity, age composition and spatial impact — since these have the greatest utility in understanding the marked differences in human population movement that exist between countries, the most value in building migration theory, and the most relevance to formulation of social and economic policy. It will be helpful to elaborate on the significance of these three dimensions.

1.2.1  Migration Intensity Migration intensity captures the overall propensity to migrate within a population. It is extremely sensitive to the spatial scale at which migration is recorded, with the probability of moving inversely related to the size of the geographic units across which migration is measured. For this reason, the IMAGE Project adopted a measure of aggregate migration intensity that measures all changes of residential address. Early work by Long (1991) revealed significant variations in the aggregate migration rate across 14 countries, with one-year migration rates ranging from 6% in Ireland to 19% in New Zealand. The IMAGE project adopted new techniques that enabled this to be extended to 96 countries (Bell et al. 2015), underlining the massive variation that exists between countries. One-year migration rates varied from just 1% in Macedonia to 19% in Iceland, while migration rates measured over 5 years ranged from 5% in India to 55% in New Zealand. Comparisons within Asia (Charles-Edwards et al. 2019) have shown similar heterogeneity, with the highest mobility intensities measured over 5 years recorded in South Korea (53%) while the lowest aggregate migration intensity was in India, with just 5% of the population changing address over a five-year period. How does one account for these variations? Explanations have been sought using both macro and micro-level approaches. The most common macro-level theory is found in the migration transition framework proposed by Zelinsky (1971), which suggests that migration intensity undergoes a systematic transition as countries progress in their economic and social development, driven partly by regional differences in economic opportunity. This initially gives rise to large rural-to-urban flows, followed by urban-to-urban flows as the urban transition comes to a close. In

1  IMAGE-Asia: An Introduction

7

super-advanced societies, migration intensities may start to decline from their peak as regional differences are ameliorated, and transport and communication technologies substitute for geographic mobility. Zelinsky’s hypothesis of a mobility transition has been criticised as Eurocentric and time-bound (see e.g., Cadwallader 1993) but empirical work confirms clear associations between migration intensities and various indicators of development (Bell et al. 2015). Structural factors such as age composition (Bell et al. 2015) and the nature of housing markets have also been implicated (see e.g., Caldera-Sanchez and Andrews 2011). Micro-level approaches to comparative analysis of migration intensity have sought explanations by referring to migrant selectivity, housing adjustment and transitions in the life course. Bernard (2017) employed a cohort perspective to explore differences across 14 countries, demonstrating that age at first migration was a key indicator and that migration intensity was lower in countries in which the first move was delayed. The cumulative evidence suggests that cross-national differences are not a simple reflection of either macro or micro-level factors but are shaped by processes at both levels. Explanation for cross-national differences therefore ultimately needs to bridge these modes of explanation. Current thinking, as further discussed in Chap. 2, also underlines the importance of regional and local contexts in understanding the factors that trigger or inhibit migration.

1.2.2  Age at Migration Age at migration is the second dimension of mobility explored in this volume. The selective nature of migration has long been established (Thomas 1938, 1958). In the 130 years since Ravenstein’s seminal papers (Ravenstein 1885, 1889), the search for universal laws of migration has largely been abandoned, but solid empirical evidence points to a positive association between migration and income, education and occupational status, with further differences according to housing tenure and marital status. The most persistent regularities, however, are found in relation to age. Rogers and Castro (1981) showed that the age profile of migration displays a remarkably consistent shape through space, time and across spatial scale, characterised by a peak among young adults, with lower rates of movement at older ages and among teens, rising again among retirees, the elderly and the very young. More recent work has revealed subtle but significant variations in key aspects of this profile, especially the age at which migration peaks, and the extent to which migratory activity is concentrated around that peak. Bernard et al. (2014a) proposed two measures to facilitate cross-national comparison of migration age profiles – age at peak and intensity of peak – and showed that together these measures capture two-thirds of the variance between countries. Applied to a global sample of 25 countries they revealed a striking regional pattern. Five countries in Asia (Malaysia, Vietnam, China, Nepal and Indonesia) formed a distinct cluster with migration strongly concentrated and peaking at younger ages than in industrialised countries or Latin America. Bernard et  al. (2014b) sought to explain these variations by

8

E. Charles-Edwards et al.

reference to differences in the life course with key events occurring earlier and in more concentrated form in Asia than in other parts of the world. Is this same pattern replicated in other parts of Asia? As in the case of migration intensity, these findings point to the importance of understanding how contextual factors such culture, religion and the level of economic development shape migration age profiles. In Chap. 2 we review the life-course framework and the way its proximate determinants, such as educational attainment and age at first marriage, are thought to shape the age profile in countries across the region. Chapter 3 explains the metrics that are used.

1.2.3  Spatial Impact Spatial Impact is the third dimension of migration explored in this volume and is arguably the most significant aspect of population movement in terms of policy and planning. Spatial movements are of longstanding scholarly interest in Asia, primarily because of their role in the rapid urbanisation, which has accelerated across much of the region since the 1950s (ESCAP 1982; Pryor 1979). In practice, however, comparative studies have been hindered by differences in the classification of urban areas, and in the way migration data are collected, which effectively preclude clear identification of the role of migration in the urbanisation process. Following Rees and Kupiszewski (1999), the IMAGE project adopted an alternative approach to this problem by using regional population density as a proxy for the level of urbanisation and proposed a conceptual model with a trajectory linking migration to population density as development proceeds (Rees et  al. 2017). We discuss this model in more detail in Chap. 2, explain its computational basis in Chap. 3, and then test its utility in subsequent chapters. As Fielding (2015) points out, however, internal migration in the countries of Asia has been driven by a variety of forces including resource development, defence, government policies and natural disasters, not just the attraction wrought by burgeoning primate cities. Moreover, redistribution of population involves a complex web of migration flows and counter-flows. As with other dimensions of population movement, understanding these migratory streams requires close consideration of the socio-political context in which migration occurs. We use maps and circular plots to visualise the impact of these flows, but reliable cross-national comparisons ultimately call for robust metrics that capture the scale and spatial impact of these movements. For this we report the Aggregate Net Migration Rate (ANMR) and show how this is ultimately shaped by the interaction between migration intensity and migration effectiveness, the latter capturing the balance between internal migration flows and counter-flows. The three dimensions of internal migration discussed above form the foundation for the exploration of internal migration in the 15 countries of Asia presented in this book. The IMAGE project made significant advances in development of metrics to facilitate empirical cross-national comparisons across multiple dimensions of migration. Statistical indicators provide the essential framework against which to situate the scale, composition and patterns of migration, but interpretation of the

1  IMAGE-Asia: An Introduction

9

dynamics, causes and consequences of such movements calls for a nuanced understanding of the context within which this mobility occurs. To echo the views advanced by Pryor (1979), there is a clear and pressing need to view quantitative metrics against the backdrop of each nation’s geography and history, having close regard for its economic, socio-cultural and political setting. Only by coupling robust metrics with a well-founded appreciation of the forces that shape migration is it possible to simultaneously understand population movements within a country and to make credible comparisons with those in other nations. This book represents an explicit attempt to achieve this goal by coupling the empirical framework established by the IMAGE Project with the in-depth knowledge supplied by scholars with country-specific expertise. In this way, the objective is to provide systematic new insights into the patterns and drivers of internal migration in the countries of Asia.

1.3  IMAGE-Asia The programme of research underpinning this book, IMAGE-Asia, began in 2017 with discussions between scholars at the University of Queensland and the Asian Demographic Research Institute (ADRI) at Shanghai University. The aim of the project was to enhance substantive understanding of internal migration across the countries of Asia and generate new insights into internal migration processes. To borrow from the language of Graeme Hugo (1975 p. 25), ‘conducting cross-national comparative research is ultimately the art of the possible’, and there were a number of obstacles to overcome. Most fundamental was access to internal migration data at the geographic scale needed to generate comparative metrics of intensity, age and spatial impact. Such data are generally accessible in the countries of East, Southeast and South Asia but less readily available moving westward into Central and Western Asia. There is a similar gradient in the depth of expertise and scholarship on internal migration across the continent. Another impediment was the recruitment of scholars with a thorough understanding of both the dynamics and drivers of internal migration in each country, who were also comfortable with the approach laid out by the editors. As discussed further below, this approach involved adopting a clearly defined structure with discrete sections devoted to the national setting, a review of the available data, a summary of prior research and a systematic discussion and interpretation of migration focussing on the three dimensions previously outlined. To assist with this task, authors of each of the country chapters were provided with a series of analytic outputs capturing migration intensity, age at migration and spatial impact, computed centrally at the University of Queensland and at ADRI using the IMAGE Studio. Maps of net migration and circular plots depicting inter-regional flows for each country were also made available. At the same time, contributors were also encouraged to incorporate information from other country-specific sources of migration data such as national surveys. A key phase in the research programme was a two-day workshop held at Shanghai University in July 2018, funded by the ADRI. Scholars from 20 countries were invited to attend and to present first

10

E. Charles-Edwards et al.

drafts of their contributions. This workshop proved to be a highly effective mechanism for reinforcing the goals of the project, building technical skills, especially among contributors with less prior knowledge of quantitative measures of migration, and for identifying common themes. Subsequently individual authors have worked closely with the editors to refine their contributions. As with all such endeavours, not all participants at the Shanghai workshop were able to follow through to completion, owing to a range of personal and professional circumstances, so sadly we are missing analyses from several countries that might have been included. Nevertheless, the book includes case studies for a final list of 15 countries with representation from all regions of this widely dispersed continent.

1.4  Structure of This Volume This book endeavours to blend a systematic approach to the measurement of migration with an in-depth understanding of the forces that shape population mobility across an array of countries with diverse bio-physical and socio-political settings. With growing recognition of the way local context shapes human behaviour, migration scholars have moved beyond the search for any single grand theory that might explain population movement. Nevertheless, simple empirical observations that prioritise uniqueness and eschew generality offer little hope of any serious advance in understanding. As Nourse (1968) would have it, facts without theory are sterile: theories are needed to link facts together and provide a framework for interpretation. We provide this framework in Chap. 2, using a tripartite conceptual schema that endeavours to account for the variation we find between countries in the three dimensions of migration that are our focus: migration intensity, age composition and spatial impact. We briefly review previous approaches to theorising migration and endeavour to sketch the links between elements of our tripartite schema. For each of these three dimensions, Chap. 3 defines the statistical indicators that we employ to measure migration, explains their rationale and outlines how they are computed. By way of background, we also describe the development of the global IMAGE project from its origins in 2002, review four major groups of impediments to cross-national comparisons and explain how these have been addressed by the IMAGE methodology. Here too we provide an overview of the IMAGE Studio, the bespoke software system that was created specifically to compute the migration metrics and address the modifiable areal unit problem, one of the longstanding issues in spatial analysis that inhibit comparative analysis. Together with this Introduction, Chaps. 2 and 3 make up Part I of this book, collectively labelled The Framework for the analysis which follows. Part II contains 15 chapters, each focussing on a single country in Asia. These are divided into four regional groupings which, for simplicity, we arrange in turn, moving from east to west across the continent. In East Asia our contributors examine migration in two of the most highly developed countries in the region, Japan and South Korea, together with the most populous, China, and perhaps the most remote,

1  IMAGE-Asia: An Introduction

11

Mongolia. These four countries alone present a remarkable diversity of landscapes, histories, cultures, politics and economic development. Southeast Asia is represented by Thailand, Cambodia and Myanmar, again illustrating variety of histories, stages of development and political organisation. In South Asia we examine population movements in countries with the largest (India) and smallest (Bhutan) territories, together with Sri Lanka, Nepal and Iran, each of which have a distinctive physical and cultural setting. West and Central Asia have long been blank spots on the atlas of population mobility but they are represented here by three countries that also present extreme contrasts, both with each other, and with their eastern neighbours: Israel, Armenia and Kazakhstan. Each chapter has been written by a scholar or team of scholars with in-depth knowledge of their subject country and each chapter follows a standard format including a country overview, a summary of prior research, a discussion of internal migration data and an overview of the spatial framework followed by substantive analysis of migration intensity, migrant characteristics, and the spatial impacts of migration. Collectively, these 15 chapters comprise The Evidence as to the nature of internal migration in the countries of Asia that constitutes Part II of the book. By way of conclusion, Part III of the book, entitled Synthesis, contained in Chap. 19, endeavours to provide a synthesis of the empirical findings, and to draw out key themes and insights from the preceding chapters. We include a series of league tables that formally compare the countries of Asia in terms of the key metrics used to measure our three dimensions of migration, and to situate Asia in its wider global context. We also revisit the conceptual frameworks advanced in Chap. 2, assess their utility and suggest possible avenues for enhancement. We hope that the coupling of a strong empirical framework situated within in-depth contextual interpretation will provide new insights into internal migration in Asia and help to further advance understanding of population mobility in Asia, and globally.

1.4.1  Caveat Following the approach adopted by the United Nations (2019; ii) the designations employed in this book and the material presented in it do not imply the expression of any opinions whatsoever on the part of the editors concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.

References Amrith, S. (2011). Migration and diaspora in modern Asia. New York: Cambridge University Press. Bell, M., Blake, M., Boyle, P., Duke-Williams, O., Rees, P., Stillwell, J., & Hugo, G. (2002). Cross-national comparison of internal migration: Issues and measures. Journal of the Royal Statistical Society A, 165(3), 435–464. https://doi.org/10.1111/1467-985X.t01-1-00247.

12

E. Charles-Edwards et al.

Bell, M., Charles-Edwards, E., Ueffing, P., Stillwell, J., Kupiszewski, M., & Kupiszewska, D. (2015). Internal migration and development: Comparing migration intensities around the world. Population and Development Review, 41(1), 33–58. Bernard, A. (2017). Levels and patterns of internal migration in Europe: A cohort perspective. Population Studies, 71(3), 293–311. Bernard, A., Bell, M., & Charles-Edwards, E. (2014a). Life-course transitions and the age profile of internal migration. Population and Development Review, 40(2), 231–239. Bernard, A., Bell, M., & Charles-Edwards, E. (2014b). Improved measures for the cross-national comparison of age profiles of internal migration. Population Studies, 68(2), 179–195. Cadwallader, M. (1993). Classics in human geography revisited: Commentary 2. Progress in Human Geography, 17(2), 213–219. Caldera-Sanchez, A., & Andrews, D. (2011). To move or not to move: What drives residential mobility rates in the OECD? (OECD Economics Department Working Papers, No. 846). Paris: OECD Publishing. Charles-Edwards, E., Muhidin, S., Bell, M., & Zhu, Y. (2016). Migration in Asia [other than China and India]. In M. J. White (Ed.), International handbook of migration and population distribution. New York: Springer. Charles-Edwards, E., Bell, M., Bernard, A., & Zhu, Y. (2019). Internal migration in the countries of Asia: Levels, ages and spatial impacts. Asian Population Studies, 15(2), 150–171. https://doi. org/10.1080/17441730.2019.1619256. ESCAP. (1982). Comparative study on migration, urbanization and development in the ESCAP region (RAS/P13/79). Population Research Leads, 6, 1–36 Revised version. United Nations, Economic and Social Commission for Asia and the Pacific. Fielding, A. (2015). Asian migrations: Social and geographical mobilities in Southeast, East and Northeast Asia. Abingdon: Taylor and Francis. Hugo, G.  J. (1975). Conducting research into population mobility in Java: The art of the possible. In R. J. Pryor (Ed.), The motivation of migration, studies in migration and urbanisation (pp. 25–27). Canberra: Department of Demography, Australian National University. Hugo, G. (2005). The new international migration in Asia. Asian Population Studies, 1(1), 93–120. https://doi.org/10.1080/17441730500125953. Long, L. (1991). Residential mobility differences among developed countries. International Regional Science Review, 14, 133–147. Nourse, H. O. (1968). Regional economics. New York: McGraw-Hill. Pryor, R. (1979). Migration and development in South-East Asia. Kuala Lumpur: Oxford University Press. Ravenstein, E.  G. (1885). The laws of migration. Journal of the Statistical Society of London, 48(2), 167–235. Ravenstein, E. G. (1889). The laws of migration. Journal of the Royal Statistical Society, 52(2), 241–305. Rees, P. & Kupiszewski, M. (1999). Internal migration and regional population dynamics in Europe: A synthesis. International Journal of Population Geography. Strasbourg: Council of Europe Publishing. Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), 1–22. Rogers, A. & Castro, L.J. (1981). Model migration schedules. International Institute for Applied Systems Analysis. Laxenburg, Austria. ISBN 3704500224. Skeldon, R. (1985). Migration in South Asia: An overview. In L.  A. Kosiński, K.  M. Elahi (Eds.), Population redistribution and development in South Asia. Geo Journal Library, Vol 3. Dordrecht: Springer. Thomas, D.  L. (1938). Research memorandum on migration differentials. New  York: Social Science Research Council.

1  IMAGE-Asia: An Introduction

13

Thomas, D. L. (1958). Age and economic differentials in interstate migration. Population Index, 24(4), 313–325. United Nations. (2019). World population prospects  – 2019 highlights. Population Division, Department of Economic and Social Affairs. New York: United Nations. ST/ESA/SER.A/423. White, M. J. (2016). International handbook of migration and population distribution. New York: Springer. Zelinsky, W. (1971). The hypothesis of the mobility transition. Geographical Review, 61(2), 219–249. https://doi.org/10.1177/030913259301700205.

Chapter 2

Understanding Internal Migration: A Conceptual Framework Aude Bernard, Martin Bell, Elin Charles-Edwards, and Yu Zhu

2.1  Introduction Theoretical approaches to understanding migration are diverse and extensive, with a literature emanating from different disciplinary perspectives, including demography, geography, economics, sociology and anthropology (White 2016). From this body of work two main streams of research can be identified: the macro-level perspective, which is concerned with areal aggregates, and the micro-level perspective where individuals are the unit of analysis. Widely popular in the 1960s and 1970s, the macro-level approach largely disappeared from migration studies in the 1980s (Skeldon 2018), mainly because of the Modifiable Areal Unit Problem (MAUP) (Openshaw 1984) and the ecological fallacy (Robinson 1950), both of which plagued early migration studies. The MAUP refers to the fact that measures of migration are affected by the number and shape of the areal units into which a nation is divided. The absence, until recently, of satisfactory solutions to these twin problems of scale and zoning contributed to a decline of interest in macro-level analysis of migration in the 1980s (Fotheringham et  al. 2000). The retreat from analysis of aggregate data was further fuelled by the ecological fallacy, which arises from using aggregate data to draw inferences about the characteristics and relationships between individuals (Voss 2007). A. Bernard (*) · M. Bell · E. Charles-Edwards Asian Demographic Research Institute, Shanghai University, Shanghai, China Queensland Centre for Population Research, The University of Queensland, Brisbane, QLD, Australia e-mail: [email protected]; [email protected]; [email protected] Y. Zhu Asian Demographic Research Institute, Shanghai University, Shanghai, China School of Geography, Fujian Normal University, Fuzhou, Fujian Province, China e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_2

15

16

A. Bernard et al.

As a result of this paradigm shift, a rich and growing literature based on individual-­level analysis has developed in the last 30 years tied to (1) improvements in data collection and accessibility, first with the advent of public use census micro-­ data files and then the emergence of longitudinal surveys, and (2) advances in research methods, including event-history analysis (Kulu and Milewski 2007; Pelikh and Kulu 2018), agent-based simulation (Klabunde and Willekens 2016) and sequence analysis (Vidal and Lutz 2018). After a lengthy absence, macro-level analysis of migration has regained not just acceptability, but also popularity, with an increasing number of studies adopting aggregate-level modes of analysis, particularly to explain variations in migration processes across space and time. Examples of such work include the IMAGE Project from which this book is derived, which sought to compare internal migration levels, patterns and processes in countries across the globe, and the DEMIG (Determinants of International Migration) and MADE (Migration and Development) projects, which take the mobility transition as their starting point. This recent shift is underpinned by a growing interest in understanding social processes within their spatial context (Goodchild and Janelle 2010) and the increasing recognition that migration is fundamentally a spatial process (Raymer et al. 2018). The re-emergence of macro-level approaches to migration has been made possible by a series of methodological advances, including efforts to address the MAUP, and increasing availability of migration data from a wide range of countries, both of which are discussed in Chap. 3. Both micro and macro- approaches have a place in migration research and should proceed in tandem (Billari 2015). Because they differ not only in the type of data and research methods employed but also in the nature of the questions addressed, micro and macro-modes of explanation provide complementary insights into migration behaviour and processes. In this book, we take an explicitly macro-level approach to measuring, understanding and explaining variations in internal migration among the countries of Asia and we do so for three main reasons. First, migration processes are embedded within a spatial context (Voss 2007), which is often missed in individuallevel analyses of migration that remain largely a-spatial. The emergence of multilevel modelling and geographically-weighted regressions has permitted consideration of the broader socioeconomic context in which individual lives are embedded and this has contributed to bridging of the gap between micro and macro-level perspectives and to the emergence of the ‘crucial meso-level’ (Faist 2010). Yet, the direction of population movement is largely left out of individual-­level approaches. More generally, micro-analyses tend to focus on particular aspects of migration, including the selectivity of migrants (Bernard and Bell 2018; Greenwood 2014), the impact of family and social networks on migration decision processes (Mulder 2018; Palloni et al. 2001) and the role of life-course transitions in triggering a change in place of residence (Mulder 1993; Sander and Bell 2014). Thus, despite their merits, current microlevel explanations of migration do not offer a conceptual understanding of migration in a way that explains cross-national variations in internal migration. While attempts have been made to compare migration behaviour between countries at a micro-level (Bernard et al. 2016; Mulder et al. 2002; Vidal et al. 2017), such approaches fail to provide a comprehensive view of population movement. As noted in the previous chapter, migration is a multi-­dimensional process, with each dimension providing

2  Understanding Internal Migration: A Conceptual Framework

17

different insights into population movement, and as such migration cannot be adequately summarised and understood with a single summary statistic. Thus, we argue that the macro-­perspective is better suited to identifying and understanding crossnational variations in migration by providing population-level insights into migration processes across a range of dimensions. By revealing similarities and differences between countries, macro-level analysis can in turn help theorisation. In that context, this chapter does not purport to provide an exhaustive review of migration theory,1 but aims instead to advance a tripartite conceptual schema as a framework for understanding the three dimensions of migration identified in Chap. 1. By way of background, we first review two broad bodies of theory that have been highly influential in the search for understanding of cross-national variations in migration, namely Zelinsky’s Hypothesis of a  Mobility Transition and Skeldon’s thesis linking Mobility and Development. Drawing on this review, the chapter then sets out a tripartite conceptual framework against which it is proposed that the three dimensions of migration can be interpreted, linking (1) migration intensity to national economic development, (2) the age profile of migration to life-course transitions and (3) migration to evolution of the settlement pattern.

2.2  A Review of Macro-level Migration Theories The classical theoretical foundations of migration research date from the late nineteenth century with Ravenstein’s Laws of migration (1885, 1889), which represent, with his earlier papers, the first detailed analyses of internal migration. Drawing on empirical regularities from the 1871 and 1881 British censuses, Ravenstein proposed 11 laws as listed in Table 2.1, which have been shown to be empirically valid, Table 2.1  Ravenstein’s laws of migration 1. Most migrants move over short distances. 2. Migration is a step by step process. 3. Long-distance migration is mainly to major industrial and commercial centres. 4. Most migration flows are from agricultural to industrial areas. 5. Individuals born in urban areas are less likely to migrate than those of rural areas. 6. Migration drives the growth of large towns to a greater extent than natural increase. 7. The size of migration flows grows with the development of industries, commerce and transport. 8. Each migration stream has a counter-stream. 9. The majority of migrants are independent adults, particularly over long distances. 10. Women migrate more than men domestically, but less internationally 11. The major causes of migration are economic. Source: Ravenstein (1885, 1889)

1  For comprehensive reviews of migration theory see for example Massey et al. (1993); Boyle et al. (1998); de Haas (2010); Fielding (2015); Skeldon (1997); White (2016).

18

A. Bernard et al.

except perhaps for the observation that urban dwellers are less mobile than their rural counterparts, which reflects the time when Ravenstein wrote. While his propositions have been criticised as descriptive, deterministic and historical (Castles and Miller 1993), they ‘provided the hypotheses upon which much future migration research and theorisation was built’ (Boyle et al. 1998 p. 59) including gravity and human capital models (Greenwood 2019). Ravenstein’s theoretical proportions are simple yet broad-ranging, and they address a wide range of questions that are still relevant today: who migrates? why do they migrate? where do they migrate? and how often do they migrate? Wright and Ellis (2016) outlined the prescience of each law in the way it has shaped migration theory and scholarship in geography. In his seventh law, Ravenstein anticipated a link between migration and development arguing that migration increases in volume as industries and commerce develop and transport improves, but he did not consider the possibility of changes in the direction of migration flows. He wrote at a time of rapid industrialisation where rural-to-urban flows dominated migration systems in Europe, yet in his eighth law, he observed that for ‘each main stream or current of migrants there runs a counter-current, which more or less compensates for the losses sustained by emigration. This counter-current is strong in some cases, weak in others, and literally compensatory in a few instances’ (1885 p.  187). In laws nine and ten, Ravenstein anticipated the selectivity of migration though he did not mention age. It is not until the seminal work of Thomas (1938) on migration differentials in the United States that the age selectivity of migrants was observed. Since then this pattern has been repeatedly verified at a range of spatial scales and across a broad spectrum of countries (Bell and Muhidin 2009; Rogers and Castro 1981). Age has become a singular focus of subsequent scholarship as migration plays a particularly significant role in altering the age structure of populations at origin and destination. Ravenstein is also credited with the first cross-national comparison of internal migration. In his 1889 paper, he extended his search for empirical regularities to over 20 European countries, Canada and the United States by examining lifetime net migration aggregates and he concluded that migratory movements follow the same principles in all countries. His scholarship extends beyond cataloguing laws; he viewed migration as a response to spatial differentials in economic opportunities, conditioned by basic geographic characteristics and transportation systems, but also by regulations and climatic conditions. Ravenstein did not, however, consider how migration behaviour is itself conditioned by the demographic structure of the population (Greenwood 2019) and it was not until Zelinsky’s hypothesis of a mobility transition (1971) that this dimension was systematically considered in migration studies. In his seminal 1971 paper, Zelinsky linked the level and direction of migration to progress through the demographic transition, arguing that there were ‘definite patterned regularities in the growth of personal mobility through space-time’ (1971 p.  221–22) and that these formed an integral part of the modernisation process, which he referred to as the mobility transition. Not only was Zelinsky’s model the first attempt to bring migration into the same framework as mortality and fertility, it also anticipated a fifth phase not included in the original formulation of the

2  Understanding Internal Migration: A Conceptual Framework

19

Table 2.2  Zelinsky’s stages of the migration transition Phase 1 – pre-modern traditional society Limited moves that relate to traditional practices Phase 2 – early transitional society Large-scale migration from rural areas to cities and growth in circulation Frontierward migration within country Emigration to attractive foreign destinations and small immigration of specialised workers Phase 3 – late transitional society Continuing rural-to-urban migration but at a decreasing rate Emigration and frontierward migration are on the decline Growth and complexification in circulation Phase 4 – advanced society Rural-to-urban migration declines further Increase in migration between cities and within urban agglomerations Large-scale immigration of unskilled and less-skilled workers from less developed countries Significant international migration or circulation of skilled migrants and professionals Phase 5 – super advanced society Better communication contributes to a decline in residential migration The majority of internal migration is inter-urban or intra-urban Immigration of unskilled labour from less developed countries continues but with strict political control Increase in some forms of circulation, including new forms Source: Zelinsky (1971)

demographic transition, coined ‘future super advanced societies’. Zelinsky considered not only changes in the volume of migration but also in the direction of flows by linking each phase of the migration transition to different forms of migration: international, frontierward, rural-to-urban, urban-to-urban and intra-urban and circulation as shown in Table 2.2. He anticipated a decline in rural-to-urban migration, an increase in intra-urban migration and an overall decrease in migration levels because of technological progress, an idea supported by Jones and Brown (1985). According to Zelinsky, cross-national variations in migration levels could be interpreted as reflecting different stages of the development ladder. Rooted in modernisation theory, the idea of single mobility transition has been criticised as Eurocentric and time bound (Woods et al. 1993) and Zelinsky himself extended and modified his original views (1979, 1983). Yet, because of the retreat of macro-level approaches, the hypothesis of a migration transition progressively fell out of favour and migration studies published in the 1990s and 2000s made scant reference to the idea of a migration transition, with the notable exceptions of Boyle et al. (1998) and Skeldon (1997). As noted by Skeldon (2018), it is not until the fifth edition of the Age of Migration (Castles et al. 2014) and the involvement of de Hass on the editorial team that the book engaged with Zelinsky’s ideas, over 20 years after the first edition. However, in recent years the mobility transition has regained traction. For example, King (2012) included Zelinsky as one of the four major theoretical contributions made by geography to migration studies along with

20

A. Bernard et al.

the work of Ravenstein, Mabogune and Hägerstrand. In recent historical accounts of migration, and the demographic and urban transitions in nineteenth-century Europe, Bocquier and Costa (2015) highlighted the inter-relationship between these processes, while emphasising the role of economic and institutional changes in shaping their evolution. In a recent review of Zelinsky’s contributions to migration, Skeldon (2018, p.1) concluded that ‘rather than an inflexible linear model, the hypothesis provides scope to incorporate multiple pathways of changing patterns of migration through time and space’. Cooke et al. (2018, p.503) reached a similar conclusion, stating that ‘the mobility transition is not an obsolete frame of reference but a prescient pliable and adaptable framework which not only informs the study of human geographic mobility today but also, perhaps, even into the future’. Recognising that there is no single pathway through a migration transition and that not all countries will process linearly from one phase to the next, subsequent scholarship turned to context-specific approaches in search of distinctive sets of migration sequences that reflect the particular economic circumstances, social structures and population and settlement systems of a country or region. An example of such approach can been seen in the global regionalisation perspective that Skeldon (1997) took in identifying five macro-level regions, each with its particular type of migration transition. For example, China falls into the expanding core where immigration and emigration occur alongside internal centralisation (i.e., rural-to-­ urban migration). On the other hand, the Philippines belongs to the ‘labour frontier’ cluster, which is dominated by out-migration and internal centralisation, while parts of central Asia are in the so-called ‘resource niche’ with weaker forms of migration. Such groupings highlight the wide variations that exist in migration processes across the Asian region. Such conceptualisation enables the integration of migration and broad aspects of development into a single spatial-temporal model (De Haas 2005, 2010) by offering a more nuanced approach to understanding migration and development interactions. As noted by Skeldon (2012, p.  136), ‘the future direction of the development of migration theory is likely to be much more specific and contextual in specific states and specific regions’. At the same time ‘while no single pathway through any migration or developmental transition exists, it nevertheless needs to be accepted that a retreat to total relativism is counterproductive’ (2012, p. 155). We also believe that there is value in searching for commonalities and identifying differences between countries, but for this endeavour to be meaningful and fruitful we need to understand migration systems in light of the wider social, economic and political transformations within which they are embedded. This calls for context-specific approaches, thus recognising that migration-development interactions are dependent on the context in which migratory systems are embedded. The current volume seeks to achieve this objective by bringing together a series of case studies from diverse Asian countries using a common suite of metrics and a unified conceptual approach.

2  Understanding Internal Migration: A Conceptual Framework

21

2.3  A  Multi-dimensional Framework for Comparing Migration The complexity and diversity of migration experiences and processes have long been obstacles to theory formulation (Castles 2010). At the same time, a comprehensive theory encompassing all aspects of migration would require a high level of abstraction, which would weaken its explanatory power (Portes and DeWind 2004) and, for that reason, there has been a progressive retreat from grand unifying theories of migration since Zelinsky’s hypothesis of a mobility transition (1971). While our aim is not to propose an all-encompassing conceptual framework, we recognise that migration is multi-dimensional and we seek to incorporate three different aspects of migration into our conceptual understanding – migration intensity, age selectivity and spatial impact – as each of these provides a unique perspective on population movement (Bell et  al. 2002). A series of papers emanating from the IMAGE project have explored each facet of migration separately and proposed conceptual frameworks for understanding and explaining variations observed between countries for each of these dimensions. The first attempt to systematically compare multiple dimensions of migration between countries was undertaken by Bell and colleagues in a series of papers comparing migration processes between Britain and Australia (Bell et al. 2002; Stillwell et al. 2000, 2001). These studies established the essential foundation for the theoretical and methodological development that emerged in the IMAGE project a decade later. Subsequently Bernard et al. (2017) examined multiple dimensions of population movement in a single encompassing analytical framework by comparing the overall intensity, age composition, spatial patterns and distance profile of internal migration in 19 Latin American countries, and Rowe et  al. (2019) provided an equivalent analysis for Europe. While these papers identified a range of migration sequences shaped by local contingencies, leading to unique migration-development interactions, they did not advance a theoretical schema for understanding the outcomes. In this section, we outline a conceptual framework for understanding cross-national variation with respect to each of the three dimensions of migration: intensity, age composition and spatial impact. We then sketch possible inter-relationships between these aspects of migration, the underlying premise being that simultaneous analysis across a large sample of countries within the same region should reveal the extent to which difference facets of migration are systematically related, while also highlighting the role of the wider socioeconomic context in shaping migration processes that reflect the specificity of the national context.

22

A. Bernard et al.

2.4  Migration Intensity and Development The overall intensity of internal migration  – the propensity to migrate within national borders – has long been thought to be linked to the level of development of a country. Recent empirical evidence suggest that countries at advanced stages of development as measured by the Human Development Index report higher internal migration intensities than those at lower levels of development (Bell et al. 2015). While this relationship appears to hold both globally and regionally (Bernard et al. 2017; Charles-Edwards et al. 2019), the presence of outliers suggest that it is not a linear, unidirectional process. In addition, some advanced economies have reported a progressive decline in migration intensities (Champion et al. 2018). This downward trend has been particularly pronounced in the United States, which recorded a 50 per cent fall in interstate migration intensities over the 30-year period to 2015 (Foster 2017), but it has also been observed in countries around the globe including Japan and South Korea (Bell et al. 2018). A series of explanations have been proposed, which relate to wider demographic, technological and economic transformations. These include population ageing (Cooke 2013; Foster 2017), progress in technology and telecommunications (Cooke and Shuttleworth 2017) as anticipated by Zelinsky (1971) and progress toward spatial equilibrium as differences reduce between regions in their employment opportunities, thus limiting the impetus for migration (Partridge et al. 2012). This downward trend is, however, not universal; some European countries have recorded stable migration intensities while other have experienced an upswing in the last two decades (Bell et al. 2018). In this book, we adopt a broad perspective on development as contributors seek to understand migration intensities by reference to a range of forces that shape mobility. In the concluding chapter we bring these forces together by setting cross-national variations against a battery of statistical indicators that capture economic, social, political and demographic differences between the countries of Asia. There are different representations of the forces acting on migration, but the simple diagram shown in Fig.  2.1 brings together the key elements operating on macro, micro and meso-­ levels. The demographic forces, however, call for earlier attention since age itself acts to shape migration directly, both through the age composition of the population, and through the age at which migration occurs.

2.5  Age Composition and Life-course Transitions Age has been adopted as a primary explanatory variable in the micro-level tradition, particularly in the life-course perspective that relates migration behaviour to particular stages of the life course (Mulder 1993). At the same time it is well-­established that migration follows a regular profile and thus offers a bridge between micro and macro-level theories. The near-universal peak in migration propensity among young adults (Rogers and Castro 1981) means that younger populations will experience

2  Understanding Internal Migration: A Conceptual Framework

23

Fig. 2.1  The drivers of migration. (Source: Adapted from Black et al. 2011)

higher overall migration intensities than older populations. While early scholarship focussed on regularities in migration age patterns in Europe and North America (Rogers and Castro 1981) and Asia (Ishikawa 2001; Kawabe 1990), recent studies have highlighted the substantial variations that exist between countries in the age and level at which migration peaks among young adults (Bell and Muhidin 2009). This work has revealed that migration tends to occur earlier and be concentrated over a narrower age range in Asia than in other world regions (Bernard et al. 2014a). This can be seen in Fig. 2.2, which compares the age profiles of migration in China, Brazil and Portugal. To explain these variations, we use a proximate determinant framework adopted from Bongaarts’ (1978) model of fertility and formulated by Bernard et al. (2014b) where life-course transitions act as an intermediary between contextual factors and migration outcomes. Stylised in Fig.  2.3 the framework suggests that contextual factors such as economic cycles and cohort sizes shape the structure and timing of the life course, which in turn determine migration age patterns, i.e., the age profile of migration broadly mirrors the age structure of key life-course transitions. Important changes in the transition to adulthood have been occurring across Asia over recent decades (Xenos et al. 2006), including increasing variability in the timing of transitions to adult roles (Tian 2016; Yeung and Alipio 2013), which are expected to result in variations between countries in the age patterns of migration. The age at first adult migration shapes the number of adult migrations: the younger the age at first migration, the higher the number of migrations through

24

A. Bernard et al. 0.05

China Brazil Portugal

Migration intensity (%)

0.04

0.03

0.02

0.01

0.00

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

Age

Fig. 2.2  Cross-national variations in migration age patterns, selected countries. (Source: Bernard et al. 2014a)

Fig. 2.3  Proximate determinants of migration age patterns. (Source: Bernard et al. 2014b)

adulthood (Bernard 2017a, b). In addition, the probability of migration increases with the number of past migrations (Bernard and Vidal 2019; Clark and Lisowski 2018; De Jong 2000). Attitudes towards migration are not fixed but evolve in a recursive manner in response to past moves. Collectively these results suggest that migration is a cumulative process that takes place over the entire life course of individuals rather than a series of discrete events independent from each other, as shown on Fig. 2.4.

2  Understanding Internal Migration: A Conceptual Framework

25

Fig. 2.4  Impact of past migration on future migration

2.6  Impacts of Migration and Urbanisation A third aspect of population movement that is expected to vary between countries and evolve over time is the role of migration in redistributing population. A primary focus of migration research in developing countries has been concerned with the urban transition, in which rural-to-urban migration has been seen as a key component. In practice, however, the specific contribution of migration to urbanisation is difficult to measure. Moreover, as Fielding (2015) makes it clear, population mobility within countries involves much more than a simple transition from rural to urban areas. A more nuanced approach, linked to a more refined theoretical framework, is needed. We adopt the theoretical model proposed by Rees et al. (2017) that anticipates links between migration and the population density of regions within countries as depicted in Fig. 2.5, suggesting that the impact of migration on population distribution follows a logistic curve with the development process. The individual graphs embedded in the larger graph plot the net migration rate against the logarithm of population density for all regions of a country, with the solid line indicating the hypothesised relationship across regions, captured empirically via linear regression. Population density is used as a surrogate for the level of urbanisation within individual regions. Positive slopes indicate that more densely populated regions are gaining population through migration, while less densely populated areas are experiencing net population losses. The steeper the slope, the greater the rate of redistribution. The logistic curve on the larger graph traces the shift from low to high levels of urbanisation (the y-axis) as development proceeds through a series of phases (the x-axis). The conceptual model indicates that migration from low to high density regions, proceeds in a progressive sequence indicated by the changing steepness of the slope. From Stages 1 to 2 migration accelerates as development takes off to reach a peak. Migration then slows down at later stages of development (Stage 3)

26

A. Bernard et al.

Fig. 2.5  Linking development to population redistribution through net migration. (Source: Rees et al. 2017)

when countries become predominantly urban. In Stage 4 and beyond, migration flows become more closely balanced with net flows potentially oscillating between net gains or losses in more urban areas, the latter corresponding to the classic process of counter-urbanisation. In practice, of course, the relationship between population density and the rate of net migration is not as clear cut as the model suggests. In most countries residuals from regression analysis indicate that more complex patterns of movement are also under way. Migration is by no means directed uniformly towards the most densely populated regions as shown by outlying regions. Nevertheless, Rees et al. (2017) found empirical support for the model across a global sample of 67 countries. It therefore provides a useful framework for comparing the spatial impact of migration in the countries examined in this book. At the level of individual countries, it also offers a useful tool to identify regions in which migration is being driven by factors other than urban expansion. Of course, this redistribution of population is the direct spatial manifestation of the very same forces that underpin the intensity of migration and shape of its age profile, as elaborated above. At a functional level, the intensity of migration, along with its efficiency, are also the key factors that determine the extent of population redistribution resulting from migration, while its age profile is at once symptomatic of the forces that attract migrants, and a key process shaping population composition at both origins and destinations. The three dimensions of migration considered in this book are therefore mutually interdependent, not only in terms of their underlying drivers, but also in their spatial consequences.

2  Understanding Internal Migration: A Conceptual Framework

27

2.7  Conclusions Simultaneous analysis of multiple dimensions of population movement across a large sample of Asian countries should reveal the way in which different facets of migration are systematically related to form broad sequences, while also providing refined insights into the particular mix of forces that shape the evolution of particular aspects of migration in each national context. One of the key strengths of comparative analysis lies in teasing out commonalities while also highlighting what is unique, or unusual, in each nation’s experience. This approach should in turn provide greater insights into the nexus between migration and development, as first sketched by Zelinsky (1971) and refined by Skeldon (1997, 2012) and thereby serve to enhance more general understanding of migration in the Asian context. It should also offer new insights into migration patterns and processes within individual countries by revealing key differences in the intensity, age selectivity and impact of migration in the context of space-time dynamics specific to particular countries. The use of a tripartite framework in which different dimensions of migration are linked to particular aspects of development allows different dimensions to evolve in a separate manner rather than in tandem. As explained by Skeldon ‘the objective is to link sequences of change in migration with other selected variables across space and time in an integrated system of migration and development’ (2012, p.  164). We believe that the conceptual approaches formulated in this chapter will enable such an endeavour.

References Bell, M., & Muhidin, S. (2009). Cross-national comparisons of internal migration (Human Development Research Paper 2009/30). New York: United Nations. Bell, M., Blake, M., Boyle, P., Duke-Williams, O., Rees, P., Stillwell, J., & Hugo, G. (2002). Cross-national comparison of internal migration: Issues and measures. Journal of the Royal Statistical Society A, 165(3), 435–464. https://doi.org/10.1111/1467-985X.t01-1-00247. Bell, M., Charles-Edwards, E., Ueffing, P., Stillwell, J., Kupiszewski, M., & Kupiszewska, D. (2015). Internal migration and development: Comparing migration intensities around the world. Population and Development Review, 41(1), 33–58 https://doi-org.ezproxy.library. uq.edu.au/10.1111/j.1728-4457.2015.00025.x. Bell, M., Charles-Edwards, E., Bernard, A. & Ueffing, P. (2018). Global trends in internal migration. In T. Champion, T.J. ,Cooke and I. Shuttleworth (Eds.), Internal migration in the developed world: Are we becoming less mobile? (pp. 76–97). London: Routledge. Bernard, A. (2017a). Levels and patterns of internal migration in Europe: A cohort perspective. Population Studies, 71(3), 293–311. Bernard, A. (2017b). Cohort measures of internal migration: Understanding long-term trends. Demography, 54(6), 2201–2221. Bernard, A., & Bell, M. (2018). Educational selectivity of internal migrants: A global assessment. Demographic Research, 39(29), 835–854. Bernard, A., & Vidal, S. (2019). Does moving in childhood and adolescence affect residential mobility in adulthood? An analysis of long-term individual residential trajectories in 11 European countries. Population, Space and Place, 26. https://doi.org/10.1002/psp.2286.

28

A. Bernard et al.

Bernard, A., Bell, M., & Charles-Edwards, E. (2014a). Improved measures for the cross-national comparison of age profiles of internal migration. Population Studies, 68(2), 179–195. Bernard, A., Bell, M., & Charles-Edwards, E. (2014b). Life-course transitions and the age profile of internal migration. Population and Development Review, 40(2), 231–239. Bernard, A., Bell, M., & Charles-Edwards, E. (2016). Internal migration age patterns and the transition to adulthood: Australia and Great Britain compared. Journal of Population Research, 33(2), 123–146. Bernard, A., Rowe, F., Bell, M., Ueffing, P., & Charles-Edwards, E. (2017). Comparing internal migration across the countries of Latin America: A multidimensional approach. PLoS One, 12(3), e0173895. Billari, F. C. (2015). Integrating macro-and micro-level approaches in the explanation of population change. Population Studies, 69(sup1), S11–S20. Black, R., Bennett, S. R., Thomas, S. M., & Beddington, J. R. (2011). Climate change: Migration as adaptation. Nature, 478(7370), 447–449. Bocquier, P., & Costa, R. (2015). Which transition comes first? Urban and demographic transitions in Belgium and Sweden. Demographic Research, 33, 1297–1332. Bongaarts, J. (1978). A framework for analyzing the proximate determinants of fertility. Population and Development Review, 4(1), 105–132. Boyle, P., Halfacree, K., & Robinson, V. (1998). Exploring contemporary migration. London: Longman. Castles, S. (2010). Understanding global migration: A social transformation perspective. Journal of Ethnic and Migration Studies, 36(10), 1565–1586. Castles, S., & Miller, M. J. (1993). The age of migration: International population movements in the modern world. London: Macmillan. Castles, S., De Haas, H., & Miller, M. J. (2014). The age of migration: International population movements in the modern world (5th ed.). London: Palgrave Macmillan. Champion, T., Cooke, T., & Shuttleworth, I. (2018). Internal migration in the developed world: Are we becoming less mobile? London: Routledge. Charles-Edwards, E., Bell, M., Bernard, A., & Zhu, Y. (2019). Internal migration in the countries of Asia: Levels, ages and spatial impacts. Asian Population Studies, 15(2), 150–171. Clark, W. A. V., & Lisowski, W. (2018). Examining the life course sequence of intending to move and moving. Population, Space and Place, 24(3), e2100. Cooke, T. J. (2013). Internal migration in decline. The Professional Geographer, 65(4), 664–675. Cooke, T. J., & Shuttleworth, I. (2017). The effects of information and communication technologies on residential mobility and migration. Population, Space and Place, 24(3), e2111. Cooke, T. J., Wright, R., & Ellis, M. (2018). A prospective on Zelinsky's hypothesis of the mobility transition. Geographical Review, 108(4), 503–522. De Haas, H. (2005). Morocco’s migration transition: Trends, determinants and future scenarios. MDR Working Paper No. 3. Centre for International Development Issues Nijmegen. De Haas, H. (2010). Migration and development: A theoretical perspective. International Migration Review, 44(1), 227–264. De Jong, G. F. (2000). Expectations, gender, and norms in migration decision-making. Population Studies, 54(3), 307–319. Faist, T. (2010). The crucial meso level. In M. Martiniello & J. Rath (Eds.), Selected studies in international migration and immigrant incoporation (pp.  59–90). Amsterdam: Amsterdan University Press. Fielding, T. (2015). Asian migrations: Social and geographical mobilities in Southeast, East, and Northeast Asia. Routledge Foster, T. B. (2017). Decomposing American immobility: Compositional and rate components of interstate, intrastate, and intracounty migration and mobility decline. Demographic Research, 37, 1515–1548. Fotheringham, A. S., Brunsdon, C., & Charlton, M. (2000). Quantitative geography: Perspectives on spatial data analysis. London: Sage.

2  Understanding Internal Migration: A Conceptual Framework

29

Goodchild, M. F., & Janelle, D. G. (2010). Toward critical spatial thinking in the social sciences and humanities. GeoJournal, 75(1), 3–13. Greenwood, M. J. (2014). Migration and economic growth in the United States: National, regional and metropolitan perspectives. New York: Academic. Greenwood, M.  J. (2019). The migration legacy of E.G.  Ravenstein. Migration Studies, 7(2), 269–278. Ishikawa, Y. (2001). Migration turnarounds and schedule changes in Japan, Sweden and Canada. Review of Urban & Regional Development Studies, 13(1), 20–33. Jones, R. C., & Brown, L. (1985). Cross-national tests of a third world development-migration paradigm: With particular attention to Venezuela. Socio-Economic Planning Sciences, 19(5), 357–361. Kawabe, H. (1990). Migration rates by age group and migration patterns: Application of Rogers’ migration schedule model to Japan, the Republic of Korea, and Thailand. Tokyo: Institute of Developing Economies. King, R. (2012). Geography and migration studies: Retrospect and prospect. Population, Space and Place, 18(2), 134–153. Klabunde, A., & Willekens, F. (2016). Decision-making in agent-based models of migration: State of the art and challenges. European Journal of Population, 32(1), 73–97. Kulu, H., & Milewski, N. (2007). Family change and migration in the life course: An introduction. Demographic Research, 17, 567–590. Massey, D. S., Joaquin, A., Graeme, H., Ali, K., Adela, P., & J. Edward, T. (1993). Theories of international migration: A review and appraisal. Population and Development Review, 431–466. Mulder, C. H. (1993). Migration dynamics: A life course approach. Amsterdam: Thesis Publisher. Mulder, C. H. (2018). Putting family Centre stage: Ties to nonresident family, internal migration, and immobility. Demographic Research, 39, 1151–1180. Mulder, C. H., Clark, W. A. V., & Wagner, M. (2002). A comparative analysis of leaving home in the United States, the Netherlands and West Germany. Demographic Research, 7, 565–592. Openshaw, S. (1984). The modifiable areal unit problem. (Geo Abstracts University of East Anglia). Palloni, A., Massey, D. S., Ceballos, M., Espinosa, K., & Spittel, M. (2001). Social capital and international migration: A test using information on family networks. American Journal of Sociology, 106(5), 1262–1298. Partridge, M. D., Rickman, D. S., Olfert, M. R., & Ali, K. (2012). Dwindling US internal migration: Evidence of spatial equilibrium or structural shifts in  local labor markets? Regional Science and Urban Economics, 42(1–2), 375–388. Pelikh, A., & Kulu, H. (2018). Short-and long-distance moves of young adults during the transition to adulthood in Britain. Population, Space and Place, 24(5). Portes, A., & DeWind, J. (2004). A cross-atlantic dialogue: The progress of research and theory in the study of international migration. International Migration Review, 38(3), 828–851. Ravenstein, E.  G. (1885). The laws of migration. Journal of the Statistical Society of London, 48(2), 167–235. Ravenstein, E. G. (1889). The laws of migration. Journal of the Royal Statistical Society, 52(2), 241–305. Raymer, J., Willekens, F., & Rogers, A. (2018). Spatial demography: A unifying core and agenda for further research. Population, Space and Place, 25(4), e2179. Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), 1–22. https://doi.org/10.1002/ psp.2036. Robinson, W.  S. (1950). Ecological correlations and the behavior of individuals. American Sociological Review, 15, 351–357. Rogers, A., & Castro, L.  J. (1981). Model migration schedules (Research Report RR-81-30). Laxenburg: International Institute for Applied Systems Analysis.

30

A. Bernard et al.

Rowe, F., Bernard, A., Charles-Edwards, E., & Bell, M. (2019). Impact of internal migration on population redistribution in Europe: Re-urbanisation, counter-urbanisation or spatial equilibrium? Comparative Population Studies, 44, 201–234. https://doi.org/10.12765/CPoS-2019-18en. Sander, N., & Bell, M. (2014). Migration and retirement in the life-course: An event-history approach. Journal of Population Research, 31(1), 1–27. Skeldon, R. (1997). Migration and development: A global interpretation. London: Longman. Skeldon, R. (2012). Migration transitions revisited: Their continued relevance for the development of migration theory. Population, Space and Place, 18(2), 154–166. Skeldon, R. (2018). A classic re-examined: Zelinsky’s hypothesis of the mobility transition. Migration Studies, 7(3), 394–403. https://doi.org/10.1093/migration/mny019. Stillwell, J., Bell, M., Blake, M., Duke-Williams, O., & Rees, P. (2000). A comparison of net migration flows and migration effectiveness in Australia and the United Kingdom. Part 1: Total migration patterns. Journal of Population Research, 17(1), 17–38. Stillwell, J., Bell, M., Blake, M., Duke-Williams, O., & Rees, P. (2001). Net migration and migration effectiveness: a comparison between Australia and the United Kingdom, 1976–1996. Part 2: Age-related migration patterns. Journal of Population Research, 18(1), 19–39. Thomas, D. S. (1938). Research memorandum on migration differentials. New York: New York Social Science Research Council. Tian, F. (2016). Transition to adulthood in China in 1982−2005: A structural view. Demographic Research, 34, 451–466. Vidal, S., & Lutz, K. (2018). Internal migration over young adult life courses: Continuities and changes across cohorts in West Germany. Advances in Life Course Research, 36, 45–56. Vidal, S., Perales, F., Lersch, P. M., & Brandén, M. (2017). Family migration in a cross-national perspective: The importance of institutional and cultural context. Demographic Research, 36, 307–338. Voss, P.  R. (2007). Demography as a spatial social science. Population Research and Policy Review, 26(5–6), 457–476. White, M.  J. (Ed.). (2016). International handbook of migration and population distribution. Dordrecht: Springer. Woods, R., Cadwallader, M., & Zelinsky, W. (1993). Classics in human geography revisited. Progress in Human Geography, 17, 213–219. Wright, R., & Ellis, M. (2016). Perspectives on migration theory: Geography. In M. White (Ed.), International handbook of migration and population distribution (pp.  11–30). Dordrecht: Springer. Xenos, P., Achmad, S., Sheng Lin, H., Keung Luis, P., Podhista, C., Raymundo, C., & Thapa, S. (2006). Delayed Asian transitions to adulthood: A perspective from national youth surveys. Asian Population Studies, 2(2), 149–185. Yeung, W. J., & Alipio, C. (2013). Transitioning to adulthood in Asia school, work, and family life. The Annals of the American Academy of Political and Social Science, 646(1), 6–27. Zelinsky, W. (1971). The hypothesis of the mobility transition. Geographical Review, 61(2), 219–249. Zelinsky, W. (1979). The demographic transition: Changing patterns of migration. In P. A. Morisson (Ed.), Population science in the service of mankind (pp. 165–168). Liege: IUSSP. Zelinsky, W. (1983). The impasse in migration theory: A sketch map for potential escapees. In P. A. Morisson (Ed.), Population movements (pp. 19–46). Liege: Ordina Editions.

Chapter 3

Comparative Measures of Internal Migration Martin Bell, Aude Bernard, Elin Charles-Edwards, and Wenqian Ke

3.1  Introduction Rigorous comparison of internal migration between countries requires a suite of robust statistical measures that simultaneously capture the multiple dimensions of human population mobility and bridge the manifold differences in the way migration data are collected. Early work provided some valuable insights into cross-­ national differences in internal migration. Long (1991) drew together census-based data on all changes of residence, demonstrating the wide differences in movement propensities among the 16 countries collecting this form of data. Rogers and Castro (1981) underlined the apparently universal shape of the age profile of migration. The role of migration in shaping patterns of human settlement was also widely examined (see e.g. Fielding 1982; Champion 1989; Rees and Kupiszewski 1999). Notwithstanding its valuable insights, two key factors constrained this work: the lack of any internationally accepted statistical indicators of migration, and the absence of readily available, comparable sources of data. As a result, most comparisons were confined to small groups of countries, chiefly in developed-country settings, primarily in Europe. The IMAGE Project (Internal Migration Around the GlobE), an international collaborative research programme with core funding from the Australian Research M. Bell (*) · A. Bernard · E. Charles-Edwards Asian Demographic Research Institute, Shanghai University, Shanghai, China Queensland Centre for Population Research, The University of Queensland, Brisbane, QLD, Australia e-mail: [email protected]; [email protected]; [email protected] W. Ke School of Geography, Fujian Normal University, Fuzhou, Fujian Province, China Asian Demographic Research Institute, Shanghai University, Shanghai, China e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_3

31

32

M. Bell et al.

Council’s Discovery Programme (2010–2015), tackled these underlying problems directly. The foundation for that project was laid more than a decade earlier in a joint UK-Australia study which argued that four discrete dimensions of migration could be recognised, each of which provided a distinctive perspective on population mobility (Bell et al. 2002): migration intensity, distance, connectivity, and spatial impact. For each of these, it defined a suite of statistical indicators and illustrated their application to cross-national comparisons (Bell et al. 2002), while also addressing computational issues (Rees et al. 2000a) and tackling the thorny problems of differences in age-time plans (Bell and Rees 2006) and in spatial structure (Blake et al. 2000; Bell et al. 1999; Rees et al. 2000b). The IMAGE project extended the comparative framework by establishing the first global inventory of internal migration data collections (Bell et al. 2014a, 2015a) and building a central repository of migration and digital boundary data for more than 100 countries (Bell et al. 2014b). It also made further advances in calculating robust comparative measures of migration intensity (Courgeau et al. 2012), age composition (Bernard et al. 2014a, Bernard and Bell 2015) and spatial impact (Rees et al. 2017), and it generated the first global league tables comparing countries on three of the four dimensions of migration identified earlier: intensity (Bell et  al. 2015b), impact (Rees et  al. 2017) and distance (Stillwell et  al. 2016). Central to these advances was development of the IMAGE Studio (Stillwell et al. 2014), a suite of software specifically designed for the analysis and modelling of internal migration. As well as providing a bespoke platform for computing the full range of migration indicators, the IMAGE Studio incorporated a unique facility for random aggregation of basic spatial units, generating flexible geographies that addressed the Modifiable Areal Unit Problem (MAUP), a longstanding issue that had hitherto undermined analysis of spatial data (Stillwell et al. 2018). The migration indicators developed during the IMAGE project, and the analytical techniques to compute them, provide the essential metrics for the statistics reported in this book. In addition to the global summaries mentioned above, concise comparisons of internal migration using IMAGE metrics have been published for three world regions: Latin America (Bernard et al. 2017), Europe (Rowe et al. 2019) and Asia (Charles-Edwards et al. 2019). As explained in Chapter One, this volume aims to offer a more comprehensive understanding by blending statistical measures with qualitative discussion that conveys the country-specific contexts within which internal migration in the countries of Asia occurs. Of all the world’s regions Asia contains perhaps the greatest diversity of geographic settings, social and cultural norms, political and administrative systems, and levels of economic development. Reflecting this diversity, it also presents a challenging array of arrangements in the way that internal migration is conceived and measured, and the way in which data are collected and made available. Understanding how these differences affect the analysis and interpretation of migration data is essential for robust cross-national comparisons. This chapter first reviews four groups of factors that hinder comparisons of internal migration. It then describes the methods used to transcend these issues and generate internationally comparable estimates of the three dimensions of migration that are central to analysis in subsequent chapters of this volume: migration intensity, age at migration, and spatial impact.

3  Comparative Measures of Internal Migration

33

3.2  Impediments to Cross-national Comparisons Four main impediments to cross-national comparisons of migration can be identified, deriving from (1) the way migration is measured, (2) the length of the observation interval, (3) the number and spatial arrangement of geographic units into which countries are divided, and (4) issues of temporal comparability, population coverage and data quality (Bell et al. 2002). We consider each of these in turn. Migration can be measured in several ways, but the two most common forms are as events or transitions. Event data capture all occurrences of relocation and are commonly associated with population registers that record individual changes of address. Transitions measure migration by comparing place of residence at two discrete points in time and are the most common form of data captured in population censuses. Over short intervals such as a single year, transitions closely correspond to the number of events, but as the observation interval lengthens transition data show a gradually slowing rate of increase because repeat and return migrations are missed, whereas these are routinely captured by collections that focus on migration events (Long and Boertlein 1990). As a result, care is needed in comparing statistics on any of the dimensions of migration where data are drawn from different sources, especially when migration is measured over a lengthy interval. Transition and event data also follow different agetime plans, so particular care is needed when comparing migration age profiles, even when measured over a single-year interval (Bell and Rees 2006). Forty-one of the 47 UN member states in Asia collect migration data using a population census while just 16 draw data from a population register. Only four — Brunei Darussalam, Kuwait, the United Arab Emirates and Lebanon — appear not to have collected information on internal migration since the mid-1990s. More than half of all Asian countries also collect migration data using regular or occasional surveys. Prior work has shown that USAID’s Demographic and Health Surveys (DHS) provide poor information on internal migration (Charles-Edwards et al. 2019) and few surveys provide reliable data on spatial patterns. Surveys can, however, provide valuable insights into the motives for migration, as demonstrated in several of the chapters which follow. A second impediment to cross-national comparisons arises from differences in the interval over which migration is measured. Many countries use a fixed interval of 1 or 5 years to capture transitions in place of residence, but other intervals are not uncommon, and some censuses simply ask for place of previous residence, irrespective of when the last migration took place. Although simple multipliers are sometimes used, there is no reliable analytical solution to comparing migration data measured over 1 year compared with that measured over 5 years (Kitsul and Philipov 1981; Rogerson 1990). This is partly because of the effects of return and repeat migration mentioned above, but also because conditions change from year to year, so no individual year is truly representative. Only at the level of net migration for individual spatial units can simple multipliers be reliably employed (Bell et  al. 2002). As shown in Table 3.1, five-year data are more common in Asian countries, and a large proportion also collect data on previous place of residence irrespective of the date of the move (labelled ‘Last Move’ in Table 3.1). Such data are difficult to use in isolation, but fortunately most countries that employ this approach also

7

15

Y

Y

Y Y

Census 1 year 5 year Y Y

Y Y 31

Y Y Y Y Y Y Y

Lifetime Y Y

Y 22

Y Y

Y Y Y Y Y

Last move

Y 28

Y Y

Y

Y Y Y 16

y Y

Y

Duration of residence Register

27

Y Y

Y Y

Y

Survey

Note 1: The 2015 Census was based on administrative datasets, supplemented by a sample survey

South Korea Japan Cambodia Myanmar Thailand India Bhutan Nepal Sri Lanka Iran Israel Armenia Kazakhstan Total Asia

Country China Mongolia

Table 3.1  Internal migration data collection, selected countries and total Asia Name and Number of zones Prefecture (374) Aimag (21) Soum (331) Gu, Si, Gun (229) Prefecture (47) Districts (183) District (74) Tambon/Municipality (~7000) State (31) Dzongkha (20) District (77) District (25) Shahrestan (429) Locality (1211) Marz (11) District (177) – 2012 2016 2008 2011 2009

20151 2015 2008 2014 2010 2011 2017

Date of latest Census 2010 2010

34 M. Bell et al.

3  Comparative Measures of Internal Migration

35

collect data on duration of residence. If the question on duration of residence refers to the same spatial units that are used to cross-tabulate place of current and place of previous residence, then duration can be used as a filter to deliver a matrix of origin-­ destination flows that closely matches that which would be obtained from a transition-­based migration question (Bell et al. 2002). Thus, it is possible to approximate a five-year transition matrix for most Asian countries, including those considered in the current volume. Fixed interval measures provide a contemporary picture of migration patterns and a sequence from successive censuses reveals temporal trends. Five-year intervals also have the benefit of smoothing out annual fluctuations due to economic or political shocks that may influence data for a single year. Globally, however, the most common form of migration data measures lifetime migration by comparing place of residence at the time of data collection with place of birth. Lifetime migration data are less readily compared over time, or between countries, and can be difficult to interpret because they obscure the multiple changes of residence that may have occurred over a person’s lifetime. Moreover, such data are usually available only at a relatively coarse spatial scale, and they may be distorted by the boundary changes that commonly occur over a long time period. Lifetime migration data do, however, provide valuable insights into the net redistribution of the surviving population in a country over an extended period, and can be used in association with contemporary transition data to reveal changes in spatial patterns of movement, as shown by Charles-Edwards et al. (2019). The third and perhaps most challenging impediment to cross-national comparisons arises from differences between countries in the spatial framework or zonal system, that is in the number and arrangement of spatial units into which each territory is divided. This is generally referred to as the modifiable areal unit problem (MAUP), in recognition of the fact that changes (modifications) to the spatial framework influence the results from any form of spatial analysis (Openshaw 1984). Migration is generally defined by reference to relocation across the boundaries of administrative units, such as provinces or municipalities. However, these vary widely in number from country to country. Among the 15 countries discussed in this volume, for example, the number of zones for which migration data are collected ranges from just 11 in Armenia to over 7000 in Thailand (Table 3.1), although in practice data are rarely made available at such fine levels of spatial scale (see below). These differences exert a fundamental effect on the measurement of migration, particularly the apparent intensity, or overall level of movement, since a greater number of migrations are captured in a more finely grained zonal system. For two hypothetical countries which are of the same geographic size and divided into the same number of zones, the shape of the zones and their relation to the settlement pattern will also affect the apparent number of cross-border migrations. In a similar way the spatial framework affects the apparent pattern of movement through the settlement system, and it influences the resulting metrics that capture the extent and nature of population redistribution through migration. The IMAGE project devised and implemented measures that overcome these issues arising from differences in zonal systems, thus addressing the MAUP. We describe these solutions later in this chapter. A final set of issues that undermine comparability between countries arise from differences in the historical juncture at which migration is measured, the nature of

36

M. Bell et al.

the information released, and the quality of the data. UN guidelines recommend that censuses should be taken at least every 10 years, around the middle or end of the decade (United Nations 2008), but not all countries follow this recommendation. As shown in Table 3.1, the most recent censuses in our sample of 15 Asian countries were closely aligned in time, but countries differ in the time taken to process and release data, not all information collected is necessarily made available, and changes may also occur in the type of information collected. For example, Iran shifted from a mid-decade census in 1996 and in 2006 to a five-yearly enumeration in 2011 and 2016, and simultaneously switched from a ten-year to a five-year observation interval over which to measure migration. Even when censuses are coordinated, economic cycles may be out of phase, thereby distorting relativities between countries in migration intensity or spatial patterns. Population coverage and data quality also vary between countries and over time (Bell et al. 2002). Data availability is particularly challenging. Inspection of the types of information collected at a census or via a population register may suggest that certain data tabulations would be possible, providing valuable insights into particular aspects of migration. In practice, however, the multi-dimensionality which is inherent to migration data hinders hard copy output and only a few countries make detailed tabulations readily available in electronic form. In this respect, the IPUMS initiative at the University of Minnesota, which holds census sample files from a large and growing number of countries, has provided an invaluable resource for migration researchers (Minnesota Population Centre 2019). The analyses presented in many of the chapters in this volume draw on IPUMS data to supplement tables and reports made available by national statistical agencies. The chapters also draw on the extensive global repository of migration data established as part of the IMAGE project (Bell et al. 2014b).

3.3  Measuring Migration Intensity While other measures have been devised, the overall level of population mobility is conventionally measured by the Crude Migration Intensity (CMI) computed as

CMI = M / P ∗ 100

(3.1)

where M represents the total number of migrants or migrations, and P represents the population at risk. The term migration intensity owes its origins to van Imhoff and Keilman (1991) and is a convenient rubric encompassing both transition probabilities (computed from censuses) and movement rates (from population registers). Computational processes and associated issues are laid out in Rees et  al. (2000a). Intensities can be calculated for any level of spatial scale for which the requisite data are available, e.g., provinces, prefectures or municipalities, but because of differences in spatial scale such estimates cannot be compared between countries. The solution adopted in the IMAGE project was to make such comparisons using the Aggregate CMI (ACMI), a measure that captures all changes of residence within a country, not

3  Comparative Measures of Internal Migration

37

simply those that cross zonal boundaries (Bell et  al. 2015b). This is similar to the approach adopted by Long (1991) in his early work on cross-national comparisons, but the IMAGE measure excludes international migrants to avoid distortions arising from differences due to this source of movement, and generates estimates of the ACMI for the many countries which do not collect this information directly. Information to measure all residential moves is sometimes derived from a discrete census question that asks whether the respondent is still living at the same address as 1 or 5 years previously, but this is the case for only three of the countries included in the current volume: Japan, South Korea and Israel. In the IMAGE Project estimates of the ACMI for other countries were therefore made using the approach developed by Courgeau et al. (2012), and these have been replicated for the current volume, using more recent data where available. To ensure consistency of approach, these estimates were computed centrally and provided to authors for inclusion in their individual country chapters. The approach devised by Courgeau et al. (2012) estimates the ACMI by leveraging the log-linear relationship that is found between Crude Migration Intensity and the Average Number of Households per Zone, computed across a series of geographic levels. The underpinning logic is that, as the number of zones into which a territory is divided increases, so does the number of inter-zonal migrants. Courgeau et al. (2012) set out the underlying theoretical rationale and explain its derivation. Bell et al. (2015b) provide an extended discussion of its implementation and provide further empirical evidence of the linear association. Formally, the relationship can be expressed as

CMI n = w + k ln ( H / n )



(3.2)

where CMI represents the Crude Migration Intensity observed for a system of n zones, H denotes the total number of households summed across all zones, k indicates the slope of the regression, and w is a constant corresponding to the y-­intercept. Note that the equation uses households rather than persons, since a household is taken to be equivalent to an occupied dwelling, and therefore represents the smallest physical unit to or from which an individual can migrate. For countries that provide migration data at several levels of spatial scale (e.g., states, provinces, counties), it is therefore possible to estimate Eq. [3.2]. Substituting H/n = 1 in the estimated equation then corresponds to a hypothetical level of spatial resolution at which there is a single household per zone and therefore captures all migrations. Since ln(1)  =  0, the corresponding ACMI can be computed directly as the constant w or can be read as the y-intercept from a graph. Figure 3.1 illustrates the method for migration in Iran 2006–2011 using data for three levels in the administrative hierarchy, together with additional information on movement between some 63,000 towns and villages. An additional point is included on the graph for the purposes of fitting Eq. [3.2], corresponding to the situation where the country consists of a single zone, so that the CMI = 0 at ln(H). This is shown as the x-intercept in Fig. 3.1. The result delivers an estimated five-year ACMI of 11.3% with a coefficient of determination (r-squared) of 0.9977. Data at various levels of the administrative hierarchy (e.g., provinces and counties) are often made available by statistical agencies or can be readily computed, but they are typically limited to three or four spatial levels, which can lead to high

38

M. Bell et al.

Fig. 3.1  Estimating the ACMI for Iran, 2006–2011

standard errors and limited confidence in regression analysis. The IMAGE studio, described below, delivers a more robust foundation for estimating the ACMI by generating estimates of migration intensity for multiple aggregations and spatial configurations of zones at user-defined spatial levels in each country (Stillwell et al. 2014). A measure like the ACMI that captures all changes of address, irrespective of differences in spatial scale encompasses both local residential mobility within the same neighbourhood and longer distance migration, such as between cities or regions. Some commentators argue that these are functionally different processes, since residential mobility simply involves a change in housing circumstances without severing local ties and networks. However, Stillwell et al. (2016) found that a spatial interaction model delivered stable beta parameters and little variation between countries in the effect of distance friction on migration, which suggests no clear division between the two forms of movement. The work of Niedomysl and Fransson (2014) provides further evidence. Using finely gridded data for Sweden, they plotted the proportion of people who changed their place of employment with increasing distance migrated. They report no clear break until this proportion levelled off at around 100  km, which again suggests there is no empirical division between migration and residential mobility. While the topic merits further research, for most countries it seems there is no practical mechanism by which to distinguish the two forms of mobility, so we maintain the ACMI as the most robust measure for cross-national comparison. In later chapters individual country authors also report CMIs for movement at other levels of spatial scale.

39

3  Comparative Measures of Internal Migration

3.4  Measuring the Age Profile Migration is highly selective and one of the most pronounced and persistent differentials is associated with age. Rogers and Castro (1981) are credited with formalising these age-related regularities in the form of model migration schedules, noting that the overall shape of the age profile was maintained irrespective of migration distance. The broad shape of the profile can be characterised as peaking sharply among young adults then falling steadily through middle age, before rising around or beyond retirement. Movement intensities also fall among teenagers but rise for the youngest children who move with their parents. While confirming this general profile, more recent work has identified subtle but important variations between countries and has devised metrics that are more readily interpreted than the parameters drawn from the Rogers schedules, and that capture these differences more reliably. Bell and Muhidin (2009) used the age and intensity at peak migration, together with a measure of the breadth of the peak (the proportion of the area under the curve falling within a band 5 years either side of the peak) and graphed a global sample of age profiles that revealed significant differences between 19 countries. Subsequently Bernard et al. (2014a) tested a suite of possible metrics and demonstrated that just two indicators, age and intensity at the peak, accounted for two-­ thirds of the variance across a sample of 25 countries. As outlined in Chapter Two, they went on to show that differences in the age at peak could largely be attributed to the timing of life-course transitions (Bernard et al. 2014b). Figure 3.2 illustrates the calculation of these two metrics for selected countries. These are the indicators that are cited in the following chapters on individual Asian countries, accompanying graphs of the migration age profile, which in each case have been smoothed using the method described by Bernard and Bell (2015).

Intensity at peak, country B

COUNTRY A

MIGRATION INTENSITY

Intensity at peak, country A

COUNTRY B

Age at peak, country A

AGE

Fig. 3.2  Measuring the age and intensity at peak migration

Age at peak, country B

40

M. Bell et al.

3.5  Measuring Spatial Impacts Perhaps the most striking facet of migration lies in the way it alters the pattern of human settlement, and most visible in this context is its contribution to urbanisation. In practice it is difficult to measure the true volume of movement between rural and urban areas. While many countries identify the rural or urban status of residents at the time of the census, few also collect information on the status of movers at the start of the transition interval. As urban areas grow, their boundaries expand to encompass settlements that were previously considered rural. Simple comparisons of urban population size therefore conflagrate the effects of migration with other processes such as the reclassification of settlements as they grow or absorb territory on the fringes of cities (see e.g., United Nations 2000). Urban growth also prevents the delineation of stable zonal boundaries distinguishing urban and rural areas, so the administrative units generally used for data collection commonly encompass territory that is both urban and rural, with the proportions altering over time (Champion et  al. 2003). Coupled with the fact that countries use widely varying definitions of ‘urban’, this approach to classification provides a poor foundation for analysing the role of migration in reshaping patterns of human settlement. In the absence of reliable data on rural-to-urban migration, conventional practice has been to calculate net migration, together with the underlying inwards and outwards flows, for the administrative units against which they were collected, and to represent this as a rate on choropleth maps to depict the spatial pattern of population redistribution. Formally, the Net Migration Rate (NMR) for any region, i, is calculated as the balance of the inwards flows from all other regions j (Dji) and the outwards flows to all other regions j (Oij), divided by the population at risk (Pi) (which is commonly represented by the population in area i at the start of the period), and expressed as a percentage [3.3].



NMRi = 100 ∗ ∑ ( D ji − Oij ) / Pi j, j ≠i

(3.3)

Maps of this type are presented for each country in subsequent chapters at appropriate levels of scale and have the advantage of being easily recognised by those familiar with the field. Because population densities are uneven, however, choropleth maps can convey a misleading impression of the scale of net movements. Moreover, they are a poor framework on which to portray the direction of the underlying migration flows. Circular plots, adapted for use in migration studies by Abel and Sander (2014), overcome these deficiencies by situating regions around the circumference of a circle, each occupying an arc corresponding to their share of the aggregate volume of inter-regional migration. Broad arrows depict the distribution of outwards movements from each region and show the origin of inwards movements, with the direction of flows denoted by the width of the gap between the start (no gap) and end (gap) of the flow arrows (Fig. 3.3). Software to generate the circular plots is available in the R package Circlize (Gu 2019).

3  Comparative Measures of Internal Migration

41

Fig. 3.3  Circular plot of interregional migration flows for a hypothetical country

Circular plots underline the fact that net migration gains or losses conceal much larger two-way flows. The relationship between these flows reveals important information about the nature of population movements and this is also readily captured in a simple metric that is used in a number of the country case studies. The Migration Effectiveness Ratio (MERi) for a single region, i, is computed as the ratio of net gains or losses by the region in exchanges with all other regions j (Dji-Oij) and the sum of the inwards and outwards flows to and from all other regions j (Dji + Oij), expressed as a percentage [3.4], with limits of +/−100. Low values of the MER (e.g., 60), on the other hand, denote strongly asymmetric migration flows leading to large gains (positive ratios) or losses (negative ratios) for the given volume of migration.

42



M. Bell et al.

MERi = 100 ∗ ∑ ( D ji − Oij ) / j, j ≠i

∑ (D

ji

j, j ≠i

+ Oij )

(3.4)

The NMR and MER offer useful metrics to explore differences in the effects of migration across regions within a country but are not well suited to comparisons between countries or over time. For this, system-wide metrics are needed that capture the same process in a pair of analogous indicators. Bell et al. (2002) proposed the Migration Effectiveness Index (MEI) [3.5] and the Aggregate Net Migration Rate (ANMR) [3.8], computed as shown below. They also identified the functional relationship between the MEI, ANMR and the CMI [3.9]. Since these metrics refer to all regions within a country, for simplicity we can dispense with subscript j, and simply designate regions by reference to subscript i. Computationally, then.



MEI = 100 ∗ ∑ Di − Oi / ∑ ( Di + Oi ) i

i

(3.5)

where Di represents the total inflows to zone i and Oi represents the total outflows from zone i. Eq. [3.5] can also be written as



MEI = 100 × 0.5∑ Di − Oi / M i

(3.6)

where M denotes the sum of all flows between the system of regions such that



M = ∑ ( Di ) = ∑ ( Oi ) i

i

(3.7)

The MEI can assume values between 0 and + 100. As with the MER, high values indicate that, overall, migration is an efficient mechanism of population redistribution, generating a large net effect for the given volume of movement. Conversely, low values denote that interzonal flows are more closely balanced, leading to comparatively little redistribution. The ANMR measures the redistributive impact of migration more directly by relating the total volume of net migration gains, summed across all regions, to the population at risk, P:



ANMR = 100 × 0.5∑ Di − Oi / P i

(3.8)

Combining Eqs. 3.1, 3.6 and 3.8, it can then be seen that, at any given spatial scale, the CMI, MEI and ANMR are functionally related such that

ANMR = CMI × MEI / 100

(3.9)

3  Comparative Measures of Internal Migration

43

The significance of this relationship lies in demonstrating that the spatial impact of migration for any systems of zones (e.g., regions, municipalities) is driven by the combined effects of migration intensity and migration effectiveness. If intensity is high but effectiveness is low, most movement will be absorbed in reciprocal flows, leading to little net redistribution. Conversely, high effectiveness will exert little spatial impact if intensity is low. It is the combination of high intensity and high effectiveness that brings about the most pronounced redistribution of population. These measures are used extensively in the following chapters to examine the impact and underlying dynamics of migration at various spatial levels. It is important to recognise, however, that the CMI, MEI and ANMR always refer to specific zonal systems within each country and are not directly comparable from one country to the next. In a further development of the IMAGE Project, Rees et al. (2017) developed an additional indicator, the Index of Net Migration Impact (INMI), based on empirical links between these three measures that transcend differences in spatial scale. The INMI therefore provides the mechanism for robust comparison of migration impact in different countries, with the added advantage of being able to distinguish the relative contributions of intensity and effectiveness (see also Bernard et al. 2017; Charles-Edwards et al. 2019). We revisit the INMI in the concluding chapter of this volume where we draw comparisons across countries in the Asian region and situate them in a global context. Building on earlier work by Rees and Kupiszewski (1999), the IMAGE project developed one further tool for analysing migration impact, using population density as a surrogate for the level of urbanisation. Earlier work classified zones at selected geographic levels into density bands, usually quintiles, and computed the aggregate net migration gain or loss in each density band. Rees et  al. (2017) refined the approach by setting the net migration rate against the population density for individual zones in a population-weighted regression framework. Figure 3.4 provides an illustrative example. The slope of the regression line indicates whether migration, on balance, is relocating the population towards higher (positive slope) or lower (negative slope) density regions. Comparison of regression parameters for a sequence of temporal intervals can serve to test the conceptual model advanced by Rees et al. (2017), as outlined in Chap. 2. At the same time, close inspection of the scattergram can help identify zones that are driving the redistribution, and those that are following a different course, apparent on scatterplots as residuals to the regression.

3.6  The IMAGE Studio The IMAGE Studio was developed as a bespoke software facility to undertake two key tasks: first to provide a standardised, readily accessible framework to compute the extensive suite of migration indicators set out by Bell et al. (2002), using the forms of migration data that are commonly available from population censuses and administrative sources; and secondly to provide a facility to explore the MAUP

44

M. Bell et al.

Fig. 3.4  Net migration rates by log population density, regions of Japan, 2016–2017. (Source: Data from Annual Report on Internal Migration in Japan derived from the Basic Resident Registers, (https://www.e-stat.go.jp/en/) Analysis by IMAGE-Asia)

effects on measures of migration. While a detailed description of the IMAGE Studio is not needed here, several aspects merit comment to aid understanding of the metrics presented in the following chapters. The IMAGE Studio calculates migration indicators both for individual regions, such as the NMR and MER (local indicators), and for the country as a whole, such as the MEI and ANMR (system-wide indicators). Local indicators were used in creation of the choropleth maps and circular plots that were prepared at the Asian Demographic Research Institute specifically for the country chapters that make up Part II of this book. Age profiles and migration age indicators were also computed centrally but these indicators do not form part of the IMAGE Studio and were calculated separately. Similarly, it should be noted that the ACMI, the indicator measuring all changes of address, and the INMI, which provides corresponding measures of redistribution as presented in the concluding chapter, utilise estimates of the CMI, MEI and ANMR, which are generated by the IMAGE Studio but are themselves calculated outside the Studio using simple spreadsheets and the R package. Exploration of the MAUP using the IMAGE Studio is reported in several IMAGE papers (see e.g., Stillwell et al. 2018) and is outside the scope of this volume. The essence of this facility, however, lies in providing multiple random configurations of basic spatial units at user-defined levels of aggregation, with all of the available indicators computed at each spatial level and configuration. One particular benefit of this facility lies in providing a series of estimates of system-wide indicators such as the CMI to supplement the observations available directly for standard levels of

3  Comparative Measures of Internal Migration

45

the administrative hierarchy in each country (e.g., provinces, prefecture, municipalities). In this way it was therefore possible to substantially increase the number of observations of the CMI and therefore establish a more robust basis for estimating the ACMI using the regression approach devised by Courgeau et al. (2012). A comprehensive exposition of the IMAGE Studio is provided by Stillwell et al. (2014, 2018) and its use in various parts of the IMAGE project is elaborated in Bell et al. (2015b), Rees et al. (2017) and Stillwell et al. (2016). The software is available for download from https://github.com/IMAGE-Project free of charge and the latest version of the user manual (Daras 2014) is available at https://imageprojectcomau. files.wordpress.com/2018/11/image_studio_1_4_2_user_manual.pdf.

3.7  Conclusions Countries vary widely in the types of migration data they collect and the form in which it is released. The central aims of the IMAGE Project were to provide a common framework for the analysis of such data and to implement a standard suite of metrics to explore migration within countries and make international comparisons. This chapter has outlined the development of this process and described the logic and computation of key migration indicators. The focus has been confined to three dimensions of migration: its intensity, age composition and spatial impact. Meaningful interpretation of these measures is ultimately dependent on situating them within their contextual setting, which calls for solid background knowledge of the geography, history, political economy, society and culture within which such migrations took place. This is the task for authors of the individual country chapters that follow. The IMAGE Studio itself generates a wide range of additional measures for other dimensions of migration including movement distance and spatial connectivity, but these are conceptually and methodologically complex, and are not considered further in this volume. Considerable scope remains to use the IMAGE studio for more detailed analysis of migration in individual countries, or to draw international comparisons on these other dimensions of migration. A central point that warrants emphasis is that a large share of the metrics presented in individual country chapters in this volume were computed centrally as part of this IMAGE-Asia Project, either at the University of Queensland or at Shanghai University (ADRI), and were made available to individual chapter authors. Base datasets, comprising origin-destination matrices and the associated digital boundary files which underpin these metrics, were drawn either from the IMAGE Repository (Bell et al. 2014a), the IPUMS facility (Minnesota Population Centre 2019) or provided by individual chapter authors. Where tabular or graphical material computed using the IMAGE facility have been used, this is indicated by inclusion of the words (IMAGE-Asia Project) in parentheses in the source line of a figure or table. Use of data from the IPUMS facility is acknowledged in a similar way.

46

M. Bell et al.

References Abel, G., & Sander, N. (2014). Quantifying global international migration flows. Science, 343(6178), 1520–1522. https://doi.org/10.1126/science.1248676. Bell, M., & Muhidin, S. (2009). Cross-national comparisons of internal migration (Human Development Research Paper 2009/30). New York: United Nations http://hdr.undp.org/sites/ default/files/hdrp_2009_30.pdf. Bell, M., & Rees, P. (2006). Comparing migration in Britain and Australia: Harmonisation through use of age-time plans. Environment and Planning A, 38(5), 959–988. https://doi. org/10.1068/a35245. Bell, M., Rees, P., Blake, M., & Duke-Williams, O. (1999). An age-period-cohort data base of inter-regional migration in Australia and Great Britain 1976–96 (Working Paper 99/02). Leeds: School of Geography, The University of Leeds https://csapblog.leeds.ac.uk/ csap-working-papers/. Bell, M., Blake, M., Boyle, P., Duke-Williams, O., Rees, P., Stillwell, J., & Hugo, G. (2002). Cross-national comparison of internal migration: Issues and measures. Journal of the Royal Statistical Society A, 165(3), 435–464. https://doi.org/10.1111/1467-985X.t01-1-00247. Bell, M., Charles-Edwards, E., Kupiszewska, D., Kupiszewski, M., Stillwell, J., & Zhu, Y. (2014a). Internal migration around the world: Assessing contemporary practice. Population, Space and Place, 21(1), 1–17. https://doi.org/10.1002/psp1848. Bell, M., Bernard, A., Ueffing, P., & Charles-Edwards, E. (2014b). The IMAGE repository: A user guide (Working Paper 2014/01). Brisbane: Queensland Centre for Population Research, The University of Queensland https://imageprojectcomau.files.wordpress.com/2017/03/imagerepositoryuserguide.pdf. Bell, M., Bernard, A., Charles-Edwards, E., Kupiszewska, D., Kupiszewski, M., Stillwell, J., Zhu, Y., Ueffing, P., & Booth, A. (2015a). The IMAGE inventory: A user guide (Working Paper 2015/01). Brisbane: Queensland Centre for Population Research, University of Queensland https://imageprojectcomau.files.wordpress.com/2017/03/imageinventoryuserguide.pdf. Bell, M., Charles-Edwards, E., Ueffing, P., Stillwell, J., Kupiszewski, M., & Kupiszewska, D. (2015b). Internal migration and development: Comparing migration intensities around the world. Population and Development Review, 41(1), 33–58 https://doi-org.ezproxy.library. uq.edu.au/10.1111/j.1728-4457.2015.00025.x. Bernard, A., & Bell, M. (2015). Smoothing internal migration age profiles for comparative research. Demographic Research, 32(33), 915–948. https://doi.org/10.4054/DemRes.2015.32.33. Bernard, A., Bell, M., & Charles-Edwards, E. (2014a). Improved measures for the cross-national comparison of age profiles of internal migration. Population Studies, 68(2), 179–195. https:// doi.org/10.1080/00324728.2014.890243. Bernard, A., Bell, M., & Charles-Edwards, E. (2014b). Life-course transitions and the age profile of internal migration. Population and Development Review, 40(2), 213–239 https://doi-org. ezproxy.library.uq.edu.au/10.1111/j.1728-4457.2014.00671.x. Bernard, A., Rowe, F., Bell, M., Ueffing, P., & Charles-Edwards, E. (2017). Comparing internal migration across the countries of Latin America: A multidimensional approach. PLoS One, 12(3), e0173895. https://doi.org/10.1371/journal.pone.0173895. Blake, M., Bell, M., & Rees, P. (2000). Creating a temporally consistent spatial framework for the analysis of interregional migration in Australia. International Journal of Population Geography, 6(2), 155–174 10.1002/(sici)1099-1220(200003/04)6:23.0.co;2-a. Champion, A.  G. (1989). Counterurbanisation: The changing pace and nature of population deconcentration. London: Edward Arnold. Champion, A., Hugo, G., & Lattes, A. (2003). Beyond the rural/urban dichotomy. Population and Development Review, 29(2), 277–298. Charles-Edwards, E., Bell, M., Bernard, A., & Zhu, Y. (2019). Internal migration in the countries of Asia: Levels, ages and spatial impacts. Asian Population Studies, 15(2), 150–117. https://doi. org/10.1080/17441730.2019.1619256.

3  Comparative Measures of Internal Migration

47

Courgeau, D., Muhidin, S. & Bell, M. (2012). Estimating changes of residence for cross-national comparison. Population-E, 67(4), 631–652. https://doi.org/10.3917/pope.1204.0631. Also published as estimer les changements de résidence pour permettre les comparaisons internationales. Population-F, 67(4), 747–770. https://doi.org/10317/popu.1204.0747. Daras, K. (2014). IMAGE studio 1.4.2 user manual. School of Geography, University of Leeds, UK. Available from the IMAGE Project website at https://imageprojectcomau.files.wordpress. com/2018/11/image_studio_1_4_2_user_manual.pdf. Fielding, A. J. (1982). Counter-urbanisation in Western Europe. Progress in Planning, 17, 1–52. Gu, Z. (2019). Circlize: Circular visualisation. R package, version 0.4.8 https://cran.r-project.org/ web/packages/circlize/. Kitsul, P., & Philipov, D. (1981). The one-year/five-year migration problem. In A. Rogers (Ed.), Advances in multiregional demography (pp.  1–34). Laxenberg: International Institute for Applied Systems Analysis. Long, L. (1991). Residential mobility differences among developed countries. International Regional Science Review, 14, 133–147. Long, J., & Boertlein, C. (1990). Comparing migration measures having different intervals (Current Population Reports, Series P-23) (pp. 1–11). Washington DC: US Bureau of the Census. Minnesota Population Center. (2019). Integrated public use microdata series, international: version 7.2 [dataset]. Minneapolis: IPUMS, 2019. https://doi.org/10.18128/D020.V7.2. Niedomysl, T., & Fransson, U. (2014). On distance and the spatial dimension in the definition of internal migration. Annals of the Association of American Geographers, 104(2), 357–372. Openshaw, S. (1984). The modifiable areal unit problem. Norwich: Geo Abstracts University of East Anglia. Rees, P., & Kupiszewski, M. (1999). Internal migration and regional population dynamics in Europe: A synthesis. Strasbourg: Council of Europe Publishing. Rees, P., Bell, M., Duke-Williams, O., & Blake, M. (2000a). Problems and solutions in the measurement of migration intensities: Australia and Britain compared. Population Studies, 54(2), 207–222. https://doi.org/10.1080/713779082. Rees, P., Bell, M., Duke-Williams, O., & Blake, M. (2000b). Harmonising databases for the cross-­ national study of internal migration: Lessons from Australia and Britain. In Working Paper 00/05. Leeds: School of Geography, The University of Leeds https://csapblog.leeds.ac.uk/ csap-working-papers/. Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), 1–22. https://doi.org/10.1002/ psp.2036. Rogers, A., & Castro, L.  J. (1981). Model migration schedules (Research Report RR-81-30). Laxenburg: International Institute for Applied Systems Analysis. Rogerson, P.  A. (1990). Migration analysis using data with time intervals of differing widths. Papers of the Regional Science Association, 68, 97–106. Rowe, F., Bell, M., Bernard, A., Charles-Edwards, E., & Ueffing, P. (2019). Impact of internal migration on population redistribution in Europe. Comparative Population Studies, 44, 201–234. https://doi.org/10.12765/CPoS-2019-18en. Stillwell, J., Daras, K., Bell, M., & Lomax, N. (2014). The IMAGE studio: A tool for internal migration analysis and modelling. Applied Spatial Analysis and Policy, 7(1), 5–23. https://doi. org/10.1007/s12061-014-9104-4. Stillwell, J., Bell, M., Ueffing, P., Daras, K., Charles-Edwards, E., Kupiszewski, M., & Kupiszewska, D. (2016). Internal migration around the world: Comparing distance travelled and its frictional effect. Environment and Planning, A., 48(8), 1657–1675. https://doi.org/1 0.1177/0308518X16643963. Stillwell, J., Daras, K., & Bell, M. (2018). Spatial aggregation methods for investigating the MAUP effects in migration analysis. Applied Spatial Analysis and Policy, 11, 693–711.

48

M. Bell et al.

United Nations. (2000). World population monitoring 1999: Population growth, structure and distribution (Population Division, ST/ESA/SER A/183, Department of Economic and Social Affairs). New York: UN. https://doi.org/10.1007/s12061-018-9274-6. United Nations. (2008). Principles and recommendations for population and housing censuses. New York: UN. van Imhoff, E., & Keilman, N. (1991). LIPRO 2.0: an application of a dynamic demographic projection model to household structure in the Netherlands. Amsterdam: Swets and Zeitlin.

Part II

The Evidence

Part II contains 15 chapters, each focussing on internal migration in an individual Asian country. These extend from Japan in the east to Israel in the west, and are arranged broadly in east-west order, starting with China, Mongolia, South Korea and Japan (East Asia), moving through Cambodia, Myanmar and Thailand (Southeast Asia) to India, Bhutan, Nepal, Sri Lanka and Iran (South Asia) followed by Armenia, Israel and Kazakhstan (Central and West Asia). Each chapter is organised according to the same structure: setting out the type of migration data available and the spatial framework used to collect it, summarising prior research, then analysing empirical evidence as to the intensity of internal migration, the age composition of migrants, and the spatial impact of migration in shaping the pattern of human settlement. Each concludes by exploring the implications of migration for that country and identifying aspects that are key to understanding its migration dynamics.

Chapter 4

Internal Migration in China Jianfa Shen

4.1  Introduction The population of China grew from 0.54 billion in 1949 to 0.96 billion in 1978 and 1.38 billion in 2016 (NBS 2017). To contain the growth rate, China began to implement nationwide family planning in 1973. The most rigorous one-child policy was adopted from 1979 to 2013. As a result, the annual population increase was reduced to 8.09 million in 2016, which corresponds to an annual growth rate of 0.59%. A two-child policy was introduced in 2016 to stimulate the birth rate and mitigate the severity of projected population ageing. However, according to the UN medium projection, the population in China will reach a peak of 1.44 billion in 2029 and then decline steadily to 1.36 billion in 2050 and 1.02 billion in 2100 (United Nations 2017). The recent history of China can be divided into two periods following the foundation of the People’s Republic of China in 1949: the socialist planned economy period until 1977 and the reform period, which has moved toward a socialist market economy since 1978. Each period is characterised by different patterns of internal migration (Shen 2013). In the 1950s through to the early 1980s, a rigid hukou or household registration system was implemented (Cheng and Selden 1994). During that period internal migration, including migration from rural to urban areas, was tightly controlled with the main migration flows being from the eastern region to the central and western regions, usually in the form of organised migration involving agricultural workers, urban residents and soldiers (Tian and Lin 1986). This period included the ‘Cultural Revolution’ from 1966 to 1976, when some 17 million urban youths were moved to rural areas in the ‘Up to the mountains and down to the countryside’ campaign. Most of these youths returned to cities after the end of ‘Cultural

J. Shen (*) Department of Geography and Resource Management, The Chinese University of Hong Kong, Hong Kong, China e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_4

51

52

J. Shen

Revolution’ (Shen and Tong 1992). According to estimates by Hu and Zhang (1984), there were about 25 million net inter-provincial migrations from 1950 to 1979. Since 1985 the rigid hukou system has been relaxed and Chinese people have been permitted to move to other places to become part of the temporary population, also called the ‘floating population’, without changing their place of hukou registration (Shen 2005, 2013). As a result, population movement in China has increased significantly. The temporary population climbed from 6.1 million to 39.6 million between 1982 and 1990. It then expanded greatly to 109 million in 2000 and reached a peak of 253 million in 2014, before declining to 244 million in 2017 (Chen and Ye 2013; NBS 2018). These estimates include all individuals living in an area for which they do not hold a hukou, regardless of when they arrived. They do not include permanent migrants who have obtained a hukou at their destination. Such permanent migrants are included in alternative migration statistics based on a change in usual residence, as discussed in the next paragraph. The growth in recent inter-provincial migration has closely followed that of the temporary population, which includes intra-province migrants. The number of migrants who changed their province of usual residence within a five-year period was 11.0 million between 1985 and 1990. It jumped by a factor of three to 33.9 million between 1995 and 2000 and then nearly doubled to 60.6 million from 2005 to 2010 (Shen 2013; Population Census Office of State Council and Department of Population and Employment Statistics 2012). While the rate of population growth in China has slowed significantly in the reform period, population movement has increased greatly due to the relaxed hukou policy and rapid industrialisation, urbanisation and economic development, especially in the eastern region of China (Shen 2018). In that context, this chapter examines the intensity, characteristics and spatial patterns of internal migration in China. The data on internal migration, its spatial framework and previous studies are discussed first.

4.2  Internal Migration Data The main and most reliable source of population data in China is the population census, which took place in 1953, 1964, 1982, 1990, 2000 and 2010. It is complemented by 1% sample surveys, which were conducted in 1987, 1995, 2005 and 2015. Migration data were first collected in the 1987 survey and the 1990 census and have been collected ever since. This section outlines how migration data were collected at the 2010 census, which is the main source of data used in this chapter. The 2010 census had four sets of questionnaires. There was a short questionnaire with 12 items (R1–R12) for the whole population. Only a 10% sample of the population had to complete the long questionnaire with 28 items (R1-R28). There was a separate short questionnaire with 11 items (R1–R11) for foreign nationals and people from Hong Kong, Macao and Taiwan living in mainland China. Finally, there was a separate questionnaire covering individuals who died in the last 12 months, with eight items (S1–S8) to be completed by a family member. The next paragraphs

4  Internal Migration in China

53

Table 4.1  Migration data available in China Type of migration Five-year interval migration Five-year interval migration Lifetime migration Place of registration (temporary) migration Place of registration (temporary) migration Place of registration (temporary) migration Place of registration (temporary) migration

2000 census Five-year interval (10%, n = 31) Five-year interval (10%, n = (31 × 4) × 31) Lifetime (10%, n = 31) Place of registration (100%, n = 31) Place of registration (100%, city, town, rural, n = 31)

Place of registration (100%, n = 31 × (31 × 3)

2010 census Five-year interval (10%, n = 31)

Lifetime (10%, n = 31) Place of registration (10%, 100%, n = 31) Place of registration (100%, male, female, city, town, rural, n = 31) Place of registration (10%, n = 4 × 3) Place of registration (10%, n = (31 × 4) × (31 × 3))

Note: Percentages refer to the sample size while the numbers in brackets refer to the number of origins or destinations and the dimensions for migration matrices. See text for further explanation

summarise the type of migration data derived from both the long and short questionnaires and Table 4.1 summarises the data available from the last three censuses. In the long questionnaire, four questions relate to migration. They are place of hukou or household registration (R7), type of hukou (rural or urban) (R10), birthplace (R12) and place of usual residence five years ago (R13). Based on these questions, the following migration data (based on a 10% sample) are available for the 2010 census: five-year interval migration between 31 provinces (based on R13); lifetime migration between 31 provinces (based on R12); place of registration (temporary) migration between 31 provinces (based on R7); place of registration migration between four types of origins and three types of destinations (based on R10); place of registration migration between 31 × 4 origins and 31 × 3 destinations (based on R7 and R10). The four types of origins in each province correspond to four types of local administrative areas, i.e., townships, residents’ committees in towns, villagers’ committees in towns, and sub-districts. The three destinations correspond to city, town and rural areas as defined by the Census Office for the 2010 census. Furthermore, migration flow data are also available for movements from provinces to counties. In the short questionnaire, question R7 on place of hukou is useful to measure migration. Place of registration migration between 31 provinces can be obtained for the entire Chinese population from the 2010 census. In addition, origin-­ destination migration matrices are available for all migrants by sex and type of destination, i.e., city, town or rural area. Similar migration data are available from the 2000 census with some differences as shown in Table 4.1. It is important to note the distinction in Chinese migration data between the concept of migration by place of registration (i.e., temporary migration) and the concept of migration by usual place of residence. Temporary migrants are so defined because their place of registration or hukou differs from their place of residence at the Census. Despite their designation as ‘temporary’, many would have moved to

54

J. Shen

their destination more than five years ago. The temporary population is therefore an accumulated migrant population, similar to lifetime migration. In this chapter this form of movement is referred to as registration migration. If such people move between provinces within a five-year period, they are counted both as inter-provincial migrants and as part of the temporary population. The size of the temporary population is therefore always larger than that of inter-­ provincial temporary migrants in a five-year period. To summarise, lifetime migration is based on the comparison of birthplace with place of enumeration at the census; registration (temporary) migration is based on the comparison of place where hukou is held with place of enumeration at the census; five-year migration is based on comparison of place of usual residence five years ago as stated by respondent with place of enumeration at the census. Inter-provincial migrants in a five-­ year period include both non-hukou migrants (temporary migrants who are not able to move their places of hukou to their destinations) and hukou migrants (permanent migrants who are able to move their places of hukou to their destinations). Generally, most inter-provincial migrants are temporary migrants as one cannot easily change place of hukou. According to a large micro-dataset from the 2000 census, which was extracted randomly from the full census data with a sampling ratio of 0.095%, a total of 123,267 migrants moved between 1995 and 2000. Among these migrants, 32,625 were inter-provincial migrants of whom 27,861 were inter-­ provincial temporary migrants, accounting for 85.4% of total inter-provincial migrants. The temporary population will continue to dominate internal migration in China unless they can get local hukou at their destinations (Shen 2013).

4.3  The Spatial Framework China’s administrative divisions have evolved into a five-tier system, with national, provincial, prefecture, county and township levels as listed in Table 4.2. As shown in Fig. 4.1, provinces are grouped into three regions: eastern, central and western, but these are not formal administrative divisions. As shown in Table 4.2, Mainland China is divided into 311 province-level divisions, including 22 provinces, five autonomous regions and four municipalities (Beijing, Chongqing, Shanghai and Tianjin). Province-level divisions are subdivided into 333 prefecture-level divisions, which include 283 prefecture-level cities, 17 prefectures, 30 autonomous prefectures and three leagues. Prefecture-level divisions are themselves subdivided into 2856 county-level divisions, which include 853 urban districts, 370 county-level cities, 1461 counties, 117 autonomous counties and 55 other units. County-level divisions are further divided into 40,906 township-level units, which include 13,379 townships, 1095 nationality townships, 19,410 towns, 6923 sub-districts and 99 other units. The smallest administrative units are grassroots administrative areas, of

 Numbers of administrative units at the 2010 census.

1

Table 4.2  Administrative areas, Mainland China, 2010 Administrative level Provincial level

Total units 31

Prefecture levela

333

County level

Township/town level

2856

40,906

Types of area 22 Five provinces autonomous regions 283 cities 17 prefectures 853 urban districts

370 cities

49 leagues

Three autonomous leagues 1095 nationality townships One nationality sumu

13,379 townships 96 sumu

Four municipalities 30 autonomous prefectures 1461 counties

Two special districts 19,410 towns

Three leagues 117 autonomous counties One forest district 6923 sub-districts

Two district offices

Source: http://www.xzqh.org/html/show/cn/4851.html a Some publications and datasets refer to 345 prefectures. This number includes four municipalities directly under the central government (Beijing, Tianjing, Shanghai and Chongqing), each of which is counted as two prefectures, four prefectures in the provinces of Henan, Hubei, Hainan and Xinjiang with counties that are directly under the provincial government, and a prefecture in the Hubei province that is not recorded in the 333 count, but excludes a prefecture in Fujian

Fig. 4.1  Regions and provinces of China

56

J. Shen

which there are of four types: villagers’ committees in townships, residents’ committees in towns, villagers’ committees in towns, and residents’ committees in sub-­ districts. Limited data for these units are collected at the census but not analysed here. The major changes to provincial administrative divisions after 1979 were the establishment of Hainan province (separated from Guangdong Province) in 1988 and the Chongqing municipality (separated from Sichuan Province) in 1997. As a result, the number of provinces increased from 29 at the 1987 survey to 30 at the 1990 census and 31 at the 2000 census. These changes are compounded by modification of lower level administrative divisions particularly at prefecture and county levels, which is a challenge for trend analysis. Thus, the spatial units related to each migration dataset depend on the year the data were collected. The number of administrative areas in mainland China in 2010 is shown in Table 4.2. At the 2000 and 2010 censuses, the same set of criteria was used to define urban (both city and town areas) and rural areas based on their functional linkages rather than their administrative classification (see Fig. 1 in Shen (2005) for details). Thus, migration matrices between 31 provinces based on place of registration are available by area type (i.e., city, town or rural area) for the 2000 and 2010 census data. For the 2005 1% survey data and the 2010 census, migration matrices are between 31 × 4 origins (31 provinces × 4 types of grassroots administrative areas) and 31 × 3 destinations (31 provinces × 3 types of city area, town area or rural area) based on place of registration. For the 2000 census, the migration matrices are between 31 origins (31 provinces) and 31 × 3 destinations (31 provinces × 3 types of city area, town area or rural area) based on place of registration (Table 4.1).

4.4  Prior Research Zhu et al. (2016) have recently reviewed the development of population geography in China since the 1980s, with a particular section devoted to internal migration by Chinese scholars published in English and Chinese. This section will therefore only provide a brief review. Previous studies on inter-provincial migration in China have focussed on the spatial patterns (Liang and White 1997; Yang 2000; He and Gober 2003; Poston and Zhang 2008; Sun and Fan 2011), regional concentration of migration flows (He and Pooler 2002; Ding et al. 2005), determinants of migration (Tian and Lin 1986; Zhao 1999; Hare 1999; Liang et al. 2002; Cai and Wang 2003; Li 2004; Fan 2005a; Liu and Shen 2017), the relationship between migration and regional development (Liang et  al. 2002; Zhu 2003; Fan 2005b; Shen and Wang 2016), and the settlement intentions of the floating population (Zhu 2003, 2007; Zhu and Chen 2010; Fan 2011). The increasing urban-rural gap and unequal regional development among eastern, central and western regions have been considered as one of the main reasons for the rise in migration in China (Fan 2005b; Zhu 2003). Hare (1999) analysed the determinants of out-migration and return migration outcomes within the context of conditions at the rural origin. Fu and Gabriel (2012) conducted an analysis of

4  Internal Migration in China

57

internal labour migration in China using the sampling data for 1990–1995. They attempted to identify the role of human capital agglomeration in destination choices of migrants considering population groups with different education and age. In summary, major factors contributing to the growth of migration in China include rural reform, hukou (household registration) reform and regional development (Shen 2012). Liu et al. (2015) analysed the changing spatial patterns of the floating population in China from 2000 to 2010 at county level, defining urban districts in the same city as one single city unit. The study included 287 cities at prefecture level or above, 370 county-level cities and 1627 counties, banners or autonomous counties. He et  al. (2016) also analysed the distinctive spatial patterns of floating population redistribution in China 2000–2010 at county level, focussing on destination rather than inter-county flows, for which data are not available. Cromley et al. (2010) estimated intercensal net interprovincial migration using census data from 1990 and 2000. Various migration models were estimated to identify the effects of spatial structure on interprovincial migration using 1990 census data (Shen 1999). Poncet (2006) studied how migratory forces evolved between 1985 and 1995 among rural-to-urban migrants in China. She found that aggregate migration costs declined between the two periods of 1985–1990 and 1990–1995, while migration restrictions were also relaxed. Shen (2016) used a new method to estimate migration modeling errors by their sources using the case of regional migration in China for the period 2005–2010. By calculating the contributions of various factors to the modeling error, he identified which factors can be modeled more or less accurately. Shen found that spatial interaction was the largest source, accounting for 31.55% of the weighted absolute mean error in predicting migration flows. Thus, the spatial interaction effect remains the most difficult to model. Given the constraints faced by rural migrants in Chinese cities, some do not intend to settle in cities, which has important implications for urbanisation policy and urban planning. Several studies have addressed this issue (Zhu 2003, 2007; Fan 2011). Zhu and Chen (2010) examined the settlement intentions of the floating population in Fujian Province in 2002 and 2006. They found that the intention to settle in cities had increased, but a high proportion of the floating population was still in the process of circulating and choosing their final destination. Based on a survey of migrants in Beijing’s urban villages, Fan (2011) found that the majority of rural migrants did not intend to stay permanently in cities. She found that migrants’ labour markets and social futures in the city were important factors in their settlement intentions. Thus, circular migration and split households became long-term practices among many rural migrants. Their economic performance in cities affected their decisions about whether to settle there. Mohabir et al. (2017) examined how gender, age and sense of belonging affect the choice of rural migrants to stay or return and found that migrant workers preferred to remain in mega-cities rather than move to smaller cities or return home to rural areas. In summary, various survey data and census data have been used in the study of internal migration in China. The migration data for destinations are available at fine spatial scales and some analyses have been made at provincial and county levels.

58

J. Shen

Migration flow data are available for inter-provincial migration and inter-prefecture flow data have been derived from micro census data. Most quantitative modelling and analysis have been based on inter-provincial migration flows. Many studies have also sought to identify the determinants of internal migration in China, but internal migration in China has rarely been compared with that of other countries (Bell et al. 2002).

4.5  How Much Movement? Migration Intensity Migration flow data for China are available for movements between provinces and from provinces to counties. However, data for the total numbers who moved between townships/towns/sub-districts are also available for registration migration in the 2010 census. The 100% short form data indicate that some 260.94 million Chinese people were living in another township-level unit where they did not have hukou in 2010, which corresponds to a crude migration intensity (CMI) of 19.58%. This is very similar to the results obtained from the 10% long form data (19.13%). As shown in Table  4.3, the CMI can be decomposed into 6.78% intra-county migrants, 6.35% inter-county migrants within the same province and 6.44% inter-­ provincial migrants. These are perhaps the most reliable indicators of crude migration intensity for China. The hukou system records all residential address changes if reported, but only an address change between townships/towns/sub-districts is considered as an official change in the place of hukou. A township/town/sub-district is a local area and migration within such local area is considered less important. However, the scale of such migration is likely to be large as many residents may change residence especially within townships/towns. County-level lifetime migration data from the 2010 census, based on 10% long form data, indicate that 21.67 million (about 216.69 million for the whole population) Chinese people had left their county or district of birth. This corresponds to a CMI of 17.02%, which is higher than the inter-county CMI of 12.80% for registration migration. This difference arises because registration migration only includes migrants who do not have hukou at their destination while lifetime data include all migrants regardless of their hukou status. This means that just one quarter of inter-­ county migrants have a hukou at destination. A similar discrepancy is observed for inter-provincial migration, with a CMI of 8.97% for lifetime migration compared with 6.35% for registration data. In other words, in 2010 only 29% of Chinese who had moved to another province had been granted hukou at their destination. Results for recent migration from the 10% long form data indicate that 5.52 million (about 55.23 million for the whole population) Chinese people had changed their usual place of residence between 2005 and 2010, which corresponds to a CMI of 4.61%. This is lower than the inter-provincial lifetime CMI of 8.05%. As shown by the registration and lifetime migration data, the crude migration intensities for intra-county migration, intra-province inter-county migration and inter-provincial migration are numerically very similar. If this is also the case for five-year interval migration, it can be estimated that the five-year interval total crude migration

4  Internal Migration in China

59

Table 4.3  Crude migration intensity by type of move and spatial scale, 2010 Migration Total Total Indicator population migrants Migration by place of registration Population 1332.81 260.94 (million) CMI (%)

not applicable

19.58

Population (million)

127.34

24.36

CMI (%)

not applicable

19.13

Intra-­ county

Inter-county same Inter-­ province province

Total inter-­ county

90.37

84.69

85.88

170.57

6.78

6.35

6.44

12.80

11.42

10.25

21.67

8.97

8.05

17.02

Lifetime migration 127.34 Population (million) CMI (%)

not applicable

Five-year interval migration by place of residence Population 119.89 (million)

5.52

CMI (%)

4.61

not applicable

14.00

4.85

4.54

9.15

Source 100% short form 100% short form 10% long form 10% long form 10% long form 10% long form 10% long form 10% long form

Source: Author’s calculation from the 2010 census Note: CMI crude migration intensity. Underlined figures were estimated by the author. See text for explanation

intensity was 14.00% (corresponding to 167.81 million migrants for the total population) for the interval 2005–2010. Compared with other countries in Asia and globally, migration intensity in China is moderate. Estimates of all changes of address (aggregate crude migration intensity) from the 2000 census indicate that nearly 13% of the Chinese population changed address in the previous five years, which is below the Asian mean of 15.85% and well below the global mean of 21.10% (Charles-Edwards et al. 2017). With an intensity on par with Iran, China has a lower migration intensity than Malaysia and South Korea, but higher than Vietnam, Indonesia and Thailand as shown in the concluding chapter of this book. As noted in the introduction, China had a low level of mobility from the 1950s to the 1970s due to hukou control, but its mobility has now increased to a medium level. It is likely that mobility will further increase, especially between and within cities, with further development of the country.

60

J. Shen

Table 4.4 presents a series of migration indicators based on the 1% sample of census data from the 2010 and 2010 censuses provided by the IMAGE-Asia Project. Five-year interval data by place of usual residence refer to migration among 347 regions. Table 4.4 is useful to show the evolution of crude migration intensity. It is clear that in the decade from the 2000 census to the 2010 census, inter-provincial crude migration intensity doubled for both registration migration and five-year interval migration. However, the inter-provincial lifetime crude migration intensity and five-year interval crude migration intensity among 347 regions increased only slightly over that period. The ratio of inter-provincial lifetime intensity to five-year interval migration intensity by place of residence (n = 31) was 1.75 in 2010, smaller than all other Asian countries, and smaller even than at the 2000 census when the ratio was 2.21. This ratio indicates that the current intensity of migration in China is high in comparison to historical experience, and that the country is experiencing a phase of migration expansion. This is in contrast to Indonesia and Malaysia, which are witnessing a contraction of migration as indicated by high ratios of 4.78 and 5.76, respectively. Other indicators in Table 4.4 (the MEI and the ANMR) are discussed later in this chapter.

Table 4.4  System-wide migration indicators by type of move, 2000 and 2010 Crude migration intensity (%) Type of migration 2010 Census Place of registration (n = 31) 6.19 Five-year interval (n = 31) by place of 4.56 residence Lifetime (n = 31) 7.98 Five-year interval (n = 347) by place 8.99 of residence Ratio of lifetime intensity to five-year 1.75 interval migration intensity by place of residence (n = 31) 2000 Census Place of registration (n = 31) 3.10 Five-year interval (n = 31) by place of 2.79 residence Lifetime (n = 31) 6.17 6.70 Five-year interval (n = 347) by place of residence Ratio of lifetime intensity to five-year 2.21 interval migration intensity by place of residence (n = 31)

Aggregate net Migration effectiveness index migration rate (%) (%) 69.79 62.27

4.32 2.84

60.66

4.84

68.11 61.21

2.11 1.71

45.47

2.81

Source: Calculated from the 2000 and 2010 Censuses (IMAGE-Asia Project) Note: The number in brackets refer to the number of regions. The crude migration intensity values are close to those shown in Table 4.3, which are based on the 100% or 10% census data

4  Internal Migration in China

61

4.6  Who Moves? The Characteristics of Migrants Migration is a selective process, particularly with respect to age and level of educational attainment. Sex-selectivity is also a dimension that has attracted scholarly attention in China (Sun and Fan 2011) and elsewhere in Asia, although there appear to be important variations between countries. This section examines the characteristics of migrants in China, paying particular attention to sex, age and education, and concludes by explaining migrant selectivity in light of the reasons for moving. The 2010 census provided the most detailed data on migration by place of registration and this section will mainly use this data source. As shown in Table 4.5, the crude migration intensities for registration migration were close, 20.1% and 19.1% for males and females respectively. The sex ratio (number of males per 100 females) for registration migration was 110.5, higher than the 104.9 for the total population in 2010, indicating that there were slightly more male migrants than female migrants in China. However, variations emerge when considering the distance moved. Females were more mobile than males in intra-county migration with a sex ratio of 97.5 while the reverse was true for inter-county and inter-provincial migration. In other words, women are more likely to move short distances but less likely to move long distances. Figure 4.2 presents age-specific migration intensities by sex at a range of spatial scales. Both male and female migration peaked at age 20 with intensities of 33.80% and 35.95% respectively. They were above 15% between ages 15–46 for males and 15–49 years for females with a sharp increase between ages 15 and 20, presumably linked to school completion. Male and female intra-provincial migration intensities peaked at age 17, at 24.33% and 26.77% respectively while inter-provincial intensities peaked at age 25 for males and 21 for females with figures of 14.64% and 12.37%. While males report higher inter-provincial migration intensity, females often engage in marriage migration (Fan and Huang 1998) and show higher intra-­ provincial migration intensity than males at young adult ages. Many students also move within a province at young ages. Such marriage and student migration produced younger peaks for intra-provincial migration than for inter-provincial migration. Overall, intra-provincial migration intensity was higher than that for inter-provincial migration in 2010. Table 4.6 compares the educational level of the total population with migrants aged six and over in 2010. Over 39% of the Chinese population has junior

Table 4.5  Crude migration intensity by sex and type of move, registration migration, 2010 Sex Total Male Female Sex ratio

Total migration 19.58 20.07 19.06 110.50

Intra-county 6.78 6.54 7.03 97.54

Inter-county within provinces 6.35 6.45 6.26 108.10

Source: Author’s calculations based on 100% short form 2010 census data

Inter-province 6.44 7.09 5.77 128.89

62

J. Shen 30.0 Intra-provincial male Intra-provincial female

Migration intensity (%)

25.0

Inter-provincial male Inter-provincial female

20.0 15.0 10.0 5.0 0.0

5

10 15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95 Age

Fig. 4.2  Age-specific migration intensities by sex and type of move, registration migration, 2010. (Source: 2010 Census) Table 4.6  Total population and migrants aged six and over by education level, 2010 Total population All migrants Intra-provincial Inter-provincial Educational level Male Female Male Female Male Female Male Female No schooling 2.76 7.33 1.05 2.87 1.14 3.12 0.87 2.32 Primary school 26.58 31.01 15.16 17.70 14.87 16.95 15.69 19.43 Junior secondary school 44.06 39.25 42.09 39.74 35.47 34.76 54.10 51.21 Senior secondary school 16.42 13.56 23.40 21.86 26.55 24.58 17.68 15.59 University non-degree 5.82 5.21 9.89 9.97 12.00 11.66 6.05 6.09 University degree 3.98 3.35 7.75 7.29 9.19 8.32 5.15 4.91 Graduate school 0.37 0.29 0.67 0.57 0.78 0.61 0.47 0.46 Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 Source: Author’s calculations based on 100% short form 2010 census data

secondary school education and a further 26% received primary school education. Seven percent of the female population had no schooling and 31% had only primary school education, which is much greater than that of the male population. Over 8% of the population received university education or above, with a higher percentage for males. On average migrants were better educated than the general population. More than 17% of both male and female migrants had received university-level education and over 21% had received senior secondary school education. Intra-provincial migrants were better educated than inter-provincial migrants and the latter had only

4  Internal Migration in China

63

Table 4.7  Registration migration by reason and type of move, 2005–2010 Migration reason Work and business Job transfer Study and training Dependents Joining relatives and friends Home move due to demolition Hukou attachment Marriage Other Total

Total Male 50.53 4.66 10.90 11.78 3.75 9.39 0.75 1.60 6.63 100.00

Female 39.14 2.95 11.98 16.81 4.73 9.21 0.68 8.39 6.12 100.00

Intra-province Male Female 35.26 25.86 5.67 3.35 14.52 15.21 14.02 19.17 4.33 5.04 14.08 12.79 1.08 0.92 2.12 9.85 8.92 7.81 100.00 100.00

Inter-province Male Female 78.51 69.74 2.82 2.04 4.28 4.55 7.68 11.36 2.69 4.01 0.79 0.95 0.14 0.13 0.65 5.02 2.43 2.22 100.00 100.00

Source: Author’s calculations based on 100% short form 2010 census data, all ages

a slightly higher education level than the total population, due to a higher proportion with junior secondary school education. This reflects the fact that many inter-­ provincial migrants were migrant workers while better educated intra-provincial migrants moved for other reasons (see Table 4.7). In terms of sex differences, there are higher proportions of males with secondary school education in the total population. This male advantage was weaker among both intra-provincial and inter-provincial migrants. There were fewer male migrants with no schooling but the proportions with university education were similar between male and female migrants. Table 4.7 presents self-reported reasons for migration. Main reasons include work and business (39–50%), moving as dependents (11–17%), and study and training (10–12%). Migration due to demolition of housing is relatively common and is thought to be related to urban renewal (9%) for both male and female migrants. Important variations emerge when considering distance moved. As in other countries, long-distance migration is mainly linked to employment with over 69% of inter-provincial female migrants and 79% of males moving for work and business compared with about 7% of males and 11% of females moving as dependents. Smaller proportions of intra-provincial migrants, just 35% of males and 25% of females, moved for work and business, while 14–15% moved for study and training, 14–19% moved as dependents and 13–14% moved due to home demolition. It is clear that intra-provincial migrants moved for more diverse reasons. Motives for moving are clearly gendered. Overall, a smaller proportion of female migrants, 39%, moved for work and business, but 17% moved as dependents and 8% moved for marriage. Note also that the extent of sex differentials is in part concealed because all ages are combined. There were similar sex differences among intra-provincial and inter-provincial migrants.

64

J. Shen

4.7  Where Do They Move? Spatial Patterns Migration does not necessarily bring about significant changes in the distribution of the population because some flows are offset by counter-flows. The net redistribution of population is the product of two parallel processes: (1) the intensity of movement, which is measured by the CMI and (2) the degree of asymmetry between flows and counter-flows, which is measured by the Migration Effectiveness Index (MEI). The aggregate net migration rate (ANMR), which is the product of these two measures divided by 100, quantifies the extent of population redistribution across all provinces in China. As shown in Table 4.4, the overall redistributive effect of migration across provinces over a five-year interval increased from 1.71% to 2.84% between the last two censuses. This is the result of an increase in migration intensity from 2.79% to 4.56%, which was boosted by a slight rise in the MEI from 61.21 to 62.27. In any case, this is a very high value, which points to highly unbalanced flows, leading to the redistributive effect of migration in China being intermediate by regional and international standards, as discussed in this book’s concluding chapter. While these measures provide summary indicators of the impact of migration on overall population redistribution within China, they provide no information as to its spatial manifestation. Figure 4.3 presents bilateral interprovincial flows and shows that the five largest flows with over one million migrants in the period 1995–2000 were from Hunan, Sichuan, Guangxi, Jiangxi and Hubei to Guangdong. From 1995–2000 to 2005–2010, the number of inter-provincial migrants nearly doubled from 33.9 to 60.6 million (Shen 2013; Population Census Office of State Council and Department of Population and Employment Statistics 2012). The number of flows of over one million migrants also increased, adding a further four to those listed above: Anhui to Jiangsu, Anhui to Zhejiang, Anhui to Shanghai and Henan to Guangdong (Fig. 4.4). Table 4.8 presents migration indicators at a provincial level for 1995–2000 and 2005–2010, and Figs.  4.5 and 4.6 map net migration rates for the same period. Shanghai, Beijing, Zhejiang, Guangdong and Tianjin had the highest net gain of over 14% in the period 2005–2010, while Jiangsu and Fujian registered gains of 7–9% in the same period. Anhui had the highest net loss of 10.36% followed by Sichuan, Chongqing, Guizhou, Guangxi, Henan, Hubei, Hunan and Jiangxi with 5–10%. For most provinces the net migration rate was smaller in the period 1995–2000 than in 2005–2010 although the spatial patterns were similar. Qinghai and Nei Mongol changed from net out-migration to net in-migration while Shanxi and Yunnan changed from net in-migration to net out-migration. As a result of high in-migration and low out-migration, Shanghai, Beijing, Zhejiang, Guangdong and Tianjin recorded net migration rates exceeding 12% between 2005 and 2010. They also exhibited very high migration effectiveness ratios (MERs), greater than 70%, indicating that inflows were poorly compensated by outflows. Jiangsu and Fujian had lower net migration rates of around 4.5%. Together these seven provinces form the group labelled ‘E’ on Fig. 4.7 comprising developed, highly urbanised, coastal areas. Following the model proposed by Rees et al. (2017),

4  Internal Migration in China

65

Fig. 4.3  Bilateral migration flows, provinces of residence, China, 1995–2000. (Source: Calculated from the 2000 Census (IMAGE-Asia project))

this plot sets net migration rates against the logarithm of population density, as a proxy for the degree of urbanisation. Xinjiang, Qinghai, Xizang and Nei Mongol form a separate group (W) of low density sparsely populated provinces in the western and northern parts of the country with positive but low net migration gains, below 3%. The third cluster (group C) brings together densely populated areas with net migration losses. This includes the provinces of Guizhou, Hunan, Jiangxi and Anhui, which recorded net migration losses of more than 6% in the period 2005–2010. According to Rees et  al. (2017), a positive slope indicates that more densely populated areas are gaining through net internal migration, while less densely populated areas are losing. China’s slope of 4.86 is well above many Asian countries. This indicates strong population gains in the more densely population regions through migration and is consistent with the fact that, during the first decade of the

66

J. Shen

Fig. 4.4  Bilateral migration flows, provinces of residence, China, 2005–2010. (Source: Calculated from the 2010 Census (IMAGE-Asia Project))

millennium, China was at a stage of rapid urbanisation with large scale rural to urban migration linked to industrialisation and economic development.

4.8  Understanding Internal Migration in China From the 1950s to the early 1980s, internal migration in China was tightly controlled. A balanced development strategy was adopted and development focussed on the central and western regions. The main flows were from the eastern region to the central and western regions (Tian and Lin 1986). Hu and Zhang (1984) estimated that there were about 25  million net inter-provincial migrations in the period 1950–1979. Eight provinces (Shanghai, Jiangsu, Zhejiang, Shandong, Henan,

4  Internal Migration in China

67

Table 4.8  Migration indicators, provinces of China, 1995–2000 and 2005–2010

Area Shanghai Beijing Zhejiang Guangdong Tianjin Jiangsu Fujian Xinjiang Qinghai Xizang (Tibet) Liaoning Nei Mongol Hainan Ningxia Shanxi Yunnan Shandong Shaanxi Hebei Jilin Gansu Heilongjiang Chongqing Guangxi Sichuan Hubei Henan Guizhou Hunan Jiangxi Anhui Total

1995–2000 Out-­ migration rate In-migration (OMR) rate (IMR) 1.33 15.32 1.74 16.57 2.57 6.51 0.68 16.34 1.14 5.38 1.82 2.71 2.06 4.20 1.19 7.02 2.88 1.82 1.99 3.48

Net migration rate (NMR) 13.99 14.83 3.93 15.66 4.23 0.89 2.14 5.83 −1.06 1.49

Migration effectiveness ratio (MER) OMR IMR NMR MER 84.02 2.77 30.53 27.76 83.34 80.95 2.75 25.89 23.14 80.78 43.32 2.90 18.63 15.73 73.09 92.00 2.06 17.06 15.00 78.49 64.88 2.35 14.34 12.00 71.87 19.67 2.77 7.32 4.56 45.16 34.16 3.70 8.05 4.35 37.04 71.05 1.34 4.24 2.89 51.83 −22.52 2.25 4.13 1.87 29.33 27.13 2.56 4.09 1.52 22.88

0.81 2.06 1.60 1.42 0.95 0.99 1.01 2.04 1.52 2.23 2.57 2.55 5.62 4.18 5.30 4.18 2.64 3.53 5.43 7.08 4.86 2.79

1.15 −0.65 0.96 0.81 0.54 0.86 −0.01 −0.66 −0.24 −1.09 −1.78 −1.63 −3.93 −3.37 −4.46 −3.11 −2.12 −2.63 −4.76 −6.41 −4.21 n/a

41.46 −18.68 23.13 22.22 22.02 30.49 −0.28 −19.28 −8.72 −32.08 −52.70 −46.73 −53.77 −67.56 −72.64 −59.20 −67.13 −59.36 −78.18 −82.59 −76.21 n/a

1.96 1.41 2.56 2.23 1.49 1.85 1.00 1.38 1.28 1.15 0.80 0.93 1.69 0.81 0.84 1.07 0.52 0.90 0.66 0.68 0.66 2.79

2005–2010

1.71 2.80 3.33 3.32 2.30 2.60 2.32 3.70 2.86 3.35 3.94 4.11 7.58 6.17 6.34 7.07 5.92 7.80 7.54 8.56 10.36 4.56

2.83 3.63 4.13 4.00 1.53 1.68 1.36 2.28 1.38 1.36 0.93 0.90 2.92 1.42 1.33 1.66 0.42 1.78 1.16 1.65 1.35 4.56

Source: Calculated from the 2000 and 2010 Censuses (IMAGE-Asia Project) Note: The data are sorted by net migration rate in 2010

1.12 0.84 0.80 0.68 −0.77 −0.92 −0.96 −1.42 −1.48 −1.99 −3.01 −3.21 −4.67 −4.75 −5.01 −5.42 −5.49 −6.02 −6.38 −6.91 −9.01 0.00

24.73 13.03 10.74 9.27 −20.06 −21.56 −26.12 −23.72 −34.88 −42.24 −61.84 −64.20 −44.46 −62.57 −65.34 −62.03 −86.63 −62.81 −73.40 −67.65 −76.88 n/a

68

J. Shen

Fig. 4.5  Net migration rates, provinces of China, 1995–2000. (Source: Calculated from the 2000 Census (IMAGE-Asia Project))

Hunan, Guangdong and Sichuan) recorded net out-migration. The net migration in Liaoning, Anhui and Fujian was negligible. Other provinces recorded net in-­ migration with Nei Mongol (3.1 million), Heilongjiang (6.5 million) and Xinjiang (2.8 million) accounting for half the total net migration gain. As revealed in the previous sections, the spatial patterns of migrations in the reform period have been characterised by rural-to-urban migration from western and central provinces to the east. This is a result of urbanisation and regional development (Shen 2018), but the majority have moved as temporary migrants without local hukou at their destinations. A coastal development strategy was adopted in the 1980s and the focus of regional development shifted to coastal areas in the eastern region. Economic reform and open-door policies were implemented much earlier and faster in the eastern region than in the rest of the county. Many areas in the eastern region, notably Guangdong, Jiangsu and Zhejiang, have benefited from rapid economic growth since the late 1970s, which has created many opportunities for migrants from other regions.

4  Internal Migration in China

69

Fig. 4.6  Net migration rates, provinces of China, 2005–2010. (Source: Calculated from the 2010 Census (IMAGE-Asia Project))

Economic reform and the open-door policy have been implemented with increasing intensity since 1978. They also have a clear pattern of spatial diffusion. Economic reforms were introduced first in special economic zones (1979) and open coastal cities (1984) and were then extended to open economic areas (1985), Pudong New District (1990) and all parts of China in 1994. Following this staged reform, rapid economic growth took place first in the Pearl River Delta (Guangdong) in the 1980s, then spread to the Yangtze River Delta (Shanghai, Jiangsu and Zhejiang) in the 1990s and Jing-Jin-Tang region (Beijing, Tianjin and Tangshan) in north China in the twenty-first century (Yeung and Shen 2008). Specific regional development polices have also targeted the western region of China since the late 1990s, and northeast and central China in the twenty-first century. As a result, some areas led as major migrant destinations followed later by others. Thus, economic reforms in China since the late 1970s have caused a reversal in the direction of migration whereby more migrants have moved from the western and central regions to the eastern region (Fan 2005b; Shen 2008). Clearly, the growth in population movement has been the result of economic reform toward a market economy and a deregulation of migration control. Rural

70

J. Shen

Fig. 4.7  Net migration rates by log population density, provinces of China, 2005–2010. (Source: Calculated from the 2010 Census. Note: E: developed coastal areas with high net migration gains; W: sparsely populated western and northern areas with low net migration gains; C: densely populated areas with low net migration gains or net losses)

reform and rural development since the 1980s have improved agricultural productivity and released many surplus rural workers. Initially, ‘leave the land but not the home town’ was encouraged, resulting in rural industrialisation and rural urbanisation. Since the 1990s, ‘leave the land and the home town’ has become popular and many rural migrants have moved to coastal areas (Shen 2018). The institutional driver, relaxation and reform in the hukou system, is considered the most important factor serving to increase mobility in post-reform China. Shen (2013) argues that such institutional factors were influential in the initial increase of migration in the 1980s. However, subsequent increases in the scale of migration after 1990 were driven largely by rapid and unbalanced economic development. This is confirmed by the results of inter-provincial migration modeling (Shen 2012). In a multilevel Poisson model, urban income and rural income per capita at the destination were found to have a positive impact on migration, confirming the importance of income in internal migration in China in the period 1985–1990 and 1995–2000 (Shen 2012). The share of tertiary sector employment in 2005 also had a positive impact on inter-provincial migration in the period 2005–2010 (Shen 2016).

4  Internal Migration in China

71

4.9  Impacts and Implications Population movement has increased significantly since the early 1980s. Most migration takes the form of temporary movement, rural-to-urban migration and inter-­ provincial migration toward coastal cities. This means that most migrants moved without a local hukou at destination. Thus, internal migration has been a key factor in the rapid urbanisation of China in recent decades. For simplicity, the increase in hukou population is often termed a formal urbanisation process while the increase in temporary population without hukou is termed an informal urbanisation process (Zhu 1998; Shen et al. 2002). According to recent population censuses (NBS 2011), the urban population of China more than tripled from 1982 to 2010, increasing from 210.82 to 665.57 million. As a result, the level of urbanisation reached 49.7% in 2010. Temporary migrants are the main driving force behind the increase in the urban population. Some 33.73% of the urban population did not have local hukou in China in 2010 compared with 24.73% in 2000. With the presence of a large number of temporary migrants without local hukou in Chinese cities, an important question remains as to whether the government should encourage, control or regulate informal but ‘legal’ migration from rural areas? With the progress of the market economy, various discriminative measures based on hukou have been abolished gradually since 1997 (Shen 2006). For example, an experiment was started in May 1997 in 450 small towns to grant full permanent resident rights, valid only locally, to migrants who have had stable employment, income and housing for two years (Zhang 1999). A significant development in hukou reform occurred in November 2011. The State Council announced new policies toward further hukou reform in a notice that described guidelines on hukou reform (General Office of State Council 2011). According to that document, local hukou should be given to people with stable jobs and accommodation and to their family members in urban areas of county-level cities and designated towns. In small and medium cities with urban districts under their administration, local hukou should be given to people who have had stable jobs and accommodation for over three years and have joined the social insurance scheme for a certain number of years, and to their family members. But the document stipulates that more rigorous requirements can be imposed if population pressure exceeds the urban carrying capacity. However, there has been no hukou policy change for cities under central administration, vice-provincial level cities and other large cities. These cities will continue to implement existing hukou policy and control population size. Indeed, Beijing Municipal Government took drastic measures to demolish illegal buildings in late 2017 and many temporary migrants had to move to other places in the city or return to their hometowns. This document also provides basic policy guidelines on rural development and rural migrants as many rural migrants are not likely to be granted local hukou in large cities. It states that policies and regulations should be improved to solve the problems facing rural migrants such as wages, child education, skill training, public health, house purchase and rental, social security, occupational safety and health. A

72

J. Shen

system of resident cards for the temporary population should be gradually implemented. Currently only a small number of migrants can obtain resident cards, which give access to welfare services similar to those available to local residents. The hukou policy reform has been shifted over time and this will remain a significant focus of urban studies and migration policy in China into the foreseeable future (Shen 2018). In general, internal migration in China has been driven by economic opportunities in the coastal regions of China as well as other social and institutional factors, including the hukou system. With diffusion of economic growth, rural development and household changes, new patterns of migration may emerge.

4.10  Conclusions This chapter has examined the intensity, characteristics, spatial patterns and impacts of internal migration in China. The data on internal migration, its spatial framework and previous studies were reviewed and discussed. China has collected migration data since the mid-1980s, primarily for migration between provinces; migration flow data at areas below province level are not usually available. There are two challenges for analysis of migration in China. First, the number of provinces increased in 1988 and again in 1997 when Hainan province and then Chongqing Municipality were established. Second, there is a distinction between hukou and non-hukou migrants, which has important social and policy implications. In 2010, the crude migration intensity based on registration migration was 19.58%. In other words, nearly one in five Chinese do not have a hukou at their destination and are part of the growing floating population. This crude migration intensity can be decomposed into 6.78% for intra-county migration, 6.35% for inter-county migration within the same province and 6.44% for inter-provincial migration. In contrast, the inter-provincial five-year crude migration intensity was 4.61%, which includes both hukou and non-hukou migrants. The five-year interval crude migration intensity for all changes of address or Aggregate Crude Migration Intensity (ACMI) was 14.00% at the 2010 census, which is higher than the estimated ACMI of 12.82% at the 2000 census and confirms an increase in the level of population movement. This is below both the Asian mean of 15.9% and global mean of 21.1%. While China currently exhibits a moderate level of migration intensity compared with other Asian countries, it has experienced a progressive increase since the low mobility of the 1950s–1970s to a medium level following economic reform and relaxation of the hukou system since the 1980s. Mobility may increase further, especially inter-city and intra-city migration, as the country further develops. Evidence of rising migration can be seen in the inter-provincial crude migration intensity, which doubled between the 2000 and the 2010 censuses for both registration and five-year interval migration. According to the place of registration migration data in 2010, there were slightly more male migrants than female migrants in

4  Internal Migration in China

73

China. Both rates peaked at age 20; 33.80% for males and 35.95% for females. Intra-provincial migrants had a much higher education level than inter-provincial migrants and compared with the total population. Inter-provincial migrants had a slightly higher education level than the total population with a larger proportion having junior secondary school education. The main reasons for migration included work and business (39–50%), moving as dependents (11–17%), study and training (10–12%) and home move due to demolition (9%) for both male and female migrants. Migration effectiveness exceeded 68% in both the 1995–2000 and 2005–2010 intervals. This ratio was the highest of the Asian countries and indicated that China was at a stage of rapid urbanisation, industrialisation and development with large scale rural to urban migration. In summary, population movement has increased significantly since the early 1980s. Most migration takes the form of temporary, rural-to-urban, inter-provincial migration, particularity to coastal cities. This means that most migrants move without local hukou at their destination. Thus, internal migration is closely related to the rapid urbanisation process in China in recent decades. Acknowledgements  I would like to thank Martin Bell, Yu Zhu, Aude Bernard, Elin Charles-­ Edwards and Gordon Kee for their comments and help with some data and figures. This work was supported by a research grant from the Chinese University of Hong Kong (Project No. 4052139).

References Bell, M., Blake, M., Boyle, P., Duke-Williams, O., & Hugo, G. J. (2002). Cross-national comparison of internal migration: Issues and measures. Journal of the Royal Statistical Society Series A, 165, 435–464. Cai, F., & Wang, D. (2003). Migration as marketization: What can we learn from China’s 2000 census data? The China Review, 3(2), 73–93. Charles-Edwards, E., Bell, M., Bernard, A., & Zhu, Y. (2017). Internal migration in the countries of Asia: Levels, ages, and spatial impacts (ADRI Working Paper Series, ADRI-WP-2017/001). Chen, B., & Ye, Y. (2013). Main characteristics of floating population and its impact on urbanization in China. Urban Issues, 3, 2–8. Cheng, T., & Selden, M. (1994). The origins and social consequences of China’s hukou system. China Quarterly, 139, 644–668. Cromley, R. G., Hanink, D. M., & Ebenstein, A. Y. (2010). Estimating and mapping the intercensal internal net migration of China, 1990–2000. Cartography and Geographic Information Science, 37(3), 229–238. Ding, J., Liu, Z., Cheng, D., Liu, J., & Zou, J. (2005). Areal differentiation of inter-provincial migration in China and characteristics of flow field. Acta Geographica Sinica, 60(1), 106–114. Fan, C. C. (2005a). Modeling interprovincial migration in China, 1985–2000. Eurasian Geography and Economics, 46(3), 165–184. Fan, C. C. (2005b). Interprovincial migration, population redistribution, and regional development in China: 1990 and 2000 census comparisons. The Professional Geographer, 57(2), 295–311. Fan, C. C. (2011). Settlement intention and split households: Findings from a survey of migrants in Beijing’s urban villages. The China Review, 11(2), 11–42.

74

J. Shen

Fan, C. C., & Huang, Y. (1998). Waves of rural brides: Female marriage migration in China. Annals of the Association of American Geographers, 88(2), 227–251. Fu, Y., & Gabriel, S. A. (2012). Labor migration, human capital agglomeration and regional development in China. Regional Science and Urban Economics, 42, 473–484. General Office of State Council. (2011). Notice of general office of state council on actively and stably moving forward the reform of hukou administration system, No 9. http://www.gov.cn/ zwgk/2012-02/23/content_2075082.htm. Accessed on 27 March 2019. Hare, D. (1999). Push versus pull factors in migration outflows and returns: Determinants of migration status and spell duration among China’s rural population. Journal of Development Studies, 35(3), 45–72. He, C., & Gober, P. (2003). Gendering interprovincial migration in China. The International Migration Review, 37(4), 1220–1251. He, J., & Pooler, J. (2002). The regional concentration of China’s interprovincial migration flows, 1982–90. Population and Environment, 24(2), 149–182. He, C., Chen, T., Mao, X., & Zhou, Y. (2016). Economic transition, urbanization and population redistribution in China. Habitat International, 51, 39–47. Hu, H., & Zhang, S. (1984). Population geography of China. Shanghai: East China Normal University Press. Li, S. (2004). Population migration and urbanization in China: A comparative analysis of the 1990 population census and the 1995 national one percent sample population survey. International Migration Review, 38(2), 655–685. Liang, Z., & White, M.  J. (1997). Market transition, government policies, and interprovincial migration in China: 1983–1988. Economic Development and Cultural Change, 45, 321–336. Liang, Z., Chen, Y. P., & Gu, Y. M. (2002). Rural industrialisation and internal migration in China. Urban Studies, 39(12), 2175–2187. Liu, Y., & Shen, J. (2017). Modelling skilled and less-skilled interregional migrations in China, 2000–2005. Population, Space and Place, 23(4), e2027 1–17. Liu, T., Qi, Y., & Cao, G. (2015). China’s floating population in the 21st century: Uneven landscape, influencing factors, and effects on urbanization. Acta Geographica Sinica, 70(4), 567–581. Mohabir, N., Jiang, Y., & Ma, R. (2017). Chinese floating migrants: Rural-urban migrant labourers’ intentions to stay or return. Habitat International, 60, 101–110. NBS. (2011). China statistics yearbook 2011. Beijing: China Statistics Press. NBS. (2017). China statistics yearbook 2017. Beijing: China Statistics Press. NBS. (2018, January 18). Economy performed stable and became better and better than expected. http://www.stats.gov.cn/tjsj/zxfb/201801/t20180118_1574917.html. Accessed 19 Jan 2018. Poncet, S. (2006). Provincial migration dynamics in China: Borders, costs and economic motivations. Regional Science and Urban Economics, 36(3), 385–398. Population Census Office of State Council and Department of Population and Employment Statistics. (2012). Tabulations on the 2010 population census of the People’s Republic of China (Vol. 3). Beijing: China Statistical Press. Poston, D. L., Jr., & Zhang, L. (2008). Ecological analyses of permanent and temporary migration streams in China in the 1990s. Population Research and Policy Review, 27, 689–712. Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), e2036 1–22. Shen, J. (1999). Modelling regional migration in China: Estimation and decomposition. Environment and Planning A, 31, 1223–1238. Shen, J. (2005). Counting urban population in Chinese censuses 1953–2000: Changing definitions, problems and solutions. Population, Space and Place, 11(5), 381–400. Shen, J. (2006). Understanding dual-track urbanisation in post-reform China: Conceptual framework and empirical analysis. Population, Space and Place, 12(6), 497–516.

4  Internal Migration in China

75

Shen, J. (2008). Population distribution and growth. In Y. M. Yeung & J. Shen (Eds.), The Pan-­ Pearl River Delta: An emerging regional economy in a globalizing China (pp. 513–548). Hong Kong: Chinese University Press. Shen, J. (2012). Changing patterns and determinants of inter-provincial migration in China, 1985–2000. Population, Space and Place, 18, 384–402. Shen, J. (2013). Increasing internal migration in China from 1985 to 2005: Institutional versus economic drivers. Habitat International, 39(1), 1–7. Shen, J. (2016). Error analysis of regional migration modeling. Annals of the American Association of Geographers, 105(6), 1253–1267. Shen, J. (2018). Urbanization, regional development and governance in China. London/New York: Routledge. Shen, Y., & Tong, C. (1992). The population migration in China – Historical and contemporary perspectives. Beijing: China Statistical Press. Shen, J., & Wang, L. (2016). Urban competitiveness and migration in the YRD and PRD regions of China in 2010. The China Review, 16(3), 149–174. Shen, J., Wong, K. Y., & Feng, Z. (2002). State sponsored and spontaneous urbanization in the Pearl River Delta of South China, 1980–1998. Urban Geography, 23(7), 674–694. Sun, M., & Fan, C.  C. (2011). China’s permanent and temporary migrants: Differentials and changes, 1990–2000. The Professional Geographer, 63(1), 92–112. Tian, F., & Lin, F. (Eds.). (1986). Population migration in China. Beijing: Knowledge Press. United Nations, Department of Economic and Social Affairs, Population Division. (2017). World population prospects: The 2017 revision (DVD Edition). Yang, H. (2000). A comparative analysis of China’s permanent and temporary migration during the reform period. International Journal of Social Economics, 27(3), 173–193. Yeung, Y., & Shen, J. (2008). Coastal China’s urban-rural spatial restructuring under globalization. In Y.  Huang & A.  M. Bocchi (Eds.), Reshaping economic geography in East Asia (pp. 294–319). Washington, DC: The World Bank. Zhang, L. (1999). The household registration system and rural to urban migration in China. In Y. M. Yeung, D. Lu, & J. Shen (Eds.), China toward the 21st century: Urban, rural and regional development (pp. 201–221). Hong Kong: Hong Kong Institute of Asia-Pacific Studies. Zhao, Y. (1999). Labor migration and earnings differences: The case of rural China. Economic Development and Cultural Change, 47(4), 767–782. Zhu, Y. (1998). Formal and informal urbanisation in china  – Trends in Fujian Province. Third World Planning Review, 20(3), 267–284. Zhu, Y. (2003). The floating population’s household strategy and migration’s role in China’s regional development and integration. International Journal of Population Geography, 9(6), 485–502. Zhu, Y. (2007). China’s floating population and their settlement intention in the cities: Beyond the hukou reform. Habitat International, 31(1), 65–76. Zhu, Y., & Chen, W. (2010). The settlement intention of China’s floating population in the cities: Recent changes and multifaceted individual-level determinants. Population, Space and Place, 16(4), 253–267. Zhu, Y., Ding, J., Wang, G., Shen, J., Lin, L., & Ke, W. (2016). Population geography in China since the 1980s: Forging the links between population studies and human geography. Journal of Geographical Sciences, 26(8), 1133–1157.

Chapter 5

Internal Migration in Mongolia Solongo Algaa

5.1  Introduction Mongolia is a landlocked country located between the Russian Federation and the People’s Republic of China. It is the nineteenth largest country in the world by territory size (1,565,100 km2) but 138th according to population (3.1 million). As a consequence, it has the lowest population density (1.9 persons per km2) of any country in the world. The population is unevenly distributed across its vast territory, with 46% living in the capital city of Ulaanbaatar in 2015 (NSO 2016). This creates a dual national character: a nomadic pastoral society on the one hand, and a highly urbanised society on the other. The population is young by global standards with a median age of 27.1 compared to the global average of 29.6 in 2015 (UN 2019). Fertility remains above replacement, with a TFR of 2.83 (2010–2015) although this has declined substantially since the 1990s. The population is growing at 2.2% per year (NSO 2016). The pattern and intensity of internal migration in Mongolia have changed radically over the course of the twenty-first century. Before the revolution of 1921, Mongolia was a nomadic pastoral society, with herders roaming seasonally to access better pasture, and goods and services from nearby settlements. The rural population began to migrate to urban areas due to the collectivisation of the pastoral sector through the formation of the so-called ‘Negdel’ or agricultural collectives. The shift to a centrally planned economy following the 1921 revolution led to the development of new towns and industries, which accelerated in the 1950s. Migration and settlement were under strict government control during this time and migrants were directed towards urban centres to meet the growing demand for labour. As a consequence, the urban population increased threefold between 1956 and 1969, while the rural population grew by just 10%. By the mid-1970s the urban population had

S. Algaa (*) School of Arts and Science, National University of Mongolia, Ulaanbaatar, Mongolia e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_5

77

78

S. Algaa

surpassed the rural population. The transition away from a centrally controlled economy in the early 1990s triggered a collapse in the industrial sector and closure of agriculture collectives. At the same time, the adoption of the new Constitution in 1992 removed controls on migration, allowing citizens to freely choose their settlement area. The privatisation of the pastoral sector in the 1990s triggered migration from cities to the countryside. At the same time, regional economic disparities accelerated in-migration to central aimags and cities, especially to the capital city of Ulaanbaatar. In the first decades of the twenty-first century Mongolia experienced rapid economic growth due to a mining boom. In 2011, GDP grew by 17.3%, the fastest in the world (World Bank 2019). This has triggered a new wave of migration to cities and resource-rich areas. This chapter describes patterns of internal migration in Mongolia using data from the 1989, 2000 and 2010 censuses and the Mongolian Civil Registration system. The chapter begins with a review of the internal migration data available in Mongolia, followed by a description of the spatial framework across which migration is measured. Following a review of prior research, the analysis then examines migration intensity, the characteristics of migrants and the spatial patterns of population movement. The chapter concludes with a discussion aimed at understanding the dynamics and causes underlying internal migration in Mongolia and explores its impacts and implications.

5.2  Internal Migration Data There are two primary sources of internal migration data in Mongolia: the Population and Housing Census and the Civil Registration System. The Mongolian Population and Housing Census is conducted decennially with a total of ten censuses since 1918. The 2010 Census collected information on both lifetime and recent migration including place of usual residence, aimag (province) of birth, aimag of residence in the last five years and 11  months before the census, and duration of stay in the aimag of residence. Migration data can be cross-tabulated with other characteristics captured by the census, including age, sex, marital status, economic status, education and ethnicity, but no information is collected on the reason for move. The Civil Registration System commenced in 1951 and captures changes in residential address along with information such as marital status, education and employment. A change in residence is recorded at the bagh/khoroo level, the smallest administrative unit in Mongolia (see section three), before being confirmed by the soum or district’s civil registration system. Monthly statistics are consolidated and delivered by the aimag and provincial municipality statistical offices to the National Statistical Office, which produces annual reports. Since 1987, information from the census has been linked to the registration system. In 2015, the National Statistical Office (NSO) mounted a register-based census from the Civil Registration System and linked databases held by various state agencies to the Population and Household Database (PHD).

5  Internal Migration in Mongolia

79

In addition to census and register-based statistics, there have been 12 surveys since 2000 on the characteristics, causes and consequences of internal migration. Of these studies, six were implemented at national or regional level, whereas the rest were confined to specific areas. These surveys have explored the characteristics of migration, factors affecting migration and the consequences of migration, and have produced a range of policy recommendations.

5.3  The Spatial Framework Mongolia is a vast country, with a strongly primate settlement distribution. The capital city, Ulaanbataar, has a population of around 1.3 million people and is home to 45% of the nation’s people. The next two most populous cities, Erdenet and Darhan, have less than a tenth of the population of the capital, with fewer than 80,000 residents each. Mongolia’s administrative geography is comprised of a four-­ tier hierarchy. The first tier consists of five regions: Western, Khangai, Central and Eastern regions and the Capital city of Ulaanbaatar (Fig. 5.1). The second tier consists of 21 provinces or aimags and one provincial municipality (Ulaanbaatar). The third-tier geography is comprised of soums and districts. Aimags are divided into 331 soums, while Ulaanbaatar is divided into nine districts. The fourth-tier geography is comprised of 1600 baghs (outside Ulaanbaatar) and 152 khoroos (within Ulaanbaatar). As noted earlier, internal migration data collected at the census refer to movements between aimags, while migration statistics from the registration system are collated for soums and districts, but disseminated at the aimag level.

Fig. 5.1  Regions and aimags of Mongolia, 2010

80

S. Algaa

5.4  Prior Research Internal migration is transforming the social, political and economic structure of Mongolia, most notably through its links to the intensive sedentarisation of Mongolian herders. There is limited scholarly work focussing on the patterns and processes of internal migration in Mongolia. However, there is a rich literature exploring the shifting livelihoods of Mongolian herders over the past century, of which migration into cities such as Ulaanbaatar has been a significant part (see, e.g. Bruun and Narangoa 2006). The impacts of government policy and climate change on internal migration have also been explored (Mayer 2016; Algaa 2012, 2017). While scholarly papers (in English) focussing on internal migration are few, there has been a concerted effort to understand internal migration in Mongolia through the deployment of multiple large scale surveys. For example, in 2000, the Population Teaching and Research Center (PTRC) at the National University of Mongolia conducted a Micro Study of Internal Migration (PTRC, MoSWL and UNFPA 2001). The survey aimed to determine the causes of migration, access of migrants to essential social services, challenges faced by migrants, strategies used to overcome hardships and the impact of migrants on destination areas. A survey on Internal Migration Dynamics and Its Consequences was conducted from July 2008 to February 2009 (PTRC, MoSWL and UNFPA 2009). The objectives included study of internal migration flows, trends, causes and consequences for social and economic development. It also sought to inform the criteria for Millenium Development Goal objectives related to migration and to develop policy recommendations for population and regional development. Aimags and urban areas of Mongolia were classified into three areas, namely areas of origin, destination areas and ‘stepping stone’ areas. The survey included participatory methods, standardised questionnaires and a document review. The survey covered 640 potential migrant households in areas of origin, 650 households in destination areas, 32 groups of managers and 364 experts. A qualitative study of internal migration in Mongolia was conducted in 2007 by the National Statistical Office with the technical and financial support of the UNFPA  (NSO and UNFPA 2007). The study aimed to understand the migration experience of individuals using in-depth interviews (n = 142) and focus group discussions (n = 11). Based on the findings of this study, recommendations were made to improve the migration registration system. These included more open data and changes to the registration institutions. The National Statistical Office has published two monographs on internal migration based on the 2000 and 2010 censuses. The monograph based on 2000 census data aimed to collect basic data for policy development and resolution of problems on migration, location of the population and urbanisation, however no policy recommendations were developed (NSO 2002). A second monograph based on the results of the 2010 census explored major patterns and trends in internal migration during the previous decade and its impacts on aimags and on Ulaanbaatar (NSO 2011). This monograph also proposed a range of policies on current and future migration, distribution and settlement.

5  Internal Migration in Mongolia

81

5.5  How Much Movement? Migration Intensity In the five years prior to the 2010 Census, 295,924 Mongolians or 12.7% of the population changed their aimag or provincial municipality of residence (Table 5.1). This inter-provincial movement has increased since the 2000 Census, which recorded 174,599 recent movers, equivalent to 8.2% of the population. Lifetime inter-provincial migration at the 2010 census was 32.6%, up from 20.5% in 2000 and 21.9% in 1989. The decline in lifetime migration between the 1989 and 2000 censuses reflects the shift to a market-driven economy following the collapse of the Soviet Union along with the removal of controls on migration. Many Mongolians returned to the countryside following the collapse of the pastoral collectives and privatisation of that sector, reducing lifetime intensities. Return migration remains significant, with data from the 2010 census indicating that, while one-third of Mongolians lived outside their place of birth, 8.3% had returned to their birthplace over the previous five years. The propensity to migrate within Mongolia has changed appreciably over the last three decades. Figure 5.2 shows the crude migration intensity (CMI) for movement Table 5.1  Crude migration intensities by type of move, Mongolia, 1989, 2000 and 2010 Type of migration Five years Lifetime

Population census 1989 N/a 21.9

2000 8.2 20.5

2010 12.7 32.6

Fig. 5.2  Crude migration intensities, movement between aimags, Mongolia, 1990–2018. (Note: Intensities are based on in-migration to aimags. Source: NSO 2018. www.1212.mn updated to 30 March, 2018)

82

S. Algaa

between aimags for the period 1990 to 2018. The inter-provincial CMI averaged 2% over this period but has declined over time. The CMI peaked in 1991 when 3.8% of the population changed their province of residence. This increase was triggered by the removal of controls on migration, the privatisation of livestock and dissolution of cooperatives following the collapse of the Soviet system. A secondary peak in the CMI occurred in 2010, underpinned by substantial rural-to-urban migration. This was a response to the devasting dzud in 2009–2010: an extremely snowy and cold winter that resulted in extensive loss of livestock. A smaller peak in the CMI was observed in 2004. This coincided with the waiver of registration fees and taxation previously imposed on migrants moving to Ulaanbaatar, resulting in the registration of formerly undocumented migrants. In 2017 restrictions on migration towards Ulaanbataar were reintroduced by decree of the Mayor’s Office. These are expected to remain in place until 2020. In response to these controls, the CMI in 2018 was just 1%, the lowest observed in three decades.

5.6  Who Moves? The Characteristics of Migrants Migration is a highly age-selective process. Around the globe, migration typically peaks in young adulthood before declining into middle age. Figure 5.3 shows the age profile of migration for males and females based on people who had spent less than a year at their current residence at the 2010 Census. For females, migration peaks between the ages of 15–19 before declining. Migration among males peaks later, between the ages of 20 and 24 and is less concentrated at the peak. Migration among males remains relatively high into middle age. Table 5.2 reports CMIs for one year, five years and lifetime by sex. The lifetime migration rate was higher for females than males. However, males are more migratory than females in the category of ‘return migration’. Five-year and one-year migration rates were higher for males than females. In general, males appear to be slightly more mobile than their female counterparts. There is a clear sex differential in migration patterns. Males are more likely to migrate to Dornogobi and Umnugobi aimags, where significant resource extraction projects are located, while females are more likely to migrate to other provinces and to Ulaanbaatar. According to a 2000 migration study (PTRC, MoSWL and UNFPA 2001), young females aged 15–19 are the most likely to migrate to Ulaanbaatar. This remained the case in 2010, and is likely associated with the continuation of education at tertiary and professional technical institutions in Ulaanbaatar. Males aged 20–24 are the next most likely group to migrate to Ulaanbaatar, reflecting a combination of employment and education motives. Based on data from the 2010 Census, the majority of migrants were found to be young and less educated. For all levels of education, males are more likely to be involved in migration than females. Young unmarried people are the most likely to migrate to Ulaanbaatar. Sex differences show that young females who have not yet married (9.8%) are more likely to migrate to Ulaanbaatar than young unmarried males (9.2%).

5  Internal Migration in Mongolia

83

Fig. 5.3  Age-specific migration intensities by sex, movement between aimags, 2009–2010. (Source: NSO 2011)

Table 5.2  Number of migrants and crude migration intensities by sex and type of move, Mongolia, 2010 Mobility status Lifetime migrants Of which return migrants Five-year migrants One-year migrants

Males Number 422,076 142,646 160,726 60,442

CMI 32.1 10.9 14.0 4.6

Females Number 455,486 77,385 157,196 53,343

CMI 34.2 5.8 13.4 4.0

Total Number 877,562 220,031 317,922 113,786

CMI 33.1 8.3 13.7 4.3

Sex ratio 93 184 102 113

Source: NSO 2011. Author’s calculations Note: CMI Crude Migration Intensity

5.7  Where Do They Move? Spatial Patterns The impact of internal migration on the geographical distribution of national populations reflects the intensity of migration, captured by the CMI and the degree to which counter-flows offset flows, captured by the migration effectiveness index (MEI). The MEI ranges from 0 to 100; a high score implies that internal migration is an efficient mechanism of population redistribution in the country. The Aggregate Net Migration Rate (ANMR), which is the product of the CMI and MEI, divided by 100 quantifies the degree to which migration redistributes population across geographic units in a country. As shown in Table  5.3, fully 5.5% of the Mongolian population was redistributed between aimags in the five years before the 2010

84

S. Algaa

Table 5.3  System-wide migration indicators by type of move, Mongolia, 1989, 2000 and 2010

Migration effectiveness index Aggregate net migration rate

4.5 4.0

CMI/ANMR (%)

3.5

Type of migration Five years Lifetime Five years Lifetime Five years Lifetime

Population census 1989 2000 N/a 8.2 21.9 20.5 N/a 49.6 65.5 67.8 N/a 4.1 14.4 13.9

ANMR CMI MEI

2010 12.7 32.6 43.8 70.9 5.5 23.1

90 80 70

3.0

60

2.5

50

2.0

40

1.5

30

1.0

20

0.5

10

MEI (%)

Migration measure Crude migration intensity

0.0 0 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 Year

Fig. 5.4  System-wide migration indicators by type of move, Mongolia, 1990 to 2018. (Source: Calculated from National Statistical Office register data)

Census. This increased from 4.1% in the five years preceding the 2000 Census. This high impact is a function of both high intensities and high levels of migration effectiveness, with 44 people redistributed for every 100 who changed their aimag of residence in 2010 and 50 people redistributed for every 100 who changed their aimag of residence in the five years to 2000. Lifetime migration impact, as captured by the ANMR, has also increased over time. At the 2010 Census, 23.1% of Mongolians had been redistributed between aimags since birth, up from 13.9% at the 2000 Census. The growing impact of lifetime migration reflects increases in intensity but also extremely high lifetime effectiveness, with MEIs of 71 in 2010 and 68 in 2000. These metrics show the substantial impact of internal migration on the settlement pattern of Mongolia in the twenty-first century. Register data provide insights into the changes in migration impact since the 1990s. Figure  5.4 reveals dramatic changes in migration effectiveness over time which, coupled with fluctuations in migration intensity, has led to significant

5  Internal Migration in Mongolia

85

volatility in migration impact over this period. In 1990, 0.32% of the population was redistributed between aimags. This increased to 1.02% in the following year, due to the dramatic increase in migration intensity described earlier. Critically the MEI in 1991 was relatively low at 27.1%, reducing the overall impact. Migration effectiveness continued to increase sharply until 2003 when it peaked at 80.1%. This degree of asymmetry is extraordinary and was driven by flows into Ulaanbaatar. High effectiveness, coupled with a peak in intensity meant that 1.65% of the population was redistributed between aimags in 2004. As noted earlier, this is partly an artefact of delayed registration in response to the scrapping of registration fees for in-­ migrants to Ulaanbaatar. After 2004 the migration system became progressively more balanced and, by 2018, the MEI was just 2%. This dramatic reduction can be attributed to the re-introduction of controls on migration into Ulaanbaatar. Figure 5.5 shows the system of bilateral migration flows between aimags in the five years preceding the 2010 Census. The colour of the chords denotes the origin

Fig. 5.5  Bilateral migration flows, aimags of Mongolia, 2005–2010. (Source: Calculated from the 2010 Population Census of Mongolia (IMAGE-Asia Project))

86

NMR

S. Algaa

< -10.0 -9.9 - -5.0 -4.9 - 0.0 0.1 - 5.0 >5.1 - 13.6

300 150

0

300

Fig. 5.6  Net migration rates, aimags of Mongolia, 2005–2010. (Source: Calculated from the 2010 Population Census of Mongolia (IMAGE-Asia Project))

of flows, while their size reflects the volume. Ulaanbaatar dominates the system, capturing a third of gross flows, and a quarter of all in-flows. The largest outflows are from Western and the Khangai regions to Ulaanbaatar and aimags in the Central region. The pattern of lifetime migration between regions based on the 2010 census is similar to recent moves (not shown), again dominated by flows into Ulaanbaatar. The net result of migration flows in the five years prior to the 2010 Census is depicted in Fig. 5.6. Of Mongolia’s 21 aimags and Ulaanbataar, only six recorded net migration gains during this period. The largest gains occurred in Ulaanbaatar, which recorded a net migration gain of 13.6% in the five years to 2010. The second largest gain was recorded in Omnogovi, a resource-rich region in the Gobi desert. Other regions recording gains include Govisumber (5.1%), Orkhon (2.4%), Dornogovi (0.4%) and Darkhan-Uul (0.1%). Net losses were highest in Dundgovi (21.7%), Zavkhan (18.5%) and Uvs (18.1%). Data from the population register suggest that there have been dramatic shifts in the pattern of inter-regional flows since the 2010 Census. While a matrix of origin-­ destination flows is unavailable, it is possible to map net migration rates based on register data. Figure  5.7 shows the pattern of gains and losses across aimags in 2018. There has been a reversal of the pattern observed at the 2010 Census, likely an outcome of the restriction on internal migration to Ulaanbaatar. Net migration gains are recorded in regions in the Eastern and Western Regions, while aimags in the Central regions recorded losses. To explore the impact of internal migration on the progress of urbanisation, we examine changes in the direction of flows between urban and rural areas. Comparison of migration between rural and urban areas over time is commonly inhibited by changes in definition, and this is also true in Mongolia. Here we adopt the approach used by Rees et al. (2017), which employs population density as a proxy for the

5  Internal Migration in Mongolia

87

NMR -0.04 - -0.02 -0.01 - 0.00 0.01 - 0.05 0.06 - 0.10

300 150

0.11 - 0.20

0

300 Kilometres

Fig. 5.7  Net migration rates, aimags of Mongolia, 2018. (Source: Calculated from National Statistical Office register data (IMAGE-Asia Project)) 20.00

Ulaanbaatar

15.00

Net Migration Rate

10.00 5.00 0.00 -5.00 -10.00 -15.00 -20.00 -25.00 0.00

Omnogovi Govisumber

Darkhan-Uul Dornogovi Selenge Tov Khovsgol Bayan-Olgii

Orkhon

Khentii Bulgan Bayankhongor Arkhangai Khovd Dornod Sukhbaatar Uvs Govi-Altai Zavkhan Dundgovi Ovorkhangai 20.00

40.00

60.00

80.00 100.00 120.00 140.00 160.00 180.00 200.00 Log of population density

Fig. 5.8  Net migration rates by log population density, aimags of Mongolia, 2005–2010. (Source: Calculated from the 2010 Population Census of Mongolia (IMAGE-Asia Project))

conventional rural/urban dichotomy. To this end, we calculate net migration rates for each of the provinces and regress these against the logarithm of their population density. The relationship between population density and net migration rates is captured by the slope parameter of a population-weighted linear regression. A positive slope indicates that migration is predominately from less densely populated to more densely populated regions (rural to urban), while a negative slope indicates that migration is predominately urban to rural. The magnitude of the slope, denoted by

88

S. Algaa

the regression coefficient, indicates the strength of the process. Figure 5.8 shows a strong positive slope (9.6) and reveals a system dominated by flows to Ulaanbaatar. Orkhon, the aimag containing Mongolia’s second-largest city, also experienced significant net migration gains.

5.8  Understanding Internal Migration in Mongolia Internal migration in the twentieth and twenty-first century has had a significant impact on the population of Mongolia, leading to a primate settlement distribution in which nearly half of the population live in Ulaanbaatar. Unlike many other countries where the process of urbanisation involved migration from small villages and towns to larger cities, internal migration in Mongolia emerged alongside the sedentrisation of the nomadic population, many of whom settled in the capital city. There has been significant volatility in the intensity and impact of internal migration over recent decades. This reflects economic and environmental factors as well as the impact of policies designed to limit settlement in Ulaanbaater. In rural areas, pastoral herding remains the main source of income, and there are few other income-­generating opportunities. Almost 70% of herders are considered poor, and except for a small number of civil servants, most people lack the job opportunities needed to provide a stable income. This lack of opportunity in rural areas has triggered significant flows to urban areas. Another factor triggering migration is the lack of education, health and other social services in rural areas. Access to quality education for their children is an important motivator for many migrants, as is access to a range of consumer and social services absent at their rural origins. Rural-urban disparities in employment and other services have deepened in recent times, contributing to the migration from rural to urban areas. Natural disasters have resulted in some herders abandoning nomadic (or rotational) pastoralism completely and many have migrated to urban areas in search of alternative livelihoods. Climate change is already starting to negatively impact rural communities. During the last dzud in 2009–2010, almost 20% of the national herd was lost. Around one third of herders lost at least half their livestock, depriving them of their major source of income. Loss of livestock was one of the push factors from remote rural to urban areas for herders. Indeed, an analysis of the relationship between rural out-migration (per 1000 population) and lost livestock (percent of total livestock) revealed a positive association. When the livestock loss increased by one percentage point, rural out-migration per 1000 population increased by 0.23 persons (p = 0.004) (Algaa 2017). Government policy aimed at limiting new migration into Ulaanbaatar has had a significant impact on internal migration flows. In-migration to Ulaanbaatar increased dramatically following the dzuds in 1999–2001. As a result, the Capital City Citizens’ Representative Khural issued a new resolution (No. 47 of year 2000) to double the in-migration registration fee to 50,000 MNT per person. In response to the National Committee for Human Rights’s petition filed in 2003 to the Supreme Court of Mongolia, the registration fee levied on in-migrants was removed in 2004.

5  Internal Migration in Mongolia

89

This resulted in an annual population growth rate of more than 6% in the main ger districts, the low-density peri-urban area around Ulaanbaatar. The most recent retrictions, introduced in 2017, have coincided with a significant drop in intensity and effectiveness, suggesting that these are having some impact.

5.9  Impacts and Implications Internal migration has exerted significant impacts at destinations and on the migrants themselves. Perhaps the most striking have been the impacts on Ulaanbaatar. Large inflows into the capital have resulted in great poverty and a range of environmental problems. The most obvious manifestation is the vast low-density peri-urban area (the so-called ger districts) that accommodates about 60% of Ulaanbaatar city’s population (about 840,000 inhabitants). Almost 56% of the population of the ger districts live in stand-alone houses, while 46% live in traditional yurts. The ger districts have expanded through successive waves of migration. Since 2010 the estimated average population growth in ger areas was 25,000 inhabitants per year. This represents more than half of the city’s 3.8% annual growth rate. These low- and medium-income household settlements (the average monthly household income in ger areas is $380) are characterised by loosely aligned plots, creating irregular pathways that remain unpaved. While the majority of households have land tenure, residents lack access to basic urban services. This has been exacerbated by rapid population growth, overwhelmining the capacity of schools, kindergartens and hospitals. Access to public transport is another issue, with many remote ger districts lacking public transportation. One outcome of growth in urban districts has been the ‘urbanisation of poverty’ as the urban poverty rate became higher than that of the countryside. Expansion of the ger districts has also resulted in significant environmental impacts arising from increased waste, open latrines, local deforestation, and air pollution due to the burning of coal and refuse for heat and cooking. To ensure the right to life in a healthy and safe living environment, a decree by the Mayor of Ulaanbaatar city was issued in 2017 with the effect of ‘Temporary discontinuation of movements for permanent residence from rural areas into Ulaanbaatar city with an exception to citizens who require access to health services or have purchased residential apartments’. While this appears to have slowed migration, it has presented a variety of challenges for unregistered migrants. Impacts on the migrants are mixed. Migration brought them closer to health and education services. Migrants also benefit from being close to entertainment, being able to learn about urban culture and a settled lifestyle. Surveys of migrant outcomes have shown both benefits and costs. In one study, migrants reported a better quality of education for their children compared to the countryside as well as moral satisfaction from being with children, decreased family expenses, and an opportunity to obtain specialised medical care (PTRC, MoSWL and UNFPA 2001; Batbaatar et  al. 2005). Getting a job and increased income were considered the greatest advantages of migration at the individual and household level. More than half of migrants in one survey (PTRC, MoSWL and UNFPA 2009) reported a lack

90

S. Algaa

of improvement in access to health care, living conditions, environment, satisfaction with their lives, purchasing capacity, relations with family, relatives, friends, active position in life and social participation following migration. Outcomes were dependent upon finding a job, access to public and social welfare services due to inadequate documentation, inability to find land to settle on, remoteness from infrastructure and, as a result, isolation from society (PTRC, MoSWL and UNFPA 2009; IOM 2018).

5.10  Conclusions Internal migration in the twentieth and twenty-first century has had a dramatic impact on the settlement system of Mongolia, producing a strongly primate settlement system and transforming livelihoods away from nomadic pastoralism to urban pursuits. The impact on the settlement system is the largest observed within Asia, reflecting both high migration intensities and high levels of effectiveness, with flows into Ulaanbaatar dominating the national system. The intensity and pattern of internal migration reflect dramatic transformations in the political and economic system of Mongolia as well as environmental factors. Collectivisation in the Soviet era triggered an initial process of urbanisation, later followed by counter-urban flows with the shift to a market economy in the 1990s. Since this time, periodic dzuds, regional inequalities and the rapid growth of the economy tied to resource extraction have driven successive waves of migration. The impacts of migration are most profound in the ger districts of Ulaanbaatar. Unlike informal settlements in many other parts of the world, residents of ger districts are generally granted land tenure; however, access to public services and infrastructure is lagging. For the migrants themselves, difficulties in securing work and accessing services have led to some dissatisfaction, particularly acute in the case of undocumented migrants. The scale of in-migration to Ulaanbaatar presents significant planning and policy challenges. Restrictions on migration are one response, but there is a need for policies to support balanced development of regions, reduce migration flows, decrease overconcentration, and ameliorate the negative consequences of migration. Reducing vulnerability and income disparities are critical developmental challenges in Mongolia. Large segments of Mongolia’s population remain vulnerable with insecure livelihoods. Supporting rural livelihoods and income diversification are potential ways to reverse this situation by reducing vulnerability and keeping people on the land.

5  Internal Migration in Mongolia

91

References Algaa, S. (2012, August 26–29). Internal migration consequences and its changes. Second conference of Asian Population Association. Thailand, Bangkok. Algaa, S. (2017). Relationship between rural to urban migration and Dzud disaster impact. Mongolian Population Journal. Scientific Paper of National University of Mongolia (470)27, 78–84. ISSN:2226-1389. Ulaanbaatar: NUM Press. Batbaatar, M., Bold, T., Marshall, J., Oyuntsetseg, D., Tamir, C., & Tumennast, G. (2005). CHIP Report 17: Children on the move: Rural-urban migration and access to education in Mongolia. Childhood Poverty Research and Policy Centre (CHIP). London, UK. Bruun, O., & Narangoa, L. (2006). Mongols from country to city: Floating boundaries, pastoralism and city life in the Mongol lands. Copenhagen: NIAS Press. IOM. (2018). Mongolia: Internal migration study (Survey report). Ulaanbaatar: International Organization for Migration. Mayer, B. (2016). Climate migration and the politics of causal attribution: A case study in Mongolia. Migration and Development, 5(2), 234–253. https://doi.org/10.1080/21632324.201 5.1022971. NSO. (2002). Migration and urbanization: 2000 population and housing census. Ulaanbaatar: National Statistical Office of Mongolia. NSO. (2011). Internal migration and settlements in Mongolia (2010 Population and housing census monograph). Ulaanbaatar: National Statistical Office of Mongolia. NSO. (2016). 2015 population and housing by census of Mongolia (National Report). Ulaanbaatar: National Statistical Office of Mongolia. NSO & UNFPA. (2007). Report of a qualitative study of internal migration in Mongolia (National Statistical Office of Mongolia). Ulaanbaatar: United Nations Population Fund. PTRC, MoSWL, & UNFPA. (2001). A micro study of internal migration 2000 (Survey report). Ulaanbaatar: Population Teaching and Research Center, Ministry of Social Welfare and Labor and United Nations Population Fund. PTRC, MoSWL, & UNFPA. (2009). Mongolia: Internal migration dynamics and its consequences (Survey report). Ulaanbaatar: Population Teaching and Research Center, Ministry of Social Welfare and Labor and United Nations Population Fund. Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), 1–22. https://doi.org/10.1002/ psp.2036. Save the Children UK. (2006). Children on the move: Rural-urban migration and children’s access education (Survey report). Ulaanbaatar. United Nations (2019). World population prospects. New  York: United Nations Development Programme. World Bank. (2019). GDP growth, Mongolia, World Bank National Accounts Data, and OECD National Accounts Data Files. Washington, DC: The World Bank https://data. worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?contextual=max&locations=MN&m ost_recent_value_desc=true.

Chapter 6

Internal Migration in South Korea Yeonjin Lee and Doo-Sub Kim

6.1  Introduction This chapter explores the levels, composition and patterns of internal migration in the Republic of Korea, hereafter termed South Korea. Situated in East Asia, South Korea is one of the most mobile countries in the world with more than half the population changing their address in the five years to 2006 (Bell et  al. 2015). South Korea is among the most economically advanced countries in Asia, with the eleventh highest GDP per capita. Internal migration in Korea in the twentieth century has been driven by colonisation, displacement due to war, and rapid industrialisation. Indeed, since the Korean War, the economy has undergone dramatic restructuring, transforming from an agrarian society to an advanced industrial and service-based economy. The early stages of this economic transformation triggered massive population redistribution from rural to urban areas, focussed on the cities of Seoul, Busan and Daegu. A migration turnaround occurred in the 1990s–2000s, with internal migrants leaving metropolitan regions to settle in suburban regions, satellite cities and new cities such as Sejong. Economic modernisation in the post-­ war period was accompanied by a rapid demographic transition, with South Korea now recording among the highest life expectancies and lowest fertility rates in the world. International migration to Korea is also low by global standards, with just 3.2% of the population born overseas (Statistics Korea 2019a). This has led to rapid ageing of the population, particularly in rural areas, and a shrinking pool of young people available to move from rural to urban areas. The extreme spatial Y. Lee Department of Social Work and Social Administration, and School of Public Health, University of Hong Kong, Hong Kong, China e-mail: [email protected] D. S. Kim (*) Department of Sociology, Hanyang University, Seoul, South Korea e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_6

93

94

Y. Lee and D. S. Kim

concentration and rapid industrialisation of the Korean economy has created a unique social context in which internal migration has occurred.

6.2  Internal Migration Data South Korea is well served with data on internal migration. There are two sources of internal migration data in Korea: the Korean Census of Population and Housing, and the Korean Population Register. The Census has been conducted on a quinquennial basis since 1925, with the last traditional Census conducted in 2010. The 2010 Census collected information on respondents’ places of residence five years prior to the Census, and it included rich detail on the characteristics of migrants. Since 2015 census enumeration has been based on administrative datasets, supplemented by a sample survey that assembles basic household and individual information including internal migration data. However, the Korean Population Register has now replaced the Census as the principal source of internal migration statistics. Annual migration flows are estimated based on registered changes of residence in the Korean Population Register. These data include annual estimates of the total number of internal migrants (i.e., all changes of address), as well as in- and out- migration for 229 municipal units—Gu (districts), Si (cities), Gun (counties)—and 17 metropolitan cities and Do (provinces). These are explained more fully in the next section. Data are available back to 1971, although they are not strictly comparable due to changes in the number of spatial units. Data from the Korean Population Register at the province level, along with data on all moves, are used in this chapter. Data used in this chapter have been extracted from the KOSIS Internal Migration Statistics Database.1

6.3  The Spatial Framework The Korean administrative geography consists of a three-tier hierarchy. The first tier is composed of 17 regions of which eight are metropolitan cities (Si) and nine are provinces (Do) (Fig. 6.1). The eight metropolitan cities include the capital Seoul, Sejong (a special self-governing city), Busan, Daegu, Incheon, Gwangju, Daejeon, and Ulsan. The nine provinces are Gyeonggi-do, Gangwon-do, Chungcheongbuk-do, Chungcheongnam-do, Jeollabuk-do, Jeollanam-do, Gyeongsangbuk-do, Gyeongsangnam-do, and Jeju (a special self-governing province). Annual internal migration statistics, including origin-destination matrices, are produced at the

 http://kosis.kr/eng/statisticsList/statisticsListIndex.do?menuId=M_01_01&vwcd=MT_ ETITLE&pa rmTabId=M_01_01

1

6  Internal Migration in South Korea

95

Fig. 6.1  Provinces and districts of South Korea, 2017

first-­tier geography back to 1971. The second tier comprise municipal divisions of which there were 229 in 2018. The metropolitan cities are composed of Gu (districts), while provinces are composed of Si (cities) and Gun (counties). The third tier consists of sub-municipal divisions. Si are composed of Dong (neighbourhoods) while Gun are composed of Eup, Myun (towns) and Li (villages), with Eup and Myun located in rural areas. Annual migration statistics are not produced at the sub-­ municipal level, however, moves within municipalities are reported.

96

Y. Lee and D. S. Kim

6.4  Prior Research In comparison with other developed countries, there have been relatively few studies on internal migration in South Korea. Most empirical analyses of internal migration have been based on census data. Some studies in the 1970s dealt with regional migration by sex and age based on the census survival ratio method (Kwon et al. 1975; Yu 1980). Since the early 1980s, various studies utilising census data on the place of residence one year or five years ago have examined the volume and characteristics of internal migration, particularly rural-to-urban migration and migration toward the capital region (Kim and Lee 1976; Kim and Sloboda 1981; Choi and Chang 2004; Kim 2017). Several studies were conducted after specific censuses and published as a part of the census monograph series (Choe et al. 1989; Choi and Choi 1993; Kim et al. 1997). With increased availability of survey data on migration since the early 1980s, the determinants, selectivity and consequences of migration at the micro-level have received increasing attention from researchers and policymakers (Kim 1982; Choe et  al. 1989; Choi and Choi 1993; Kwon and Lee 1997; Lee 2014). Many of the demographic studies on internal migration and urbanisation have also paid attention to over-urbanisation, the importance of rural-to-urban migration in the urbanisation process, and inter-city inequalities and unequal development (Yu 1980; Kim 1988; Moon 1993; Kang 1993; Kwon and Kim 2002; Choi and Chang 2004; Kim 2004). From the end of the Second World War until the mid-2000s, rural-to-urban migration was the dominant migration process in South Korea. This was reflected in increased urbanisation of the population from 28.0% in 1960 to 81.6% in 2015 (Kim 2018; UN 2018). Drivers and patterns of rural-to-urban migration changed over this period. In the 1960s, rural-to-urban migration was largely driven by push factors in rural areas, arising from adjustments in the rural economy, demographic growth and a surplus of rural labour. Later, following a period of rapid industrialisation, demand for labour in urban areas became the major pull factor. A key feature of rural-to-urban migration at this time was the focus of migrant flows towards the largest metropolitan regions, at the expense of small and medium-sized cities (Kwon and Kim 2002; Statistics Korea 2019a). This resulted in ‘truncated’ urbanisation whereby ‘massive permanent village-to-city migration was not preceded or accompanied by short-term circulation between urban places of varying sizes’ (Mobrand 2012 p.  389). The population of Seoul grew from one million in 1950 to reach 2.4 million in 1960, and 5.4 million in 1970. Busan grew from 0.9 million in 1950 to 1.9 million in 1970, while Daegu grew from 0.3 million to 1.1 million over the same period. Rural-to-urban migration is estimated to account for around three-­ quarters of urban growth during this period (Kwon and Kim 2002). An outcome of this extreme spatial concentration of rural-to-urban flows was the primacy or megalopolitanisation of the Korean settlement system. Internal migration rates increased throughout the 1970s and became increasingly concentrated on the two urban axes of the country, Seoul and Busan, (Kim 2017, 2018; Statistics Korea 2019a). By contrast, cities in the provinces of Honam and

6  Internal Migration in South Korea

97

Gangwon, generally excluded from the benefits of industrialisation, continued to record low rates of internal migration. By the beginning of the 1980s, national rates of internal migration began to slow due to the falling size of rural populations. Urban-to-urban migration emerged as an important stream, leading to growth in intermediate cities. In Seoul, migration flows towards the metropolitan core also slowed as the economic links between Seoul and neighbouring cities increased and suburbanisation began. This led to rapid growth in the regions immediately surrounding Seoul. Population growth in the core of the Seoul metropolitan region turned negative in the 1990s, with the population decreasing by 709,000 between 1990 and 2015 (Kim 2018). This was despite continued demand for industrial labour. Localities close to Seoul, such as Sungnam, Bucheon and Anyang, became increasingly attractive destinations for migrants due to lower housing and living costs, leading to strong suburbanisation.

6.5  How Much Movement? Migration Intensity South Korea is among the most mobile countries in the world. In 2018, 14.1% of the Korean population changed their place of residence as measured by the Aggregate Crude Migration Intensity (ACMI). Of these moves, 5.5% were within the same municipal area (Si/Gun/Gu), 4.1% were between municipal areas within the same province, while 4.8% were between provinces. In terms of absolute flows, 2.4 million people changed province of residence, another 2.1  million people changed municipality within provinces and 2.8  million changed residence within their municipality. While mobility in Korea remains high by international standards, the ACMI has declined steadily over time, from a peak of 25.5% in 1975, to 21.5% in 1985, 20.1% in 1995 and 18.3% in 2005, as shown in Fig. 6.2. One explanation for the decline is the dramatic ageing of the Korean population. However, the age standardised migration intensity for 2018, based on the 1975 population age structure, was 15.5%, only 10% higher than the value observed in 2015. This suggests that there has been a real drop in migration intensities over time, especially at peak ages. This drop is explored further in the next section. There has been some annual volatility in mobility over this period tied to economic cycles, with a dip recorded following the Asian financial crisis in 1997. It can also be partly attributed to the following factors: (1) the construction and establishment of large industrial cities and housing complexes in the course of socioeconomic development; (2) residential and employment suburbanisation of the major metropolitan cities; and (3) territorial expansion of city boundaries. The extremely high intensity of internal migration in South Korea in the 1970s can be attributed to rapid urbanisation driven by massive and centralised industrialisation. This process attracted large numbers of in-migrants to the metropolitan region of Seoul, as well as to Busan and Daegu. In 1975, when migration intensity was at its peak, inflows to Seoul accounted for 33.2% of all inter-provincial movements (Kwon and Kim 2002). By 2018, this share had declined to 18.9%. While

98

Y. Lee and D. S. Kim

Aggregate crude migration intensity (%)

30.0

25.0

20.0

15.0

10.0

5.0

2018

2016

2014

2012

2010

2008

2006

2004

2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

1970

0.0

Year

Fig. 6.2 Aggregate crude migration intensities, South Korea, 1970–2018. (Source: KOSIS Database on Internal Migration)

inter-provincial migration has been high throughout this period, accounting for around a third of all moves, there has also been significant intra-municipality migration (Kwon and Lee 1997). This suggests that housing, education, job seeking, economic opportunity and other sociocultural factors are also contributing to the high levels of internal migration in South Korea (Yu 1980; Park and Fullerton 1980; Kim 1982; Choi and Chang 2004).

6.6  Who Moves? The Characteristics of Migrants Internal migration is highly selective with respect to age, with migration peaking in young adulthood. As shown in Fig. 6.3, this is also the case for South Korea where migration peaks between the ages of 25 and 29 for both males and females. This is later than what is observed in most Asian countries where migration peaks tend to occur in the early twenties. Migration is also less concentrated around the peak age than in other parts of Asia. Thus, the age profile of internal migration in South Korea is more akin to the late and dispersed patterns observed in Europe and North America (Bernard et al. 2014). Figure 6.3 shows very little in the way of sex differentials in the migration age profile. Migration intensities are slightly higher for females in their early twenties, and lower in later adulthood. This probably reflects differences in the age of key life-course transitions tied to family formation; women are more likely to marry earlier than men and choose older partners (Kwon et al. 1975; Kwon and Kim 2002). Lower levels of migration among young adult men may reflect a later exit from home due to mandatory military service, as those in

6  Internal Migration in South Korea

99

30.0 Female

25.0

Migration intensity (%)

Male 20.0

15.0

10.0

5.0

80+

75-79

70-74

65-69

60-64

55-59

50-54

45-49

40-44

35-39

30-34

25-29

20-24

15-19

10-14

5-9

0-4

0.0

Age

Fig. 6.3  Age-specific migration intensities by sex, all moves, South Korea, 2018. (Source: KOSIS Database on Internal Migration and Lee (2014))

military service do not change their resident registration. Females also demonstrate an increase in moves at very old ages, which can be explained by their residential adjustments related to widowhood and health. The age profile of migration has become less concentrated over time. Figure 6.4 shows age specific migration intensities in 1995 and 2018. In 1995, migration peaked at ages 25 and 29 before declining sharply. In 2018, high intensities were observed among those aged 30–34. By contrast, children and older adults were moving less than in 1995. While it is not possible to explain these shifts directly, the age pattern of migration has been shown to mirror the age structure of key life-­ course transitions (Bernard et al. 2014). A dispersion of migration around the peak suggests a more complex set of life-course transitions tied to higher education, entry into the labour market, and family formation. Stabilisation of mobility in middle age may also reflect the emergence of suburbanisation as an important migration process in South Korea. In the absence of data on reasons for move, it is difficult to reach any more definitive conclusions about the observed shifts in the age profile of migration.

6.7  Where Do They Move? Spatial Patterns The impact of internal migration on the geographical distribution of national populations reflects the intensity of migration, measured by the CMI, and the degree to which flows are offset by counter-flows. The latter is captured by the Migration

100

Y. Lee and D. S. Kim

40.0 1995

35.0

2018

Migration Intensity (%)

30.0 25.0 20.0 15.0 10.0 5.0

80+

75-79

70-74

65-69

60-64

55-59

50-54

45-49

40-44

35-39

30-34

25-29

20-24

15-19

10-14

5-9

0-4

0.0

Age

Fig. 6.4 Age-specific migration intensities, South Korea, 1995 and 2018. (Source: KOSIS Database on Internal Migration) Table 6.1  System-wide migration indicators, provinces of South Korea, 1975, 2000 and 2018 Year 2018 2000 1975

Crude migration intensity (%) 4.63 6.05 7.98

Migration effectiveness index (%) 10.32 7.34 26.29

Aggregate net migration rate (%) 0.48 0.44 2.10

Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project) Note: There is a small discrepancy between the CMIs calculated for provinces based on the origin-­ destination matrices, reported in the table above, and the number of movers identified as inter-­ provincial movers in the South Korean statistical tables

Effectiveness Index (MEI). The MEI ranges from 0 to 100; a high score implies that internal migration is an efficient mechanism for population redistribution in the country. The Aggregate Net Migration Rate, which is the product of the CMI and MEI, divided by 100 quantifies the degree to which migration redistributes population across spatial units in a country. As shown in Table 6.1, 0.48% of the population was redistributed between 17 Korean provinces in 2018. The high migration intensity is offset by low migration effectiveness, with just ten people being redistributed for every 100 who changed their province of usual residence. The impact of internal migration on population distribution has declined markedly over time. When migration intensity was at its historic peak in 1975, 2.1% of the population were redistributed between provinces in a single year. This is the combined outcome of extraordinarily high migration intensity with high migration effectiveness. It should be noted that there were six fewer provinces in 1975 than in 2018, which has the

6  Internal Migration in South Korea

101

effect of lowering the measured CMI. Thus, the true impact of migration on population redistribution was likely even higher than these figures suggest. The observed shift in migration impact accords with the model proposed by Rees et al. (2017), which posited a U-shaped relationship between migration impact and development. In the early stages of industrialisation, such as that observed in Korea in the 1960s and 1970s, large economic differentials between regions trigger both rising migration intensities and greater imbalance in inter-regional flows. As economic development progresses and processes such as suburbanisation emerge, both the intensity and effectiveness of migration decline. Internal migration has exerted a significant impact on the overall redistribution of population in South Korea over many decades, but the patterns of migration have changed over time. Figure  6.5 shows net migration gains and losses in Korean Provinces in 2018. The highest gains in 2018 were by the city of Sejong, which recorded a net migration rate of over 10%. Sejong is a new city, founded in 2007, as the de facto administrative capital of South Korea. All other metropolitan cities experienced net migration losses, with the largest loss recorded in Gwangju. Seoul recorded a net migration loss of 1.3%. Net migration gains were experienced in the provinces surrounding the metropolitan cities. This suggests suburbanisation processes are currently dominant in Korea. The proximate drivers of this shift are high housing and living costs in the metropolitan areas in contrast to medium-sized cities, which provide better residential circumstances for migrant households (Lee 2014; Kim 2018). These movements have been facilitated by improvements in transportation, communication and residential environments, supported by a deliberate national policy to promote balanced settlement outside the Seoul metropolitan region (Kwon and Kim 2002). Figures 6.6 and 6.7 show the system of bilateral migration flows between provinces in 1975 and 2018. In 1975 (Fig. 6.6), migration to and from Seoul accounted for a quarter of all migration flows. Flows to Seoul were drawn from neighbouring provinces (e.g., Gyeonggi-do) as well as from less urbanised provinces in the south and west. The largest outflows from Seoul were to the neighbouring province of Gyeonggi-do, with smaller counter-flows to other parts of the country. There was very little exchange between Seoul and Busan, Korea’s second largest city. The migration system in 2018 is markedly different (Fig. 6.8). Gyeonggi-do, the province immediately surrounding Seoul, now accounts for the largest share of gross migration flows and interchanges between Seoul and Gyeonggi-do constitute the largest migration stream. Flows across the rest of the settlement system are more balanced than in 1975, with exchanges between provinces and neighbouring regions offset by counter-flows, for example between Busan and Gyeongsangnam-do (Lee 2014; Abel and Heo 2018; Kim 2018). The fall in metropolitan populations due to the development of an industrial complex in the neighbouring province, coupled with suburbanisation, has been observed in Busan since 1995, and in Daegu, since 2000. Analyses of data from the Resident Population Register reveal increasing population outflows from the capital region to Sejong city and to the provinces of Chungcheongnam-do, Gangwon-do and Chungcheongbuk-do since 2015 (Kim 2018).

102

Y. Lee and D. S. Kim

Fig. 6.5  Net migration rates, provinces of South Korea, 2018. (Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project))

As described previously, South Korea underwent a rapid urban transition in the twentieth century, from 21.4% of the population living in towns and cities in 1950 to 81.6% in 2015. In the early stages of the transition, urban growth was concentrated in the Seoul metropolitan region, a so-called ‘truncated’ urban transition, whereby migrants from rural areas bypassed small towns to go directly to the Seoul metropolitan region (Mobrand 2012). The impact on population growth can be seen in Table  6.2, which shows average annual growth for Seoul and the surrounding metropolitan region since the 1960s. Population growth in Seoul peaked at an annual rate of 8.1% in the 1960s. The growth rate halved in the 1970s and halved again in the 1980s, but it was compensated by growth in the surrounding regions. A

6  Internal Migration in South Korea

103

Fig. 6.6  Bilateral migration flows, provinces of South Korea, 1975. (Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project). Note: Each province is assigned a colour (for example Seoul: Olive green) and flow arrows are assigned the same colour as the state of origin. The width of the arrows indicates the relative size of the migration flows. The volume of the flows is indicated by the tick marks on the circumference of the plot)

migration turnaround occurred in the 1990s at which time Seoul’s population growth turned negative, although the wider Seoul metropolitan continued to grow. To explore the urban transition in Korea and its subsequent turnaround we would ideally examine changes in the direction of flows between urban and rural regions. However, comparison of migration between rural and urban areas over time is inhibited by changes in definitions. To circumvent this issue, we adopt the approach used by Rees et al. (2017), which employs population density as a proxy for the conventional rural/urban classification. To this end, we calculate the net migration rates for each of the provinces and regress these against the logarithm of their population density. The relationship is captured by the slope parameter of a population-­ weighted linear regression. A positive slope indicates that migration is predominately from less densely populated to more densely populated regions (rural-to-urban),

104

Y. Lee and D. S. Kim

Fig. 6.7  Bilateral migration flows, provinces of South Korea, 2018. (Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project))

while a negative slope indicates that migration is predominately urban-to-rural. The magnitude of the slope indicates the strength of the process. Table 6.3 shows the slope coefficients at the province level in 1975, 2000 and 2018. In 1975 the coefficient was strongly positive and highly significant, indicating a strong redistributive effect from sparsely populated to more densely populated provinces (Fig. 6.8). The two largest cities, Seoul and Busan, experienced strong positive net migration rates of 6.5% and 4.0% respectively. Gyeonggi-do (which at this time included the city of Incheon, now Korea’s third largest city) also experienced high net migration gains of 4.3%. All other provinces registered net migration losses. The pattern of redistribution indicates a strong urbanisation process. By 2000 (Fig.  6.9), there is evidence of a migration turnaround. The strongly positive association between population density and net migration has been replaced by an inverted U-shaped curve. The large cities of Seoul, Busan and Daegu all

105

6  Internal Migration in South Korea

Fig. 6.8  Net migration rates by log population density, provinces of South Korea, 1975. (Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project)) Table 6.2 Average annual population growth, Seoul, Seoul Metropolitan Area, Gyeonggi Province, and Capital Region, 1960–2000 Interval 1960– 1970 1970– 1980 1980– 1990 1990– 2000

Seoul (%) 8.12

Seoul metropolitan area (%) 7.80

Gyeonggi province (%) 1.85

Capital region Total country (%) (%) 5.28 2.15

4.28

5.25

4.00

4.17

1.91

2.38

4.30

4.80

3.35

1.48

−0.70

1.64

3.63

1.39

0.61

Source: Kim (2018) and Statistics Korea (2019a)

experienced net migration losses, while the intermediate cities of Daejeon, Incheon and Ulsan all gained internal migrants. The growth of these metropolitan regions reflects deliberate government actions including the movement of government functions away from Seoul to cities lower down the urban hierarchy. It is also affected in part by the Korean government’s revision of the urban hierarchy system in 1995, which resulted in territorial expansion of the boundaries of major metropolitan

106

Y. Lee and D. S. Kim

Table 6.3  Net migration rate and population density, provinces of South Korea, 1975, 2000 and 2018

Year 1975 2000 2018

Slope coefficient 5.50 −0.007 −0.72

R-squared 0.672 0 0.088

Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project)

2.5

2.0

Gyeonggi-do

1.5

Net migration rate (%)

1.0 Daejeon

0.5

Ulsan

Incheon

Gyeongsangnam-do Gwangju

0.0

Chungcheongbuk-do Daegu Chungcheongnam-do

-0.5

Jeju Seoul

Gangwon-do

-1.0

Gyeongsangbuk-do Busan

Jeollabuk-do

-1.5 Jeollanam-do

-2.0

0.0

0.5

1.0

1.5

2.0 2.5 Log population density

3.0

3.5

4.0

4.5

Fig. 6.9  Net migration rates by log population density, provinces of South Korea, 2000. (Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project))

cities. The highest net gains in 2000 were in Gyeonggi province, part of the wider Seoul-­Incheon metropolitan agglomeration. This growth can be attributed to the suburbanisation of Seoul’s population. By 2018 (Fig. 6.10), a migration turnaround is clearly evident. Moderate losses from Korea’s largest cities of Seoul, Busan and Daegu continue unabated, while less densely populated provinces in the south (e.g., Jeollanam-do, Jeju-do) gained

6  Internal Migration in South Korea

107

12

10

Sejong

Net migration rate (%)

8

6

4

Jeollanam-do

2

Gyeonggi-do

Jeju

Incheon

Chungcheongnam-do Chungcheongbuk-do

0

Daegu

Gangwon-do Gyeongsangbuk-do Jeollabuk-do

-2

Ulsan

Gyeongsangnam-do

-4

0.0

0.5

1.0

1.5

2.0 2.5 Log population density

Busan Seoul

Daejeon Gwangju

3.0

3.5

4.0

4.5

Fig. 6.10  Net migration rates by log population density, provinces of South Korea, 2018. (Source: Calculated from KOSIS Database on Internal Migration (IMAGE-Asia Project))

migrants. A substantial share of migrants who moved to the southern provinces from the capital region reported the relocation of their workplace due to government policy and the housing environment/cost as their major reasons for migration (Statistics Korea 2019b). The highest net gain (10%) in 2018 was by Sejong city, the de facto administrative capital of South Korea, founded in 2007.

6.8  Understanding Internal Migration in South Korea From the end of the Korean War until the 1980s, internal migration in South Korea was dominated by massive rural-to-urban flows, concentrated on the mega-­ metropolitan region of Seoul, and to a lesser extent, on the second largest city of Busan. Such was the concentration of migration flows that scholars have described a compressed or truncated urban transition in South Korea, whereby mid-sized towns and cities were bypassed in favour of the largest metropolitan regions. The rapid growth of the Seoul metropolitan region ultimately led to significant spill-over

108

Y. Lee and D. S. Kim

to the adjacent provinces and the growth of neighbouring cities such as Incheon. Mobility throughout the 1970s was extraordinarily high by international standards with 25% of Koreans changing their place of usual residence in 1975. This may be among the highest levels of voluntary internal migration observed at any time in human history. The introduction of a range of government policies aimed at decentralising populations away from Seoul to new regional cities, coupled with rising costs of living, has led to increased suburbanisation of the Korean population since the 1990s. Decentralisation of industry along with the growth of tertiary industries has also increased inter-urban flows. Migration intensity has declined, and the peak age at migration has risen as Korean youth increase their participation in higher education (Yu 1980; Choi and Chang 2004; Lee 2014). Population ageing, particularly in rural areas has also had an impact on migration, with the pool of rural workers declining markedly over time. A consequence of sustained rural-to-urban migration over many decades coupled with ageing has been the increased socioeconomic divergence between rural and urban localities. Rural areas have increasingly sought international migrants from less developed Asian countries to address this demographic deficit. The contemporary migration system of Korea is marked by continued high migration intensity, but a more balanced system of migration flows, reflecting the overall high level of urbanisation in the country and growth of intermediate cities. The growth of new cities such as Sejong, and satellite cities such as Hanam, Hwaseong and Gimpo has led to high net migration gains in specific parts of the country (Statistics Korea 2019b). As a consequence of socioeconomic development and industrialisation during the past half-century, internal migration in Korea has changed its volume, direction and intensity (Kim and Sloboda 1981; Kwon and Lee 1997; DeWind et al. 2012).

6.9  Impacts and Implications South Korea stands apart from other countries in Asia and around the globe with respect to both its high migration intensity, and the rapid urbanisation process driven by massive rural-to-urban migration from the 1960s to the 1980s. Rising standards of living, demand for urban industrial labour, and better educational and administrative infrastructure increased rural-to-urban migration flows over this period, with young migrants from rural areas contributing to highly compressed industrialisation and urbanisation processes. Internal migration and urbanisation contributed strongly to the growth of the Korean economy throughout the twentieth century. The combined benefits of market proximity and the economics of agglomeration offered considerable labour market opportunities as well as meeting the service and infrastructure needs of large numbers of people. Urban businesses generated jobs and resources that influenced global markets. The impacts of internal migration on individuals and households was also significant. Internal migration facilitated social

6  Internal Migration in South Korea

109

mobility for young adults in many traditional rural communities. Industrialisation allowed people with specific skills and higher education to leave their hometowns, while individuals with lower levels of human capital were left behind. The dramatic growth in internal migration and uncontrolled urbanisation in the 1960s to 1980s led to economic decline in rural areas. The volume of internal migration expanded steadily until the early 1980s, at which time it began to stabilise then decline. This was due both to demographic factors and government policy. At this time, urbanisation reached saturation level and population growth in rural provinces began to slow due to below replacement fertility, reducing the potential pool of rural-to-urban migrants (Kwon and Kim 2002; Statistics Korea 2019a). The national government also became concerned with the level of population concentration in metropolitan areas and introduced a range of measures to encourage decentralisation. These included the transfer of government offices and industrial facilities to local areas, the development of new industrial cities, prohibition on the construction of factories in big cities, and the attraction of educational and cultural facilities to local regions (Kwon and Kim 2002; Kim 1988). The Second National Territory Comprehensive Development Plan (1982–1991) focussed on encouraging migrant settlement in local areas and increasing rates of land utilisation. To this end, the government strengthened the political, cultural, administrative, and management functions of large local cities and placed industrial complexes around small and medium-sized cities to promote local industries and population settlement. In addition, policymakers made efforts to narrow the gap between urban and rural areas by placing amenities and social overhead capital (SOC) facilities in a balanced manner among local regions and pursued comprehensive development policies for rural areas (Kwon and Kim 2002). In Seoul, the introduction of green belt areas in the early 1980s contributed to restricting industrial expansion. This was coupled with the construction of subway connections to satellite cities to encourage population dispersal (Kim 1988) along with other policies such as restricting school transfers into Seoul. The Third National Territory Comprehensive Development Plan (1992–1999) focussed its strategy on construction of the west coast industrial zones, decentralisation of infrastructure in the country, and construction of comprehensive highspeed exchange networks for workers and their families. All these policies were aimed at redistributing population by changing the flow of market resources and deconcentrating socioeconomic and recreational infrastructure. The outcome of these policies has been large-scale outward movement of urban residents from the largest cities to areas on the periphery and to intermediate cities. Younger urban residents, particularly those of lower economic status, have been a major component of the migration into new regional cities, particularly in recent years. The outcome of these policies has been a less intense and more balanced migration system.

110

Y. Lee and D. S. Kim

6.10  Conclusions Internal migration in the twentieth century has transformed the settlement system of South Korea more rapidly and more dramatically than anything seen in almost any other part of Asia and perhaps the world. Extremely high levels of migration intensity coupled with high migration effectiveness led to mass redistribution of the population throughout the 1960s and 1970s, and created an extreme primate settlement distribution dominated by Seoul. Later the impact of demographic ageing and decline in the population of rural areas, alongside strong government intervention, led to an overall decline in migration intensity and a rebalancing of the migration system to one in which inter-urban and suburban flows are now the dominant migration processes. The Korean mobility transition closely follows the empirical model set out by Rees et al. (2017), whereby as countries modernise, rural-to-urban migration is replaced by inter-urban, suburban and counter-urban flows. Unique to the South Korean case is the scale of the impact of government policies aimed at restricting movement to the capital, and the decentralisation of industry and population towards intermediate cities, all facilitated by highly efficient transportation systems. The example of South Korea is relevant to the development and migratory trajectories of other developing Asian countries. Less developed countries need to pay attention to the distribution of national populations across regions to achieve equal and efficient economic development. Comparative study of internal migration patterns and their consequences provides a valuable framework for discussing the role of spatial mobility in shaping urban systems. Acknowledgements  This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2017S1A3A2065967).

References Abel, G. J., & Heo, N. (2018). Changing internal migration flows patterns in South Korea. Regional Studies, Regional Science, 5(1), 78–80. Bell, M., Charles-Edwards, E., Ueffing, P., Stillwell, J., Kupiszewski, M., & Kupiszewska, D. (2015). Internal migration and development: Comparing migration intensities around the world. Population and Development Review, 41(1), 33–58. Bernard, A., Bell, M., & Charles-Edwards, E. (2014). Life-course transitions and the age profile of internal migration. Population and Development Review, 40(2), 213–239. Choe, E. H., Yoon, J. J., Kim, S. B., Chung, K. W., & Huguet, G. (1989). Patterns and determinants of internal migration. A report on population analysis of the 1985 population and housing census. Seoul: National Bureau of Statistics, Korean Government. (In Korean) Choi, J. H., & Chang, S. H. (2004). Population distribution, internal migration and urbanization. In D.-S. Kim & C.-S. Kim (Eds.), The population of Korea (pp. 225–251). Daejeon: Korea National Statistical Office. Choi, J. H., & Choi, B. S. (1993). Causes and effects of unequal population distribution between regions. Seoul: Korea National Statistical Office. (In Korean)

6  Internal Migration in South Korea

111

DeWind, J., Kim, E. M., Skeldon, R., & Yoon, I. J. (2012). Korean development and migration. Journal of Ethnic and Migration Studies, 38(3), 371–388. Kang, M. K. (1993). Capitalistic spatial division and its political process. Economy and Society, 20, 230–252. (In Korean) Kim, H. K. (1982). Social factors of migration from rural to urban areas with special reference to developing countries: The case of Korea. Social Indicators Research, 10(1), 29–74. Kim, W. B. (1988). Population redistribution policy in Korea: A review. Population Research and Policy Review, 7(1), 49–77. Kim, D. S. (2004). Population growth and transition. In D. S. Kim & C. S. Kim (Eds.), The population of Korea (pp. 1–32). Daejeon: Korea National Statistical Office. Kim, K.  T. (2017). Changes in regional population due to low fertility and internal migration. Korean Social Trends 2017 (pp.  36–42). Daejeon: Statistical Research Institute, Statistics Korea. (In Korean) Kim, D. S. (2018). Major demographic changes. Korean Social Trends 2018 (pp. 24–36). Daejeon: Statistical Research Institute, Statistics Korea. (In Korean) Kim, D.  Y., & Lee, H.  K. (1976). Characteristics of internal migration in Korea, 1965–1970. Seoul: Korea Development Institute. (In Korean) Kim, D. Y., & Sloboda, J. E. (1981). Migration and Korean development. In R. Repetto, T. H. Kwon, S. U. Kim, D. Y. Kim, J. E. Sloboda, & P. J. Donaldson (Eds.), Economic development, population policy, and demographic transition in the Republic of Korea (pp. 36–138). Cambridge: Harvard University Press. Kim, N. I., Choi, S., Park, W. S., & Yang, K. S. (1997). Internal migration and changes in characteristics of rural population. Seoul: Korea National Statistical Office. (In Korean) Kwon, T. H., & Kim, D. S. (2002). Understanding population. Seoul: Seoul National University Press. (In Korean) Kwon, Y., & Lee, J. (1997). Residential mobility in the Seoul Metropolitan Region, Korea. GeoJournal, 43(4), 389–395. Kwon, T. H., Lee, H. Y., Chang, Y., & Yu, E. Y. (1975). The population of Korea. Seoul: Population and Development Studies Center, Seoul National University. Lee, H. Y. (2014). Characteristics and patterns of internal migration. Korean Social Trends 2014 (pp. 43–51). Daejeon: Statistical Research Institute, Statistics Korea. (In Korean) Mobrand, E. (2012). Reverse remittances: Internal migration and rural-to-urban remittances in industrialising South Korea. Journal of Ethnic and Migration Studies, 38(3), 389–411. Moon, S.  N. (1993). Urbanization in Korea: Its types and characteristics. Studies on Regional Development, 29, 29–66. Gwangju: Cheonnam National University. (In Korean) Park, H. Y., & Fullerton, H. H. (1980). Rural-urban labor migration: The case of Korea. The Annals of Regional Science, 14(1), 72–90. Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), 1–22. https://doi.org/10.1002/ psp.2036. Statistics Korea. (2019a). Korean Statistical Information Service (KOSIS). http://kosis.kr/. Accessed 23 Feb 2019. Statistics Korea. (2019b). Annual report on internal migration statistics, 2018. Press release on January 28, 2019. Division of Social Statistics, Statistics Korea. United Nations. (2018). World urbanization prospects: The 2018 revision. New  York: United Nations Population Division. Yu, E. Y. (1980). Internal migration and development in cities. In Y. Chang (Ed.), Korea: A decade of development (pp. 143–175). Seoul: Seoul National University Press.

Chapter 7

Internal Migration in Japan Yoshitaka Ishikawa

7.1  Introduction Since Japan’s population peaked at 128.1 million in 2008, population decline has become a chief concern of the nation. The country has subsequently witnessed a sustained decline in its population and is predicted to remain on this path, reaching a population below 100 million around 2050 (National Institute of Population and Social Security Research 2017). Underpinned by an ageing population coupled with limited immigration, this demographic trend has profound economic implications, including a rapid reduction of the labour force. As the main agent of population change, migration shapes patterns of human settlement and thus plays a particularly significant role at a regional level, exacerbating depopulation in some regions while mitigating population ageing by attracting migrants in others. In this context, a comprehensive understanding of migration processes and their role in redistributing population in Japan is essential as the country embarks on this inexorable demographic transformation. Japan’s national territory has a long and slender shape. Its three largest metropolitan areas, Tokyo, Osaka and Nagoya, are located in the central part of the main island (Honshu) while the four regional capitals are situated in northern (Sapporo and Sendai) and western Japan (Hiroshima and Fukuoka). While the country is highly urbanised, non-metropolitan areas tend to face depopulation and ageing. This population distribution is closely linked to increasing socioeconomic disparities in contemporary Japan. Considerable literature on internal migration, mainly by Japanese population geographers, has accumulated since the 1960s (Otomo 1998; Ishikawa 2001b, 2008). Comparative analyses include migration age patterns (Nanjo et  al. 1982; Kawabe 1991), the migration turnaround (Ishikawa 1999, 2001a), spatial Y. Ishikawa (*) Faculty of Economics, Teikyo University, Hachioji City, Tokyo, Japan e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_7

113

114

Y. Ishikawa

interaction modelling (Yano et al. 2003), the link between migration and life-course transitions (Fielding and Ishikawa 2003) and newly constructed ‘model-mobility’ measures (Hayashi 2014). These studies have focussed on particular aspects of migration and thus a systematic understanding of Japan’s migration processes encompassing the different dimensions of migration remains elusive. To maximise comparability with other countries, this chapter adopts the measures and analytical techniques developed as part of the Internal Migration Around the GlobE (IMAGE) project. After outlining the sources of migration data in the second section and discussing the spatial framework used to analyse migration data in the third section, Sect. 7.4 reviews prior research in the post-war era. The following sections examine the three major aspects of internal migration based on national and regional indicators: migration intensity in Sect. 7.5, characteristics of migrants in Sect. 7.6, and spatial patterns in Sect. 7.7. Sections 7.8 and 7.9 discuss the determinants and consequences of internal migration, respectively. Section 7.10 concludes this paper.

7.2  Internal Migration Data This section sets out the nature of the migration data collected in Japan by identifying key data sources and types of data and by reviewing the spatial scales and frequency at which they are collected. The two major sources of internal migration data in Japan are the population census and the resident registers (Shimizu 2011). Data from the migration survey conducted by the National Institute of Population and Social Security Research are also available. The census has been conducted on October 1 approximately every five years since 1920, but a question related to migration has been included only every ten years, namely in 1960, 1970, 1980, 1990, 2000 and 2010. While the 2015 census was not scheduled to include a question on migration, the Great East Japan Earthquake and the resulting tsunami that occurred on March 11, 2011, prompted the country’s Statistics Bureau to collect migration data in order to assess the impact on Japan’s population of the earthquake and subsequent accident at Tokyo Electric Power’s Fukushima Nuclear Power Station. The population census includes a question about each person’s current address and their address five years prior to the census. Specifically, the 2015 census included the following questions: Where did you live five years ago (as of October 1, 2010)? • A person born after October 1, 2010, should mark the place at which he/she lived after his/her birth as ‘Same as present’, ‘Another address within the same municipality’, ‘Another place within the same prefecture’, ‘Another prefecture’, or ‘Outside of Japan’. • Respondents who selected ‘Another place within the same prefecture’ or ‘Another prefecture’ were then asked to write the name of the municipality or prefecture respectively. For the Tokyo Metropolitan area or one of 20

7  Internal Migration in Japan

115

Ordinance-­designated cities1 with more than 500,000 residents, respondents were also asked to write the name of the ward. Internal migration data derived from these questions are published by the Statistics Bureau, Ministry of Internal Affairs and Communications. All changes of address within the country, regardless of spatial scale, can be derived by subtracting residents who responded ‘Same as present’ from all residents, discounting those previously residing outside Japan. Outputs are tabulated by various attributes including sex, age, educational attainment, nationality and occupation, thus providing a major source of information on migration in Japan. Count data and origin-destination (OD) matrices of inter-prefectural migration flows are also available by age group. A limitation of census data is the rate of non-response, which has increased in recent years, particularly for migration-related questions. The rate of non-response for the question concerning place of residence five years ago was 6.5% at the 2010 census (Koike and Yamauchi 2014) and rose to 8.8% in 2015. Two other census questions with a non-response rate greater than 8% were about nationality, other than Japanese (8.9%), and duration of residence (8.6%). Since 1954, annual internal migration data in Japan have also been available from the resident registers. When a Japanese resident moves to a new municipality, the person is legally required to notify the municipality office of their new address, which means that data on migration within a municipality are not collected. Based on this registration procedure, the Annual Report on the Internal Migration in Japan Derived from the Basic Resident Registers is published by the Statistics Bureau, Ministry of Internal Affairs and Communications. The main strength of these data is their annual availability over a period now extending back almost two-thirds of a century with publication each year of inter-prefectural OD matrices. Such matrices are also available by sex since 1999. It is important to note that the resident registers were limited to Japanese residents until 2011. Since 2012 they have also included foreign residents. In addition, since Okinawa remained under U.S. military control for 19 years until its reversion to Japan in 1972, migration in and out of that prefecture up to 1972 is not included in this data collection. Another limitation is under-­ reporting by young adults, although no accurate estimates of this omission are available. Migration data from the census and the resident registers are available free of charge at the e-Stat website (seifu tokei no sogo madoguchi), mostly in the form of Excel files (https://www.e-stat.go.jp/). It is notable that Japan is one of just ten countries (with Australia, Canada, Greece, Israel, Malta, Portugal, Spain, Switzerland, and the United States) that collect both one and five-year migration data (Bell et al. 2015). The National Institute of Population and Social Security Research migration survey has been conducted every five years since 1976, with the eighth national survey in 2016 covering approximately 123,000 respondents. Unlike the population census 1   Sapporo, Sendai, Saitama, Chiba, Yokohama, Kawasaki, Sagamihara, Niigata, Shizuoka, Hamamatsu, Nagoya, Kyoto, Osaka, Sakai, Kobe, Okayama, Hiroshima, Kitakyushu, Fukuoka, and Kumamoto.

116

Y. Ishikawa

and the resident registers, this survey places particular stress on investigating migration careers (National Institute of Population and Social Security Research 2018). This chapter draws on these three sources of data: the Census, Resident Registers, and the Migration survey.

7.3  The Spatial Framework This section describes the administrative geography used to collect and disseminate data on internal migration, including the numbers of zones at various levels of the spatial hierarchy. The five levels of geography are region (chiho), prefecture (to, do, fu, and ken), municipality (shi, machi, and mura), census tract (kokusei-tokei-ku) and neighbourhood (cho-cho and aza). There are several possible methods of dividing regions, but Fig. 7.1 is based on a typical regional division and shows a map of Japan with the names of regions and prefectures. There are eight regions (Hokkaido, Tohoku, Kanto, Chubu, Kinki, Chugoku, Shikoku and Kyushu). Each region (except

Fig. 7.1  Metropolitan areas and prefectures of Japan

7  Internal Migration in Japan

117

Hokkaido) consists of several prefectures, totalling 47 nationally. The number of prefectures has been stable over time. Each prefecture is divided into municipalities whose total number has been declining from 3280 (April 1, 1970) to 3235 (April 1, 2000) and 1724 (April 1, 2019), which is the result of a large-scale merger of municipalities between 2000 and 2010. The three largest metropolitan areas consist of 11 prefectures: the Tokyo area (consisting of the four prefectures of Saitama, Chiba, Tokyo and Kanagawa) with 36.1 million inhabitants at the 2015 census, the Nagoya area (Gifu, Aichi and Mie) with 11.3 million inhabitants, and the Osaka Area (Kyoto, Osaka, Hyogo and Nara) with 18.3 million inhabitants. The binary division of Japan into these three metropolitan areas versus the remainder of the country, which are often referred to as the core and periphery respectively (Vining and Pallone 1982; Ishikawa 1999), has long been a topic of debate. The periphery consists of 36 prefectures located in the regions of Hokkaido, Tohoku, Kanto (excluding the Tokyo area), Chubu (excluding the Nagoya area), Kinki (excluding the Osaka area), Chugoku, Shikoku and Kyushu. In terms of Japan’s urban system, four cities function as regional capitals, namely Sapporo in Hokkaido, Sendai in Tohoku, Hiroshima in Chugoku, and Fukuoka in Kyushu, which are important secondary cities. These cities are located in the prefectures of Hokkaido, Miyagi, Hiroshima and Fukuoka.

7.4  Prior Research A summary of prior research on internal migration within Japan is given in this section by drawing attention to its historical development, key themes and types of data used. In considering internal migration trends in the post-war period (Shimizu 2011; Sakuno 2011; Fielding 2016, 2018), it is convenient to consider three periods: the first characterised by high economic growth that lasted until the early 1970s, the second period featuring the migration turnaround that lasted from the early 1970s to the mid-1990s, and the third period from the mid-1990s to the present. This division generally corresponds to major changes in the economic situation. Figure 7.2 shows the trends in absolute net migration to the three major metropolitan areas of Tokyo, Osaka and Nagoya over the post-war period. In the first period, Japan recovered from the chaos immediately following World War II with high economic growth from the 1950s. This growth was supported by massive migration of high school graduates from regions peripheral to the three largest metropolitan areas of Tokyo, Osaka and Nagoya, chiefly due to abundant employment opportunities and income differences between origins and destinations. Annual net migration to these three areas as a whole exceeded 400,000 (Kuroda 1976; Shimizu 2011). This high level of migration contributed to solving labour shortages in metropolitan areas, where heavy and chemical industries were developing rapidly. This large human movement, however, resulted in intense urbanisation and environmental pollution in destination areas. To cope with such issues, policies restricting the location and expansion of firms and universities in the

118

Y. Ishikawa 700 Total 600

Tokyo Area Nagoya Area

500 Net migration (thousands)

Osaka Area 400

300

200

100

0

-100

1955

1960

1965

1970

1975

1980

1985 Year

1990

1995

2000

2005

2010

2015

Fig. 7.2  Inter-prefectural migration, major metropolitan areas of Japan, 1954–2017. (Source: Annual Report on Internal Migration in Japan derived from the Basic Resident Registers (https:// www.e-stat.go.jp/en/))

metropolitan areas and promoting depopulated municipalities were implemented (Hama 1982; Sakuno 2011). In this context, key research themes focussed on internal migration towards the three major metropolitan areas and the resulting rapid urbanisation using inter-­ prefectural migration data from resident registers. Partly because of the lack of detailed information on the attributes of migrants other than sex, major interests in the literature dealing with this period were changes in spatial patterns and determinants at origin and destination (e.g., Kuroda 1976; Okada 1976; Ishikawa 1978). The analytical methods used were descriptive statistics and multivariate analysis, including regression. The second period from the early 1970s to the mid-1990s was triggered by the first oil crisis in 1973 and the resultant economic stagnation. The attractive power of the three major metropolitan areas declined sharply; net migration in the three areas as a whole reversed from a gain of 410,000 in 1970 to a loss of 10,000 in 1976. In addition to this reduction in periphery-to-core migration, a small decrease in core-­ to-­periphery migration was also observed. This resulted in balanced migration flows between the periphery and the core. Although peripheral regions lost many young people as a result of massive out-migration to the major metropolitan areas in the first period, this was greatly weakened in the second period (Ito 1984). This implies relative improvement of the peripheral regions’ position compared with the major

7  Internal Migration in Japan

119

metropolitan areas in terms of internal migration. The term ‘Era of peripheral regions’ (chiho no jidai) was born in this context. Importantly, net migration to the three major metropolitan areas began to follow different trends in the period of the bubble economy (mid-1980s to early 1990s). The Tokyo area became a world city through successful industrial restructuring into service industries, including the financial sector. This caused a revival of population absorption from all over the country (Ishikawa and Fielding 1998) as shown in Fig. 7.2. In contrast, since the mid-1970s Osaka and Nagoya have shown slightly negative and nearly zero net-migration, respectively. Of major concern during this period were the two migration turnarounds, in which net migration to the core showed a sharp decrease in the second half of the 1970s followed by a recovery in the second half of the 1980s. Similar turnarounds have been observed in many advanced countries, chiefly through the work of Vining and Pallone (1982), and there have been attempts to search for common reasons for these turnarounds in Japan (Ishikawa 1999, 2001a, b), including the impact of economic restructuring (Wiltshire 1992). Return migration to peripheral regions also attracted attention (Okada 1976; Wiltshire 1979). Other key topics in this period include age-specific migration following the Migration and Settlement Project at IIASA (Rogers and Castro 1986). Census data on age-specific inter-prefectural and inter-regional migration were used to investigate these topics (Kawabe 1991; Ishikawa 1999, 2001a). An important finding was the absence of a retirement peak in Japan’s migration schedule. The third period was characterised by the expansion of regional disparities. Only the Tokyo area continued to record net migration gains despite large fluctuations as shown in Fig. 7.2: −16,900 in 1994, 155,200 in 2007, 62,800 in 2011 and 119,800 in 2017. Meanwhile, depopulation in most peripheral regions has become a serious problem in contemporary Japan, which is often referred to as ‘Tokyo ikkyoku shuuchu’ (mono-polar concentration into Tokyo). While the key theme in this third period is population decline, a range of research interests can be found. Common topics include migration behaviour over the life course (Fielding and Ishikawa 2003; Inoue and Liaw 2004; Nakazawa 2008), population re-concentration in central parts of major cities since the mid-1990s and suburban decline (Esaki 2006; Koike 2015, 2017), population ageing and migration of the elderly (Hirai 2007; Inoue and Watanabe 2014; Nakazawa 2017), destination choice of foreign residents (Ishikawa and Liaw 2009; Hanaoka et  al. 2017) and replacement migration at the municipality scale (Shimizu 2017). Fielding and Ishikawa (2003) demonstrated that economic factors are more significant in driving migration in Japan than in Britain, where social factors seem to be more important. More recently, Shimizu (2017) showed that net migration loss of Japanese residents was ‘fully or substantially offset’ by a net migration gain of foreign residents in only 159 of the 1893 municipalities (including wards) and these gains were not sustained over time, suggesting the instability of international migration as a domestic population replacement mechanism. This line of inquiry has led to calls for increasing policy intervention to mitigate population decline (Masuda 2015; Ishikawa 2018; Nakazawa 2017). The studies cited above have used migration data from all three

120

Y. Ishikawa

major sources and have increasingly drawn on GIS data, which have been made available by the national government as digital files since the mid-1990s. This innovation has enabled researchers to quickly conduct detailed analysis based on maps of, for example, net migration by prefecture and municipality, exemplified by Yano et al. (2000, 2003) and Shimizu (2017). Despite the wide range of migration studies in Japan, there have been few attempts to provide a systematic analysis of internal migration processes. The following sections seek to address this gap by focussing on the intensity, selectivity and impact of migration.

7.5  How Much Movement? Migration Intensity We use the Crude Migration Intensity (CMI), which is defined as the sum of migrants divided by the population at risk, expressed as a percentage. Figure 7.3 shows intraand inter-prefecture CMIs from 1954 to 2017. The inter-prefecture CMI in 1954 was 2.7%, which gradually increased to a peak of 4.1% in 1970. It has shown a near monotonic decrease since the 1970s, with the latest value in 2017 being 2%. Intra-­ prefecture migration has followed the same downward trend and the rate of decline has been similar at both spatial scales. Based on the periodisation discussed in section four, migration intensity (especially inter-prefectural migration) rose rapidly in the first period of high economic growth, which lasted until the mid 1970s. This reflects the massive migration of 9.0

Intra-prefectural migrants

8.0

Inter-prefectural migrants

Migration intensities (%)

7.0 6.0 5.0 4.0 3.0 2.0 1.0 2015

2010

2005

2000

1995

1990

1985

1980

1975

1970

1965

1960

1955

0.0

Year Fig. 7.3  Intra- and inter-prefectural migration intensities, Japan, 1954–2017. (Source: Annual Report on Internal Migration in Japan derived from the Basic Resident Registers (https://www.estat.go.jp/en/))

7  Internal Migration in Japan

121

young people, particularly new graduates from peripheral regions, to major metropolitan areas where demand for labour was very high. Migration intensity declined gradually in the second and third periods, reflecting a decline in the proportion of young people in source areas and a gradual process of population ageing. Thus, the downward trend of migration intensity in Japan is the result of both economic and demographic factors, the influence of which has been aggravated by consistently low levels of international migration. While the CMI depends on the number of spatial units over which migration is measured, the Aggregate Crude Migration Intensity (ACMI) captures all changes of address, and thus allows direct cross-national comparisons (see chapter three and Bell et al. 2015). In Japan, the five-year ACMI based on census data decreased from 28.13% in 2000 to 21.26% in 2010 and 19.92% in 2015. This gradual decrease is consistent with the downward trend observed for both intra- and inter-prefecture migration. The Asian mean for the ACMI is 17.9% and the global mean is 21.0%, revealing that Japan’s level of population movement is similar to the global average and higher than most countries in Asia (Charles-Edwards et al. 2019). Japan’s ACMI is on par with the more developed countries of Asia, Israel and South Korea, which suggests that it is a product of the high level of national development (Bell et al. 2015).

7.6  Who Moves? The Characteristics of Migrants This section examines the key characteristics of internal migrants, focussing on age, sex, education and nationality, for which data are readily available in Japan. Whereas discussion of age, education and nationality focus on all moves, irrespective of spatial scale as derived from the census, sex differentials relate to inter-prefectural migration as derived from the resident registers. Figure 7.4 charts 2015 Age-Specific Migration Intensities (ASMIs), defined as the number of migrants of age x as a percentage of the population of age x at the time of census. The shape of the profile closely matches that observed widely across the world, but with a number of subtle but differences. As age advances, the intensity decreases sharply from 39.6% at age five and reaches its lowest level of 14.1% at age 17. It then increases, reaching a peak of 44.8% at age 31, decreasing rapidly thereafter. The sharp rise in intensity at age 18 reflects migration associated with entry to universities, while migration in the early thirties is mainly due to marriage and housing adjustment, according to the eighth national survey on migration conducted in 2016. The ASMIs of Asian countries like China, India and Armenia are strongly concentrated around an Age at Migration Peak (AMP) of 20–25 years, but Japan’s peak occurs much later in the life course, closely resembling that of European and North American countries (Bernard et  al. 2014), which has been explained by transitions to key adult roles occurring later in life. Figure 7.4 confirms results from the Migration and Settlement Project (Rogers and Castro 1986) that Japan does not exhibit evidence of retirement migration in the mid-sixties at a

122

Y. Ishikawa 50.0 45.0 40.0

Migration intensity (%)

35.0 30.0 25.0 20.0 15.0 10.0 5.0 0.0

5

10

15

20

25

30

35

40 Age

45

50

55

60

65

70

75

Fig. 7.4 Age-specific migration intensities, all moves, Japan, 2010–2015. (Source: 2015 Population Census, all changes of address)

national level. However, analysis for certain prefectures and municipalities does reveal an upsurge in migration intensity at retirement ages. All peripheral prefectures, with the exception of Hiroshima, showed net migration gains between the ages of 60 and 64 at the 2010 census (Ishikawa 2016), indicating a movement of retirees from the core to the periphery. With regard to sex selectivity, Liaw (2003) pointed out that inter-prefecture migration intensity is lower for females than males, which contrasts with other developed countries such as Canada where sex differentials are minimal. Patterns of sex selectivity are mixed in Asia; in countries such as China males tend to be more mobile than females, as shown in the chapter by Shen in this volume, whereas the reverse is true in India, where females are more likely to engage in permanent internal migration and males are more likely to move temporarily, as shown by Bhagat and Keshri, also in this volume. Following Liaw (2003), Fig. 7.5 uses an indicator of the sex ratio that measures the number of male migrants per 100 female migrants for one-year inter-prefectural migration for Japanese nationals. The ratios are consistently and substantially above 100, confirming the persistence of a strong sex selectivity in Japanese migration streams. According to Liaw (2003), male dominance is rooted in Japan’s traditional sense of family values but fluctuates with economic conditions. The relatively higher ratios in the late 1980s and mid-2000s

7  Internal Migration in Japan

123

145 140

Sex ratio

135 130 125 120 115 110 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015 Year Fig. 7.5  Sex ratio of inter-prefectural migration, Japan, 1960–2015. (Source: Annual Report on Internal Migration in Japan derived from the Basic Resident Registers (https://www.e-stat.go.jp/ en/). Note: Males per hundred females)

correspond to economic booms, whereas lower ratios are observed in periods of economic bust. Migrants are a select group not only with respect to age and sex but also education. In countries around the world the propensity to move increases progressively with the level of education (Bernard and Bell 2018). To examine the educational selectivity of migrants we use data from the 2010 census because educational attainment was not collected at the 2015 census. Migration intensity varies from 13.4% for individuals who did not graduate from high school to 19.% for high school graduates, 28.1% for college graduates and 31.6% for university graduates. These results show a clear, positive relationship between education and migration—the higher the level of education the higher the migration intensity. Finally, attention is devoted to variation by nationality. Data from the 2015 population census show that the crude migration intensity is broadly similar for foreign residents (19.6%) and Japanese residents (19.7%). This is surprising given that in a large number of countries, including Australia, the United Kingdom and the United States, the overseas-born have been found to be more mobile as they adjust their housing needs and employment status upon arrival (Bell and Hugo 2000; Raymer and Baffour 2018; Laukova et al. 2020). However, in Japan some foreign-born residents display intensities higher than 20%: 33.3% for Brazilians, 31.3% for Peruvians, 24.8% for Filipinos, 22.6% for British, 20.4% for Koreans, and 20.3% for Chinese. The higher migration intensity of foreign residents partly reflects differences in age

124

Y. Ishikawa

structure. In addition, Brazilian and Peruvian residents show high migration rates because of the recency of their immigration to Japan and their willingness to move to obtain better incomes (Ishikawa and Liaw 2009).

7.7  Where Do They Move? Spatial Patterns This section, based on the indicators developed in the IMAGE project, turns to the dimension that is most prominent in the eyes of policymakers and planners, that is, the way migration serves to redistribute population and change the pattern of human settlement within Japan. Although migration flows in a country are complicated, they can be effectively summarised by various indicators. The first indicator is the Migration Effectiveness Index (MEI), which reveals the relative strength of migration streams and counter-­ streams. This index is defined as the sum of the absolute net migration gains and losses aggregated across all prefectures divided by total volume of migration flows across all prefectures (see chapter three of this volume). It indicates the efficiency of migration in redistributing population between prefectures with values ranging between 0 and 100. High values indicate that, overall, migration is an efficient mechanism for population redistribution, generating a large net effect for the given volume of movement. Inter-prefectural migration in the 1950s and 1960s was highly effective in terms of this indicator, reflecting the strong contribution of peripheryto-­core migration during the country’s high economic growth at that time (Liaw 2003): combined annual net migration in the three major metropolitan areas totalled 400,000 or more. Conversely, low values denote that inter-prefectural flows are more closely balanced, leading to comparatively little redistribution. The MEI for 2017 inter-prefectural migration was 5.39 based on 2017 resident register data and 4.19 based on 2015 census data, suggesting relatively inefficient redistribution. This represents one of the lowest values in the world (Rees et  al. 2017), but conceals considerable differences between age groups, as shown in Fig.  7.6. These MEI values, calculated using data from the 2015 census, vary between 0.41 at ages 10–14 and 28.94 at age 25–29 and rise again at the oldest ages. The higher index value itself does not tell us in which direction migrants move, but this can be determined from age-specific OD matrices. For the five largest age groups, with 500,000 inter-prefectural migrants or more, the 20–24 (856,000 persons) and 25–29 (862,000) age groups exhibit a periphery-to-core migration pattern, while those aged 30–34 (811,000), 35–39 (659,000) and 40–44 (535,000) show the reverse pattern with a net core-to-periphery migration. Consequently, these two flows with different directions largely cancel each other out, resulting in the very low all-age MEI for Japan. This suggests that the term ‘mono-polar concentration into Tokyo’ is particularly relevant to the migration of young adults. The next indicator is the Aggregate Net Migration Rate (ANMR), which is defined as half the sum of the absolute net migration flows aggregated across all regions, divided by the population at risk. The ANMR thus measures the overall

7  Internal Migration in Japan

125

35

Migration effectiveness index

30 25 20 15 10 5

85+

80-84

75-79

70-74

65-69

60-64

55-59

50-54

45-49

40-44

35-39

30-34

25-29

20-24

15-19

10-14

5-9

All Ages

0

Age

Fig. 7.6 Age-specific migration effectiveness index, Japan, 2010–2015. (Source: 2015 Population Census)

impact of migration on population redistribution: it identifies the net shift of population between the country’s regions per 100 residents and is the product of the CMI and MEI. The value of ANMR for 2017 inter-prefectural migration is 0.05. This value is extremely low by both global and Asian standards. This is the result of the low MEI combined with Japan’s moderate CMI. Despite relatively high mobility, there is little net redistribution of population between prefectures. Turning to regional variations in the impact and spatial patterns of migration, Fig. 7.7 maps the net-migration rate by prefecture in 2017. Only seven prefectures (Saitama, Chiba, Tokyo, Kanagawa, Aichi, Osaka and Fukuoka) experienced population gains. Among them, four prefectures (Saitama, Chiba, Tokyo and Kanagawa) lie within the Tokyo metropolitan area. Thus, the largest metropolitan area in Japan retains a strong attraction for people from other parts of the country. Aichi and Osaka are central prefectures of the Nagoya and Osaka areas, respectively, while Fukuoka prefecture is home to Fukuoka City, the regional capital of the Kyushu region. Meanwhile, the other 40 prefectures have suffered net migration losses. The largest losses (below −0.25% per annum) are observed in 14 prefectures of which 12 are situated in the periphery (Aomori, Iwate, Akita, Yamagata, Fukushima, Niigata, Yamanashi, Wakayama, Yamaguchi, Tokushima, Kochi, and Miyazaki) and only two are in the core (Gifu and Nara). The metropolitan areas of the four regional centres (Sapporo, Sendai, Hiroshima and Fukuoka) are significant destinations of inter-prefectural migration within the regions where they are located. Therefore, the four prefectures having these regional capitals reveal less negative net migration (−0.25 to 0%) for the three prefectures of Hokkaido, Miyagi and Hiroshima and slightly positive net migration (0–0.25%) for

126

Y. Ishikawa

Fig. 7.7  Net migration rates, prefectures of Japan, 2017. (Source: Annual Report on Internal Migration in Japan derived from the Basic Resident Registers (https://www.e-stat.go.jp/en/))

Fukuoka prefecture. Interestingly, Morikawa (2016) demonstrated that the positions of Osaka and Nagoya, in terms of internal migration, have declined in recent decades and become closer to the level of the regional capital cities. This pattern is clearly visible on the circular plot shown in Fig.  7.8. The overwhelmingly large volume of internal migration centred on the four prefectures in the Tokyo metropolitan area (Saitama, Chiba, Tokyo and Kanagawa) is clearly apparent, and the intra-metropolitan migration is also remarkable. As mentioned in the literature review, the importance of the Osaka area (Kyoto, Osaka, Hyogo and Nara) and the Nagoya area (Gifu, Aichi and Mie) has progressively declined since the period of high economic growth and this lower level of movement can be clearly seen in Fig. 7.8. Researchers have long shown a strong interest in understanding rural-to-urban and urban-to-rural migration flows as they play an important role in shaping patterns of settlement within a country. Japan, however, does not use ‘urban’ and ‘rural’ as a form of classification in its population census or resident registers. To circumvent the absence of such data and address limitations inherent in the conventional

7  Internal Migration in Japan

127

Fig. 7.8  Bilateral migration flows, prefectures of Japan, 2017. (Source: Annual Report on Internal Migration in Japan derived from the Basic Resident Registers (https://www.e-stat.go.jp/en/) (IMAGE-Asia Project). Note: Each prefecture is assigned a colour and flow arrows are assigned the same colour as the prefecture of origin. Arrow width indicates the relative size of the migration flows. The volume of the flows is indicated by the tick marks on the circumference of the plot)

rural/urban classification, we follow the approach adopted by Rees et  al. (2017) using population density as a proxy for urbanisation and plot the net migration rate against its logarithm, as shown in Fig. 7.9. Population weighted regression of net migration rates against the logarithm of population density is used to calculate the slope, which provides a summary indicator of the direction of population redistribution through net migration. The sign of the regression parameter indicates whether migration is taking place from predominantly rural to predominantly urban areas (less settled to more settled = a positive coefficient) or from urban to rural areas (= a negative coefficient), while the value of the parameter indicates the strength of the movement. The coefficient of 0.36 (R-Squared 0.65) indicates a moderate process of population concentration, whereby more urbanised prefectures with high population density show population gains, particularly Tokyo, while more rural prefectures show clear population losses. In the Asian context, Japan’s limited redistribution

128

Y. Ishikawa

Fig. 7.9  Net migration rates by log population density, prefectures of Japan, 2017. (Source: Annual Report on Internal Migration in Japan derived from the Basic Resident Registers (https:// www.e-stat.go.jp/en/) (IMAGE-Asia Project))

toward densely populated regions is similar to other highly urbanised countries such as Malaysia as well as countries with a smaller proportion of their population living in urban areas such as India (Charles-Edwards et al. 2019). Moderate population gains in urban areas are characteristic of countries at both early and late stages of the urban transition whereas larger gains tend to be observed in countries in the midst of their urban transition (Rees et  al. 2017). While some highly urbanised countries such as South Korea record flat or negative slopes indicative of net gains in favour of less densely settled areas, this is clearly not the case in Japan where large cities remain the main destinations.

7.8  Understanding Internal Migration in Japan In this section, we discuss and interpret these results to provide an understanding of the forces driving migration in Japan. The most significant of these forces is the economy, exemplified chiefly by fluctuations in the business cycle (see Ishikawa and Fielding 1998; Yano et al. 2000, 2003; Ishikawa 2011). When the economy was booming from the mid-1950s to the early 1970s and again in the late 1980s,

7  Internal Migration in Japan

129

population movement accelerated, but in periods of recession in the 1970s, it diminished. The business cycle in Japan has been linked to trends in unemployment, active opening ratio (explained below) and new housing construction (Ishikawa 2011). These factors can, to a degree, explain the changes in migration to the three major metropolitan areas (mainly the Tokyo area) as shown in Fig. 7.2. Important reasons for population gains in the core of the three major metropolitan areas (in particular, the Tokyo area) lie in their abundant employment opportunities and income differences. A well-known variable associated with employment opportunities is the active opening ratio, defined as the number of active jobs openings divided by the number of active applicants, indicating the balance of supply and demand in Japan’s labour market. These data are derived from public employment security offices. The value of this measure has been confirmed, for example, by the its close correlation with trends in net migration in the Tokyo area over the post-war period (Ishikawa and Fielding 1998; Regional Revitalisation Office 2014). Income differentials are also important for explaining internal migration in Japan. The prefectures in the three major metropolitan areas display high income levels. People born and raised in the periphery leave their hometowns for the core due to the income difference between the two. A significant relationship between net migration in the three major metropolitan areas (particularly the Tokyo area) and the income differential has been confirmed (Tabuchi 1988; Regional Revitalisation Office 2014). The influence of business fluctuations was underlined by the 2008 Global Financial Crisis. Its specific impact on human mobility, including internal migration, was investigated by Ishikawa (2011) who confirmed the influence of the crisis on internal migration of Japanese, although foreign residents suffered more than their Japanese counterparts. The role of demographic factors in driving Japanese migration is also noteworthy. The baby boom occurred in Japan in 1947–1949, but fertility dropped sharply in the 1950s and reached replacement level at the end of that decade. As a result, the size of the post-baby-boom cohorts was much reduced, and potential outmigrants from peripheral regions to the major metropolitan areas decreased (Ito 1984). This change, in close combination with restructuring of the labour market across both industrial and occupational dimensions, caused the migration turnaround in Japan (Ishikawa 1999, 2001b). In the early 1970s, it was believed that future fertility rates would remain stable. Against this expectation, fertility began to decline from the mid-1970s; this is often interpreted as a second demographic transition. The Total Fertility Rate (TFR) has ranged between 1.40 and 1.45 in recent years and population aging has been underway, with population decline since in 2009. At the time of writing, Japan is experiencing one of the most marked levels of population decline in the developed world, and the country’s proportion of elderly is the highest in the world. This demographic situation has affected many spheres of Japanese life including internal migration. The fall in the migration intensity, as shown in Fig. 7.3, can be explained in part by the decreasing proportion of the population who are of

130

Y. Ishikawa

labour-­force age, since this group are typically more mobile than the rapidly expanding group of retirees. One of the most notable features of the Japanese migration system in the past two decades is the remarkable in-migration to central parts and out-migration of young adults and the ageing/death of old residents from the suburbs of major metropolitan areas, including Tokyo. This is mainly caused by the weakening of suburbanisation arising from the population decline (Esaki 2006; Koike 2015, 2017). The preference for a good environment is a third factor influencing internal migration in Japan. Retirement migration from the cities to peripheral regions, mentioned earlier, is a typical example. Previous research has sought to elucidate whether such migration is popular. Retirement migration at age 60–65 is not common, as Fig. 7.4 reveals. Nevertheless, at a prefecture/municipality scale, retirement migration to many peripheral areas is evident from analysis of age-specific net migration rates using 2010 census data, and most of this can be attributed to environment-­oriented movement (Ishikawa 2016). It seems that a superior living environment has become a magnet that attracts people. The countryside (den’en) has also been attracting younger migrants, generally those in their 30s and 40s (Odagiri 2014). Favoured destinations are places with lower population density. It is not possible to determine the precise numbers involved but retirement migration and migration to the countryside serve to reduce the mono-polar concentration into Tokyo. The East Japan Great Earthquake, the resulting tsunami, and subsequent accident at Tokyo Electric Power’s Fukushima Nuclear Power Station in 2011 had a significant influence on internal migration in the country. The damage caused by these events exerted the single greatest influence on internal migration in the post-war period. Particularly important was the migration of displaced residents from the damaged coastal municipalities of the three prefectures most affected (Iwate, Miyagi and Fukushima) to inland municipalities. The number of refugees immediately following the disaster reached 470,000; in July 2019 some 50,000 were still displaced from their homes (Reconstruction Agency 2019). The inclusion of a question about address change in the 2015 census suggests the enormity of this disaster and there is a considerable literature on the disaster itself. In terms of internal migration, themes such as the places of refuge of the victims and their return to places of previous residence merit further investigation (e.g., Isoda 2011a, b; Ishikawa 2012; Oda 2013; The Tohoku Geographical Association 2018). The data needed to elucidate the impact of the 2011 disaster are available from the population census and the resident registers. Damaged areas are now on the way to revitalisation, and thus a detailed and comprehensive examination using the available data should be made. Finally, it is useful to highlight the ‘business bachelor’ (tanshin funin) as a popular category of intra-organisational transfer migration that represents a distinctive feature of Japan’s migration landscape (Ishikawa 1994; Fielding 2016, p. 244). In Japan, when employees of organisations are ordered to transfer to other locations distant from their current addresses, they (mostly male heads of household) often move alone to the area of the new posting, leaving their wives, children and elderly

7  Internal Migration in Japan

131

parents at their current places of residence. This is because migration that takes along all family members may have negative effects on the children’s school education and on the physical and mental health of the elderly. While such transfer migration is well known among Japanese people, the existing literature from a geographical perspective is limited due to the lack of comprehensive data.

7.9  Impacts and Implications This section discusses the consequences of internal migration. The focus is the impact on origins and destinations, and on policy responses in regard to urbanisation and the decline of peripheral regions. Since post-war period, internal migration has been intense with conspicuous urbanisation in the three major metropolitan areas expanding their spatial extent. Fielding (2016, p.  228) terms this situation ‘hyper-urbanisation’. As a result, the population share of the three areas increased. From 1965 to 2015 the population of the Tokyo area as a share of the national population grew from 21.2% to 28.4%, the Osaka area from 14.0% to 14.4%, and the Nagoya area from 8.1% to 8.9%. Thus, the share of the three major areas increased from 43.3% to 51.8%. The rapid population increase in the Tokyo area during the past half century is particularly noteworthy and with its 36.1 million residents, Tokyo is one of the world’s largest metropolitan areas. However, Tokyo area will soon start to experience population decline. A substantial problem in the area is the increasing number and proportion of older people, most of whom arrived from peripheral regions. Related to this, the possible rapid increase in the social security budget, particularly in suburban municipalities, is an issue of great concern (Inoue and Watanabe 2014; Nakazawa 2017). This situation is also leading to restructuring of the housing market (Sato et al. 2018). In contrast, the majority of peripheral prefectures and municipalities, which were the main origins of internal migration, particularly towards the Tokyo area, have suffered serious depopulation, aggravated both by low fertility and ageing (Sakuno 2011). Their rapid depopulation will continue throughout this century and aggravate economic decline. To address these difficulties, most peripheral prefectures and municipalities are eager to attract young people. For example, it is not unusual for a student approaching graduation, who came from a peripheral prefecture to study at a university in the Tokyo area, to be eagerly invited to return to the prefecture of origin after graduation by staff of the prefectural office. Policies for mitigating the problems of urbanisation in the core and depopulation in the periphery have been implemented (Hama 1982; Sakuno 2011) but have failed to reverse these trends. To relieve the difficulty of peripheral population decline, a new section (machi hito shigoto sosei honbu) was established within the Cabinet Office of the Japanese government in September 2014 but there has been no decline in the Tokyo area’s net migration gain in recent years (Fig. 7.2). This suggests the need for internal migration researchers to become actively involved in

132

Y. Ishikawa

policymaking to mitigate the current problem of spatial population disparity (Masuda 2015; Morikawa 2016; Ishikawa 2018; Nakazawa 2018). Meanwhile, retirees and seniors aged 55–75 have a tendency to leave for peripheral municipalities where living costs are generally cheaper. These migration flows are not widely known but should be supported by, for example, raising awareness of their significance and promoting policies to increase them (Ishikawa 2016). Although no overall trend to counter-urbanisation has yet been identified in Japan (Yamagami 2003), a policy to support core-to-periphery migration is desirable for well-balanced development. For example, the introduction of a policy to direct new immigrants to peripheral regions deserves examination, since foreign nationals tend to concentrate in major metropolitan areas. Such policies have been executed in Canada and Australia since the 1990s, and successful results for Manitoba Province and for the State of South Australia have been reported (Carter et al. 2008; Hugo 2008). Japanese policymakers should seriously consider those results (Ishikawa 2018).

7.10  Conclusions This chapter has presented the historical and current situation of internal migration in Japan, examined a series of national and regional migration indicators and described the determinants and consequences of internal migration, with the aim of placing the Japanese experience in an Asian or global context. It adds new findings to our current understanding of Japan’s internal migration. Distinctive features of the country’s internal migration derived from the analysis and their implications can be summarised as follows. First, migration intensity in the country has been decreasing, probably due to population ageing combined with the decreasing proportion of the population who are of working age. The nationwide MEI for inter-prefectural migration is one of the lowest in the world. However, results from the 2015 census show that this is due to a combination of periphery-to-core flows by young adults in their twenties being offset by core-to-periphery flows among people in their thirties and forties. Secondly, with respect to migrant attributes, the sex selectivity (male-dominated) of inter-prefectural migration is still apparent. Contemporary opinion suggests that migration of young females (aged 20–39) will play a key role in population redistribution in the near future (Masuda 2015). This non-traditional perspective and its implications need to be actively investigated. Third, the mono-polar concentration into Tokyo, particularly observed for the young adult population, has been underway since the 1980s, and the country has therefore continued to experience urbanisation. Meanwhile, most peripheral prefectures have suffered severe depopulation. This expanding spatial disparity is undesirable in terms of well-balanced development of the national territory, so it should be rectified in an efficient manner. In this context, polices such as the support of retirement migration and other forms of migration to the countryside, with people leaving

7  Internal Migration in Japan

133

the major metropolitan areas and their high population densities for peripheral regions, seem promising. There are two major concerns related to better understanding of internal migration in Japan. The first of these is the increased incidence of non-response to questions in the population census. If this continues to increase, data quality will further deteriorate. A second concern is the delay in exploring the impact on internal migration of the East Japan Great Earthquake and the accident at Tokyo Electric Power’s Fukushima Nuclear Power Station. Damaged areas are now on the way to recovery but more than eight years after the later an estimated 50,000 people have still been unable to return to their homes and are forced to live as evacuees. Detailed and comprehensive investigation of the post-disaster migration is needed. Acknowledgements  I would like to thank Aude Bernard for quickly preparing the files used in this paper and making helpful comments on an earlier draft. Appreciation is also owed to Takashi Inoue, Masato Shimizu, Shiro Koike, and Yuzuru Isoda for their help with document retrieval.

References Bell, M., & Hugo, G. (2000). Internal migration in Australia: 1991–1996 overview and the overseas-­born. Canberra: Department of Immigration and Multicultural Affairs. Bell, M., Charles-Edwards, E., Ueffing, P., Stillwell, J., Kupiszewski, M., & Kupiszewska, D. (2015). Internal migration and development: Comparing migration intensities around the world. Population and Development Review, 41(1), 33–58. https://doi. org/10.1111/j.1728-4457.2015.00025.x. Bernard, A., & Bell, M. (2018). Educational selectivity of internal migrants: A global assessment. Demographic Research, 39, 835–854. Bernard, A., Bell, M., & Charles-Edwards, E. (2014). Life-course transitions and the age profile of internal migration. Population and Development Review, 40(2), 213–239. Carter, T., Morrish, M., & Amoyaw, B. (2008). Attracting immigrants to smaller urban and rural communities: Lessons learned from the Manitoba provincial nominee program. Journal of International Migration and Integration, 9(2), 161–183. Charles-Edwards, E., Bell, M., Bernard, A., & Zhu, Y. (2019). Internal migration in the countries of Asia: Levels, ages and spatial impacts. Asian Population Studies, 15(2), 150–171. Esaki, Y. (2006). Future image of the Tokyo metropolitan area population: Population geography of the city centre and the suburbs. Tokyo: Senshu University Press. (In Japanese) Fielding, A. J. (2016). Asian migrations: Social and geographical mobilities in Southeast, East and Northeast Asia. Abingdon: Routledge. Fielding, A. J. (2018). Internal migration trends and processes since the 1950s. In A. Champion, T. Cooke, & I. Shuttleworth (Eds.), Internal migration in the developed world: Are we becoming less mobile? (pp. 173–202). Abingdon: Routledge. Fielding, A.  J., & Ishikawa, Y. (2003). Migration and the life course in contemporary Japan. Geographical Review of Japan, 76(12), 246–257. Hama, H. (1982). Regional analysis of Japan’s population structure. Tokyo: Chikura Shobo. (In Japanese) Hanaoka, K., Ishikawa, Y., & Takeshita, S. (2017). Have destination choices of foreign residents contributed to reducing regional population disparity in Japan? Analysis based on the 2010 population census microdata. Population, Space and Place, 23(1), e1975. https://doi. org/10.1002/psp.1975.

134

Y. Ishikawa

Hayashi, R. (2014). International comparison of migration: A construction of model-mobility using Japanese indicators. Jinko Mondai Kenkyu (Journal of Population Problems), 70(1), 1–20. (In Japanese) Hirai, M. (2007). Regional characteristics of inter-prefectural migration of the elderly population. In Y. Ishikawa (Ed.), Population decline and regional imbalance: Geographical perspectives (pp. 129–147). Kyoto: Kyoto University Press. (In Japanese) Hugo, G. (2008). Australia’s state-specific and regional migration scheme: An assessment of its impacts in South Australia. Journal of International Migration and Integration, 9(2), 125–145. Inoue, T., & Liaw, K.-L. (2004). Life-course perspective on some distinctive features of migration from non-metropolitan areas in Japan. Geographical Review of Japan, 77(12), 765–782. Inoue, T., & Watanabe, M. (2014). Population aging in the Tokyo metropolitan area. Tokyo: Hara Shobo. (In Japanese) Ishikawa, Y. (1978). Internal migration in post-war Japan. Chirigaku Hyoron (Geographical Review of Japan), 51(6), 433–450. (In Japanese) Ishikawa, Y. (1994). An introductory study of transfer migration. NUPRI Research Paper Series, 64, 1–34. Ishikawa, Y. (1999). Contribution of the demographic factor to the migration turnarounds in Japan, Sweden and Canada. International Journal of Population Geography, 5(1), 1–17. https://doi. org/10.1002/(SICI)1099-1220(199901/02)5:13.0.CO;2-G. Ishikawa, Y. (2001a). Migration turnarounds and schedule changes in Japan, Sweden and Canada. Review of Urban and Regional Development Studies, 13, 20–33. Ishikawa, Y. (Ed.). (2001b). Studies in the migration turnarounds. Kyoto: Kyoto University Press. (In Japanese) Ishikawa, Y. (2008). Progress in Japanese population geography: Retrospect and prospect. Geographical Review of Japan, 81(5), 247–261. https://doi.org/10.4157/grj.81.247. Ishikawa, Y. (2011). Impact of the economic crisis on human mobility in Japan: A preliminary note. Belgian Journal of Geography, 2011(3&4), 129–147. Ishikawa, Y. (2012). Displaced human mobility due to March 11 Disaster. http://tohokugeo.jp/ articles/e-contents29.pdf. Accessed 28 Feb 2019. Ishikawa, Y. (2016). Internal retirement migration revisited. Kyoto Daigaku Bungakubu Kenkyu Kiyo (Memoirs of the Faculty of Letters, Kyoto University), 55, 135–166. (In Japanese) Ishikawa, Y. (2018). New immigration and Japan: Solution to population decline. Otsu: Kaisei-­ sha. (In Japanese) Ishikawa, Y., & Fielding, A. J. (1998). Explaining the recent migration trends of the Tokyo metropolitan area. Environment and Planning A, 30(10), 1797–1814. Ishikawa, Y., & Liaw, K.-L. (2009). The 1995–2000 inter-prefectural migration of foreign residents of Japan: Salient features and multivariate explanation. Population, Space and Place, 15(5), 401–428. https://doi.org/10.1002/psp.521. Isoda, Y. (2011a). Fukushima diaspora: Age and sex of evacuees from the nuclear crisis. http:// tohokugeo.jp/articles/e-contents12.html. Accessed 28 Feb 2019. Isoda, Y. (2011b). Refuge life of evacuees from the Fukushima nuclear accident: Analysis of Okuma town survey based on human capital theory. http://tohokugeo.jp/articles/e-contents26. html. Accessed 28 Feb 2019. Ito, T. (1984). Recent trends of internal migration in Japan and potential lifetime out-migrants. Jinko Mondai Kenkyu (Journal of Population Problems), 172, 24–38. (In Japanese) Kawabe, H. (1991). Migration in developing countries. Tokyo: Ajia Keizai Kenkyu-sho. (In Japanese) Koike, S. (2015). A comparative analysis on the population re-centralization of metropolitan areas in Japan’s prefectural capitals. Keikaku Gyosei (Planning and Public Management), 38(2), 45–52. (In Japanese) Koike, S. (2017). Demographical analysis of ‘re-centralization’ trend in Tokyo ward area. Jinkogaku Kenkyu (Journal of Population Studies), 53, 23–45. (In Japanese) Koike, S., & Yamauchi, M. (2014). Unreported migration data in the 2010 population census of Japan. Jinko Mondai Kenkyu (Journal of Population Problems), 70(3), 325–338. (In Japanese)

7  Internal Migration in Japan

135

Kuroda, T. (1976). Structure of Japanese population in transition. Tokyo: Kokon Shoin. (In Japanese) Laukova, D., Bernard, A., & Sigler, T. (2020). Settlement and migration patterns of immigrants by visa class in Australia. In K. Kourit, B. Newbold, K. NijKamp, & M. Partridge (Eds.), The economic geography of cross-border migration. Basingstoke: Palgrave Macmillan. Liaw, K.-L. (2003). Distinctive features of the sex ratio of Japan’s inter-prefectural migration: An explanation based on the family system and spatial economy of Japan. International Journal of Population Geography, 9(3), 199–214. https://doi.org/10.1002/ijpg.274. Masuda, H. (2015). Disappearance of Tokyo: Care, bankruptcy and migration to periphery. Tokyo: Chuo Koron Shinsha. (In Japanese) Morikawa, H. (2016). An overview of Japanese urban systems based on an analysis of internal migration in 2010 and regional development policies. Jimbun Chiri (Japanese Journal of Human Geography), 68(1), 22–43. (In Japanese) Nakazawa, T. (2008). Spatial trajectory of occupational careers: Life course of research and development engineers and information engineers. Okayama: Daigaku Kyoiku Shuppan. (In Japanese) Nakazawa, K. (2017). Welfare-induced migration of the elderly and long-term care insurance system. Shakai Hosho Kenkyu (Journal of Social Security Research), 2(2&3), 332–348. (In Japanese) Nakazawa, T. (2018). Towards politico-economic population geographers: A critique of shukushou nippon-no shougeki (the shock of a shrinking Japan). Keizai Chirigaku Nenpo (Annals of the Japan Association of Economic Geography), 64(3), 165–180. (In Japanese) Nanjo, Z., Kawashima, T. & Kuroda, T. (1982). Migration and settlement:13. Japan, RR-82-85. Laxenburg: International Institute for Applied Systems Analysis. National Institute of Population and Social Security Research. (2017). Population projections for Japan, 2016–2065. http://www.ipss.go.jp/pp-zenkoku/j/zenkoku2017/pp29_ReportALL.pdf. Accessed 19 Apr 2019. (In Japanese) National Institute of Population and Social Security Research. (2018). Report of the eighth national survey on migration (2016). http://www.ipss.go.jp/ps-idou/j/migration/m08/ido8report.pdf. Accessed 28 Feb 2019. (In Japanese) Oda, T. (2013). The geographical space of evacuation: A record of field experience and practices from the multiple disasters of 2011  in Northeast Japan. Shirin (Journal of History), 96(1), 167–207. (In Japanese) Odagiri, T. (2014). Farm and mountain villages do not disappear. Tokyo: Iwanami Shoten. (In Japanese) Okada, M. (1976). Population U-turn and Japanese society. Tokyo: Taimeido. (In Japanese) Otomo, A. (1998). Twenty years of regional demography in Japan. Jinkogaku Kenkyu (Journal of Population Studies), 23, 55–66. (In Japanese) Raymer, J., & Baffour, B. (2018). Subsequent migration of immigrants within Australia, 1981–2016. Population Research and Policy Review, 37(6), 1053–1077. Reconstruction Agency. (2019). The number of refugees in Japan. https://www.reconstruction.go.jp/topics/main-cat2/sub-cat2-1/20190710_hinansha.pdf. Accessed 16 Aug 2019. (In Japanese) Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), 1–22. https://doi.org/10.1002/ psp.2036. Regional Revitalisation Office. (2014). Supplementary Document. http://www.kantei.go.jp/jp/ singi/sousei/meeting/souseikaigi/h26-09-19-sankou.pdf. Accessed 28 Feb 2019. (In Japanese) Rogers, A., & Castro, L. J. (1986). Migration. In A. Rogers & F. J. Willekens (Eds.), Migration and settlement: A multiregional comparative study (pp. 157–208). Dordrecht: D. Reidel. Sakuno, H. (2011). Geographies of peripheral Japan. In Y. Ishikawa, T. Inoue, & Y. Tahara (Eds.), The population geography of contemporary Japan (pp.  99–106). Tokyo: Kokon Shoin. (In Japanese)

136

Y. Ishikawa

Sato, H., Shimizu, C., & Karato, K. (2018). Role of secondhand condominiums in creating a walkable city: Case study in the Tokyo metropolitan area in the 2000s. Jimbun Chiri (Japanese Journal of Human Geography), 70(4), 477–497. https://doi.org/10.4200/jjhg.70.04_477. (In Japanese) Shimizu, M. (2011). Internal migration. In Y. Ishikawa, T. Inoue, & Y. Tahara (Eds.), The population geography of contemporary Japan (pp. 29–35). Tokyo: Kokon Shoin. (In Japanese) Shimizu, M. (2017). Net in-migration of foreigners and net out-migration of Japanese at the municipal level. E-journal GEO, 12(1), 85–100. (In Japanese) Tabuchi, T. (1988). Inter-regional income differentials and migration: Their interrelationships. Regional Studies, 22, 1–10. The Tohoku Geographical Association. (2018). The 2011 Great East Japan Earthquake and geography. Sendai: Sasaki Shuppan. (In Japanese) Vining, D. A., & Pallone, R. (1982). Migration between core and peripheral regions: A description and tentative explanation of the patterns in 22 countries. Geoforum, 13, 339–410. Wiltshire, R. (1979). Research on migration in Japan: (I) reverse migration and the concept of U-Turn. The Science Reports of the Tohoku University, 7th series, Geography, 29(1), 63–68. Wiltshire, R. (1992). Inter-regional personnel transfers and structural change: The case of the Kamaishi steelworks. Transactions of the Institute of British Geographers, 17(1), 65–79. Yamagami, T. (2003). The relation between population size and population increase in metropolitan areas in Japan. Jinkogaku Kenkyu (Journal of Population Studies), 33, 73–83. (In Japanese) Yano, K., Nakaya, T., & Ishikawa, Y. (2000). An analysis of inter-municipal migration flows in Japan using GIS and spatial interaction modeling. Geographical Review of Japan, 73B(2), 165–177. https://doi.org/10.4157/grj1984b.73.165. Yano, K., Nakaya, T., Fotheringham, A. S., Openshaw, S., & Ishikawa, Y. (2003). A comparison of migration behaviour in Japan and Britain using spatial interaction models. International Journal of Population Geography, 9(5), 419–431. https://doi.org/10.1002/ijpg.297.

Chapter 8

Internal Migration in Cambodia Jean-Christophe Diepart and Chanrith Ngin

8.1  Introduction From the early 1970s until the late 1990s, internal population movements driven by war and conflicts were central in the life of most Cambodians. In the first half of the 1970s, civil war and bombing displaced people from their hometowns. During the Khmer Rouge (KR) regime (1975–1979), cities were literally emptied and most of the rural population were forceably moved out of their villages to live and work in the countryside (Kiernan 1996; Vickery 1984). An estimated two-thirds of the population were involved in migration to rural areas during the early KR period (Desbarats 1995). After the KR regime was toppled in 1979, masses of Cambodians made their way across the country – usually returning to their home villages – in search of relatives and land. But ensuing war and political upheavals continued to displace people within and across national borders. Large-scale international migration to Vietnam and Thailand soared during periods of crisis and were followed by periods of return migration. When the country achieved peace in 1998, internal migration did not stop, but the nature of mobility changed. The liberalisation and growth of the economy opened new avenues for the development of services and industries, which triggered structural transformation from a primarily rural and agricultural-based economy to one that was increasingly urban, industrial and service-based. These opportunities have lured people to work in urban areas and incentivised migration from rural areas (MoP 2012). However, the capacity of non-agricultural sectors to create [unskilled] J.-C. Diepart (*) Gembloux Agro-Bio Tech, University of Liège, Liège, Belgium e-mail: [email protected] C. Ngin Development Studies, School of Social Sciences, Faculty of Arts, The University of Auckland, Auckland, New Zealand e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_8

137

138

J.-C. Diepart and C. Ngin

jobs remains limited and employment growth lags behind growth of the economically active rural population (Diepart 2016). This limitation lies at the core of the labour question in contemporary Cambodia and, combined with development constraints in core agricultural areas, it has resulted in the resurgence of internal and international labour migration (Diepart et al. 2014). The magnitude of internal migration is significant. In 2008, the number of lifetime migrants – using village as the spatial unit measuring migration – was 3,239,184 representing a migration rate of 24.2% for Cambodia as a whole (NIS 2010). This is more than the total number of people who have been involved in international migration in their lifetime (NCDD 2017), even though cross-border migration has increased substantially over the last decade. Despite its significance, studies of internal migration in Cambodia are scant. This impedes policymaking to address the role internal migration plays in the redistribution of population across the territory, and the resulting socioeconomic transformation of both rural and urban regions. To address this gap, we focus on internal migration by analysing data from the 1998 and 2008 censuses. Our objective is twofold. We first examine internal migration quantitatively with metrics that describe the intensity of migration, the demographic characteristics of migrants, and the spatial impacts of internal migration in Cambodia. We do so by situating Cambodia in a broader Asian and global context. Second, we aim to understand the contribution of internal migration to development, particularly with regard to the contemporary labour market.

8.2  Internal Migration Data During the prolonged period of war and political instability that stretched from the early 1970s to the late 1990s, the collection of demographic information was neglected because of internal security problems as well as the dearth of material and human resources (Desbarats 1995). Demographic data have been collected systematically since the 1998 population census. That census was a milestone because the migration-related questions asked in 1998 have been used in all subsequent national population censuses and surveys (Table 8.1). The nature of the migration data collected in Cambodia mirrors most other Asian countries. Both the 1998 and 2008 population censuses captured lifetime migration, which is the most common form of data collected on migration in Asia. Likewise, migration measured over a discrete period of one or five years, which is less frequent in Asia, is not captured in Cambodia (Charles-Edwards et al. 2019). However, both censuses included questions on each individual’s latest migration, which can be combined with information on duration of residence to generate measures of migration over a defined interval. In addition to the population censuses, inter-censal sample surveys were organised in 2004 and 2013 to update information on key demographic parameters. Both surveys included migration-related questions that are identical to those asked in the 1998 census.

8  Internal Migration in Cambodia

139

Table 8.1  Migration questions, 1998 Census, Cambodia Birthplace 9 Place of birth If in this village, enter code 1. If in another village, give name of district of that village and write name of province within brackets. If outside Cambodia, write name of country.

Previous residence 10 Where have you been living before? If always lived in this village, enter code 1 and skip to 13. If in another village, give name of district of that village and write name of province within brackets. If outside Cambodia, write name of country.

Duration of stay 11 How long have you lived in this village? (Enter Code from list below)

Reason for migration 12 Give reason for change of residence, if present residence is different from previous residence. (Enter Code from list below)

Source: NIS (2002)

In 2012, the Ministry of Planning conducted a comprehensive survey to examine the characteristics of migrants and the links between migration and the welfare of individuals, families and communities. The survey is rich in detail but limited to the population that has been streaming into Phnom Penh, the capital city of Cambodia, which is the main destination of rural-to-urban migrants (MoP 2012). In addition to census-like datasets collected and analysed by the National Institute of Statistics, local governments (at commune level) update a so-called ‘Commune Database’ annually, which comprises a large number of socioeconomic indicators, including job-related migration. The information, available at village level for rural and urban areas throughout the entire country, includes the number of people who have left their household for a period during the year (duration unspecified) to seek jobs elsewhere. Unlike previous data, the information about this migration does not imply a permanent change of residence but rather captures the mobility of labour over an extended period. As such, it provides information to complement the migration data obtained in censuses. However, the accuracy of the information recorded by village chiefs depends on their knowledge of the labour situation in their village. As a result, the headcount of job-migrants needs to be considered with caution. The main advantage is that the data are collected annually by the same person, which should therefore provide reliable information on trends. Bringing these sources together, Table 8.2 provides an overview of the data on internal migration available in Cambodia.

8.3  The Spatial Framework The spatial hierarchy of administrative entities in Cambodia has undergone considerable change since 1998, and there have also been changes in the classification of these entities as rural or urban areas. In 1998, the country consisted of 24 provinces,

140

J.-C. Diepart and C. Ngin

Table 8.2  Internal migration data collected in Cambodia, 1998 onwards

Type of datasets Information collected

Geographic coverage

1998 and 2008 Census

2004 and 2013 Intercensal survey Lifetime migration Last-move migration Previous residence (district) Duration of residence in current location Reason for migration National National

2012 One-off survey

2011 to 2016 Commune database

Demographic and socioeconomic characteristics of migrants to Phnom Penh

Number of people involved in job-­ related internal migration

Phnom Penh

National

of which Phnom Penh had the special status of capital. Each province was divided into districts (n = 183), communes (n = 1609) and villages (n = 13,406) (NIS 2002). At that stage, there was no functional classification of ‘urban areas’. The designation of places as ‘urban’ or ‘rural’ was based only on administrative criteria: all provincial towns (whole district), four districts in the capital city of Phnom Penh, and the entire provinces of Sihanoukville, Kep and Pailin (also known as krong or municipalities) were considered as ‘urban’. There were a number of problems with this because the purely administrative classification did not take into account the urban or rural conditions observed on the ground. To address the problem, in 2004 the National Institute of Statistics (NIS) adopted an additional set of criteria consistent with the Cambodian context, which classifies an area as ‘urban’ if it has a population density greater than 200 people/km2, a total population of more than 2000 people and less than 50% of employment is in agriculture. This reclassification was conducted at the commune level because villages in Cambodia do not have administrative boundaries (NIS 2012). In 2008, the law on the Administrative Management of the Capital, Provinces, Municipalities, Districts and Khans, and the sub-decrees relating to the application of this law, reorganised the sub-national administration systems (Royal Government of Cambodia 2008a). Sub-decree 18 (Royal Government of Cambodia 2008b) divided the capital city of Phnom Penh (equivalent to a province) into eight Khans (equivalent to districts) and upgraded the status of Sihanoukville, Kep and Pailin from municipalities to provinces. Including the capital Phnom Penh, the total number of provinces remained at 24. Each province consisted of districts (n = 159) and cities (also known as municipalities) (n = 26). In 2011, the NIS revised the classification of ‘urban area’ to take the 2008 census results into account. Consequently, areas declared as ‘urban’ included the communes (sangkat) of Phnom Penh and of the 26 cities (also known as krong) identified in subdecree 18 as well as any other communes that met the criteria set out above. In addition, some semi-urban communes close to major cities were reclassified as ‘urban’ based on local knowledge and field visits. Figure 8.1 represents the spatial framework of administrative entities in Cambodia in 2011 and the classification of communes as ‘urban’ and ‘rural’.

8  Internal Migration in Cambodia

141

Fig. 8.1  Provinces, districts and urbanisation, Cambodia, 2011. (Source: Mapping by the authors)

Changing numbers of spatial units at various levels of the hierarchy is a common issue in developing countries, particularly in a post-conflict context. In this case, it is problematic because it is difficult to compare the results of the 1998 and 2008 censuses at district and provincial levels. In both the 1998 and 2008 censuses, people are considered to be migrants if the village in which they were enumerated differs from their birthplace (life-time migrants) or from their previous place of residence. However, the location of the birthplace and place of the previous residence is coded only at district level, so the district is the lowest level of the hierarchy at which out-migration can be measured. As with many other countries, this limits the analysis that can be undertaken because it does not match the classification of urban and rural areas, which are identified at commune level. In order to compute net migration rates at district level and identify the different migration streams, all districts identified as municipalities or having a population density higher than 200 people/km2 were considered as urban areas.

8.4  Prior Research The literature on internal migration in Cambodia consists of a rich collection of case studies that examine particular aspects or geographies of the phenomenon. These studies are insightful but, with the notable exception of the thematic reports of national population censuses that provide a complete statistical description of the

142

J.-C. Diepart and C. Ngin

process, they fail to document the issue in a comprehensive manner. Migration studies that are framed by key national development issues, such as the labour question, are absent from the policy research landscape. Studies of rural-to-urban migration comprise a large part of the research effort. One outstanding piece of research was a detailed examination of mobility from rural areas to Phnom Penh, which is the main destination for rural-to-urban migrants. Comprehensive in scope, the study shows that in 90% of the villages investigated, the total population fell as a result of out-migration (MoP 2012). Migrants to Phnom Penh are mainly young adults seeking economic and employment opportunities. As a result, the capital’s population more than doubled between 1998 and 2010 – from 570,000 to 1,240,000 people – 80% of whom were lifetime migrants. The migrant population in Phnom Penh is young (median age of 25). Women migrate somewhat more frequently than men (MoP 2012). About 30% of migrants are women aged between 15 and 30. This is driven by the dominance of the garment industry in the city (and its outskirts) and in the national economy, which chiefly employs female workers. In 2008, the garment sector employed 41% of female migrants to the capital (NIS 2009). Women are less likely to migrate for education than for work. According to the 2008 census (NIS 2009), 25% of male migrants were students compared with only 17% of female migrants. Women are also less likely to migrate if they have children (MoP 2012). According to the 2008 census (NIS 2009), 51% of the migrant population had engaged in rural-to-rural migration, while those undertaking rural-to-urban migration accounted for only 28% of the total. However, despite its significance, little is known about rural-to-rural migration. One reason for this is that these migrants are far less visible and are much more diffused throughout the country. However, the few case studies that focus on these migrations have suggested that they contribute significantly to the redistribution of the population within the national territory (Maltoni 2010), particularly from lowland areas to upland frontier regions, fuelling the formation of post-forest agrarian systems (Chheang and Dulioust 2012; Diepart et al. 2014). However, not much is known about the intensity, characterisation, and spatial patterns of this particular stream of internal migration. Qualitative, small-scale research on internal migration provides insights into the cycles, needs and networks of migrants. Parsons and Lawreniuk (2016) assert that internal migration is not always a lengthy, one-off phenomenon, but consists, to a significant extent, of smaller cycles lasting from a few weeks (for begging migrants) to a few months or years (for construction and garment workers). They show that the mean period of stay for migrants has fallen significantly during the past half-decade (Parsons and Lawreniuk 2016) and suggest that rural and urban areas have become linked by more nuanced systems of movement and remittances. In addition to economic incentives, rural-urban links are induced by changing patterns of social relations (Kheam and Treleaven n.d.; Lim 2007) and cultural norms (Czymoniewicz-Klippel 2013; Parsons and Lawreniuk 2016) partly influenced by women’s economic empowerment thanks to their employment in the garment industry (Cuyvers et al. 2009; Derks 2008).

8  Internal Migration in Cambodia

143

On migratory networks, Parsons (2017) argues that poor households lack the networks needed to undertake the initial migration, and that this affects them throughout the migration process. Even if poor household members find employment in an urban area, the long-term benefits of their labour are difficult to sustain because their jobs are usually temporary. In such circumstances, rural assets and income are helpful and sometimes vital to support a migrant’s livelihood. There is no real physical disconnect between migrants and their parents or family members in their home village. Contacts are maintained through the sending of food (rice and dried fish) to migrants, through migrants visiting their hometown during special occasions such as Khmer New Year and Pchum Ben,1 and through return visits to hometowns to help with rice planting and harvesting, and income transfer. Remittances sent by migrant workers to their families represent an increasingly important part of the income stream in rural areas. Females are likely to remit larger amounts and a higher share of their earnings to their rural households (MoP 2012). The level of indebtedness is important in Cambodia, and debt management is a critical mechanism that supports the decision to migrate, particularly in rural areas where livelihoods are highly dependent on agriculture. In rural Cambodia, people borrow money to cover upfront costs for agricultural production but also to cover a number of other ‘non-productive’ expenditures (education, food, ceremonies, transportation etc.). However, income from agriculture has become increasingly vulnerable due to climate hazards and market fluctuations (Bylander 2013a, 2014; Oeur et  al. 2012). The risk of over-indebtedness is reinforced when expenditure is incurred for medical care. When over-indebtedness becomes unmanageable, migration is seen as a coping strategy. Debt as a push factor has been identified in both internal and international migration (Bylander and Hamilton 2015; Chan 2009; Deelen and Vasuprasat 2010). These studies show that a significant part of the remittances sent by migrant workers is spent on debt repayment. Qualitative, interview-based research in the Siem Reap province by Bylander (2013b, 2017) found that loans were sometimes used to fund the migration of a household member. More often, household members sought loans for ‘non-productive’ purchases with the expectation that these would be repaid by their own or a family member’s plan to migrate for employment.

8.5  How Much Movement? Migration Intensity To make sense of the intensity of internal migration in Cambodia, it is useful to step back and examine its evolution since the fall of the Khmer Rouge regime. Calculated using the latest migration data from the 2008 census, Fig. 8.2 presents the annual number of people involved in internal migration since 1979, as indicated by the data

 An important religious festival during which Cambodians pay their respect to deceased relatives.

1

144

J.-C. Diepart and C. Ngin

Numer of migrants (thousands)

350 300 250

4

1

5

200

2 150

3

100 50

1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

0 Year of last move

Fig. 8.2  Year of last move, residents of Cambodia, 2008. (Source: 2008 Census. Note: for legend see text)

on duration of residence collected at the 2008 census. As such, it does not include migrants who had passed away or moved overseas before 2008. Since only the date of the last move is captured, it may also understate movement intensities for earlier years by migrants who have made multiple moves. However, Fig. 8.2 provides an overall impression of recent internal migration history as it highlights years with peak migration, which are specifically related to important socio-political events. Because the Khmer Rouge forcibly displaced most of the Cambodian population from 1975 to 1979, the fall of the regime in early 1979 initiated country-wide return migration (1). In the late 1980s, the conjunction of the Vietnamese withdrawal and introduction of a more liberal constitution created the conditions for increased population movement (2). The period 1991–1993 corresponds to the return of refugees from Thai border camps (3) and 1998 marks the reintegration of the Khmer Rouge army and guerrilla forces (4). Starting in 2001, the liberalisation and structural transformation of the economy and the promulgation of a more liberal Land Law intensified migratory movements that increased steadily until 2008 (5). The volume and mobility of internal migration (lifetime and latest-move) in 1998 were relatively high as a result of the historical events outlined above. As shown in Table 8.3, between 1998 and 2008, the total number of internal migrants (lifetime and latest-move) increased to 3.2  million and 3.4  million, respectively. However, due to population growth, the intensity of migration decreased; lifetime migration intensity fell from 25.3% to 24.2% and the latest-move intensity declined from 29.6% to 25.8%. Although migration intensity decreased by only a small margin, the mobility of the population remained quite high. This can be attributed to movement from 2001 onwards in the context of post-war reconstruction and the

8  Internal Migration in Cambodia

145

Table 8.3  Number of migrants and internal migration rate, inter-village moves, Cambodia, 1998 and 2008 Total population Lifetime

Latest-move data (no defined duration)

Internal migrants (persons) Internal migration rate (percent) Internal migrants (persons) Internal migration rate (percent)

1998 11,437,656 2,890,117 25.3

2008 13,395,682 3,239,184 24.2

3,387,140 29.6

3,457,228 25.8

Source: NIS (2010) Table 8.4  Migration intensities by type of move, Cambodia, 1998 and 2008 Time interval Lifetime

Indicators ACMI CMI – Inter-district CMI – Inter-province Five-year ACMI CMI – Inter-district CMI – Inter-province Ratio lifetime/five-year

1998 N-A 18.0 (n = 149) 11.6 (n = 24) 18.4 7.3 (n = 149) N-A 2.4 (inter-district)

2008 N-A N-A 13.6 (n = 24) 16.1 7.7 (n = 184) 5.3 (n = 24) 2.5 (inter-province)

Source: NIS (2002) and NIS (2009) Notes: CMI Crude Migration Intensity, ACMI Aggregate Crude Migration Intensity, N-A not available

rapid economic transformation of the country. Between 1998 and 2008, the proportion of people who had changed residence within the previous decade rose, while the proportion with longer residence durations fell (NIS 2010). Bell et al. (2002) have argued that cross-national comparison of migration intensity is best achieved using the Aggregate Crude Migration Intensity (ACMI), which estimates all permanent changes of residence within a country, irrespective of the distance moved. For countries that do not collect this information directly, the ACMI can be determined using the procedure proposed by Courgeau et al. (2012), which involves fitting a linear regression equation to Crude Migration Intensities (CMI) generated at different spatial scales (See Chapter three of this volume). ACMI values calculated using the IMAGE studio offer a reliable basis to situate internal migration in Cambodia within the broader Asian and global context (Table 8.4). In 1998, lifetime inter-district migration intensity, which captures the cumulative migration history of the population, is higher than the five-year inter-district migration. Likewise, in 2008, lifetime inter-provincial migration intensity is higher than the five-year migration intensity at the same level of geography. In other words, the influence of historical movements on migration continues to prevail in the migration signature of Cambodia. However, the ratio of lifetime to five-year intensity exhibits lower values than those of many other Asian countries, although it is similar to those of China, Mongolia, Kyrgyzstan and Myanmar (Charles-Edwards et al. 2019). In

146

J.-C. Diepart and C. Ngin

the wider Asian context this suggests that the current internal migration intensity in Cambodia is relatively high in relation to historical movements. In 1998 and 2008, the ACMIs for five-year data were 18.4 and 16.1 respectively, close to the Asian mean (16.5) (Charles-Edwards et al. 2019). However, these values were lower than the global mean (21.0), and well below the values recorded in countries of the New World (Canada, New Zealand and the United States), where geographical mobility remains high despite declines tied to completion of the urban transition, economic maturation and population ageing (Bell and Charles-­ Edwards 2014). The five-year ACMI decreased between 1998 (18.4) and 2008 (16.1). Comparison of migration intensities at district level is affected by changes in the spatial framework between 1998 and 2008 but it is notable that the CMI rose only marginally from 7.3% to 7.7%, despite a marked rise in the number of districts, suggesting that the underlying migration intensity actually fell. However, at the provincial level, values are comparable and indicate that between 1998 and 2008, the lifetime CMI increased from 11.6% to 13.6% (Table 8.4). Altogether, this suggests that migration intensities declined between 1998 and 2008, but changes in residence now occur over longer distances than in the past. This increase reflects post-conflict economic development and the progressive urbanisation of the population, and also relates to agrarian development, particularly in frontier areas (see below).

8.6  Who Moves? The Characteristics of Migrants The age structure of migration in Cambodia follows a pattern observed elsewhere in Asia. The propensity to migrate rises quickly and peaks in young adulthood, then steadily declines and rises again at age 80 or above (Fig. 8.3). This pattern is consistent for both inter-provincial and inter-district migration. Like most other Asian countries (particularly those in Southeast Asia), the migration age profile of Cambodia is best described as ‘early and concentrated’ (Charles-­ Edwards et al. 2019). Five-year migration intensity peaks among 22 year olds for inter-province migration and 23  year olds for intra-province migration. This is slightly earlier in the life course of individuals than the global average. Figure 8.3 also reveals the concentration of migration at peak ages: the normalised migration intensity at the peak is 0.5 standard deviations higher than the global mean (Charles-­ Edwards et al. 2019). This concentration of migration at an early age can be attributed to the eagerness of young people to grasp job opportunities when they enter the labour market. It is also underpinned by different life-course events, such as marriage or the arrival of a first child, which are encouraged by social norms in Cambodia that support a quick transition to adulthood in the early twenties. However, it is becoming increasingly common for a young married couple to cohabit with their in-laws until they have sufficient means to form a new independent family unit (Heuveline 2017). These pragmatic adjustments also influence the migration decisions of young couples.

8  Internal Migration in Cambodia

147

0.030

Crude migration intensity (normalised)

Inter-district (n=184) Inter-province (n=24)

0.025

0.020

0.015

0.010

0.005

0.000

5

10

15

20

25

30

35

40

45

50

55 60 Age

65

70

75

80

85

90

95 100

Fig. 8.3  Age-specific migration intensities, Cambodia, 2003–2008. (Source: calculated from the 2008 census)

The age structure of migration shows important differences by sex and between rural and urban areas. For the country as a whole, the five-year migration rate was 9.34% but it was higher for males (9.95) than for females (8.77). The dominance of male migration is manifested in the 25–49 age group, corresponding to the core working age. Migration is more important for females in the 15–19 age group than for males, corresponding to the difference in high school dropout. Although the profile is similar, the in-migration rate in urban areas is distinctly higher than in rural areas. In urban areas, the in-migration rate (at ages 20–24) peaks at 40.1%, but is higher for females (41.4%) than for males (38.5%). In contrast, at ages 25–54 the rate is higher for males. In rural areas, the in-migration rate peaks at ages 20–24 at just 11.0% but is higher for men in all age groups (NIS 2010). To shed light on these differences, we present the age and sex composition of all migrants currently living in Phnom Penh (an urban area) and Pailin (a rural area), which are the two provinces with the highest in-migration rates in the country (NIS 2010). In 2008, the five-year migration rates in Phnom Penh and Pailin were, respectively, 32.0% and 31.4%. In Phnom Penh, the female and male migration rates were, respectively, 32.8% and 31.2%. The age structure of migrants (Fig. 8.4) is typically ‘early and concentrated’ but is also characterised by the dominance of women in the 15–24 age group, while men dominate the 25–54 age group. In contrast, the age structure of migrants in Pailin follows a sinusoidal pattern (Fig. 8.5). In the main, the migration rate of men (32.1%) is higher than that of women (30.6%). In addition to a first peak at a young age (20–24), with men dominating from the age of 15 to 49, there is a secondary peak at age 6569. The mobility of older people is sometimes

148

J.-C. Diepart and C. Ngin 50.0 Male Female

In-migration rate (%)

40.0

30.0

20.0

75 +

70-74

65-69

60-64

55·59

50-54

45-49

Age

40-44

35-39

30-34

25-29

20-24

15-19

5·9

0-4

0.0

10-14

10.0

Fig. 8.4  Inter-village moves to Phnom Penh by age and sex, 2003–2008. (Source: NIS 2010)

a result of return-migration but in most instances it is due to the tied migration of parents with their adult children. Educational attainment is low in Cambodia. In 2008, 41% of people aged 15 and over had completed education at primary level or above. Only 3.1% had completed education at secondary level or above. This characteristic is broadly reflected in the educational attainment of life-time migrants but education among the migrant population is slightly higher, with more people having completed lower secondary or secondary school level, or above (Table 8.5).

8.7  Where Do They Move? Spatial Patterns The internal mobility of people in Cambodia is multi-directional. Inter-district migrants (latest move) who left a rural area to travel to another rural area in the five years to 2008 represent more than half the total number of migrants (Fig. 8.6). This stream is significantly more important than the rural-to-urban flow (27.5%), which is typically associated with the urban transition. The counter-flow of people – from urban to rural areas – is less than a quarter of this value (6.5%). Redistribution of population from one urban area to another also occurs but at a much lower level (15.1%). Overall, most migration is rural-to-rural flow, but the net redistribution is towards urban areas, a characteristic Cambodia shares with other countries of mainland Southeast Asia (Charles-Edwards et al. 2019).

8  Internal Migration in Cambodia

149

50.0 Male Female

In-migration rate (%)

40.0

30.0

20.0

75 +

70-74

65-69

60-64

55·59

50-54

45-49

40-44

35-39

30-34

25-29

20-24

15-19

5·9

0-4

0.0

10-14

10.0

Age

Fig. 8.5  Inter-village moves to Pailin by age and sex, 2003–2008. (Source: NIS 2010) Table 8.5  Educational attainment, total population and life-time migrants aged 15 and over, 2008 Level of education No education/illiterate Primary not completed Primary Lower secondary Secondary or above Other Total

Men Total (%) 16.7 30.1 27.4 21.4 4.3 0.1 100

Migrants(%) 14.4 27.6 25.5 24.4 7.9 0.2 100

Women Total(%) 31.4 32.7 20.9 13.0 2.0 0.0 100

Migrants(%) 29.1 31.6 20.3 15.2 3.7 0.1 100

Source: NIS (2010)

The Migration Effectiveness Ratio (MERRU)2 is a simple metric that captures the balance between rural-to-urban flows and counter-flows. The value of this index in Cambodia is 40.9%, which is relatively high when compared with other Asian countries. It shows a net balance in favour of urban areas but also suggests that Cambodia is at an early stage in its urban transition (Charles-Edwards et  al. 2019). This is certainly true given the very low rate of urbanisation in Cambodia: 18.6% in 1998 and 19.5% in 2008 (NIS 2009). In fact, the MERRU offers a useful window through which to examine internal migration, that is, in relation to the flow and counter-flow between urban and rural areas. However, rural-to-rural and urban-to-urban flows are 2  MERRU = 100 × (MRU − MUR)/(MRU + MUR) where MRU represents migration flows from rural to urban areas and MUR re-migration flows from urban to rural

150

J.-C. Diepart and C. Ngin Rural to rural

Female

Rural to urban

Urban to rural

26%

53%

All

28%

51% 0%

10%

20%

6%

30%

48%

Male

Urban to urban

30%

40%

50%

60%

70%

16%

7%

14%

6%

15%

80%

90%

100%

Fig. 8.6  Composition of migration streams, Cambodia, 2003–2008. (Source: NIS 2009)

not factored into MERRU. And for countries where these flows are important, the MERRU does not fully capture the dynamics of internal migration. This is the case in Cambodia where rural-to-rural flows are of paramount importance. Net Migration Rates (NMRs)3 were calculated with five-year migration data derived from the 2008 census. As indicated earlier, NMRs are only available at district level. As such, they do not capture intra-district movements, which can be important in the Cambodian context. In aggregate, the Migration Effectiveness Index (MEI)4 – a measure of the overall effect of migration in redistributing population  – is 49.69. This suggests that, overall, migration is a fairly efficient mechanism for population redistribution (Bell et  al. 2002). The Aggregate Net Migration Rate (ANMR),5 which measures the redistributive impact of migration is 3.8%. In Cambodia, 109 of the 184 districts registered a negative NMR (with a mean of −5.48) with just 75 displaying a positive value (mean = 10.09). The geography of Net Migration Rates is distinctive (Fig. 8.7). A large number of districts – typically located in the rural lowland central area of the country (around the Tonle Sap Lake) – form a large region clearly characterised by a net outflow of migrants. On the periphery of this central area, a large number of districts form an arc-shape region characterised by an inflow of migrants. Some areas display strong net gains. Scattered throughout the national territory, some districts registered positive NMRs; these are primarily major urban centres, including Phnom Penh (PP). The relationship between the NMR and population density sheds light on this spatial distribution. For Cambodia, Charles-Edwards et al. (2019) identified a strong relationship between population density and the net migration rate, suggesting a high level of migration from low-density to high-density regions. However, they note that internal migration in Cambodia – like that in many other Asian countries – has been driven by forces more complex than the urban transition pathways  NMR = [(Inflow-Outflow)/population × 100]  MEI = 100 iDi − Oi/iDi + Oi where Di represents total inflows in zone i and Oi represents total outflows in zone i 5  ANMR = CMP × MEI 3 4

8  Internal Migration in Cambodia

151

Fig. 8.7  Net migration rates, districts of Cambodia, 2003–2008. (Source: 2008 population census, data computation and mapping by the authors. Note: Stars (∗) indicate main urban centres)

theorised by Rees et al. (2017) who take population density as a proxy for urbanisation. Indeed, the situation is more intricate in Cambodia. The relationship between population density and NMR suggests that different migration dynamics co-exist in the Kingdom (Fig. 8.8). Four groups of districts can be readily identified, each characterised by a particular pattern of internal migration: • Districts with low population density and very high positive NMRs are rural and located in the peripheral upland region • Districts with moderate population density and negative NMRs are mostly rural and located in the lowland central region • Districts with high population density and high levels of NMR are all districts (khan) of the capital city of Phnom Penh • Districts in which population density is lower than in Phnom Penh but higher than rural lowland and upland districts. These have a range of NMR values. Those with positive NMRs are important secondary cities such as Siem Reap, Battambang and Krong Preah Sihanouk, while those with negative NMRs are urban centres in less economically dynamic and mainly rural provinces such as Pursat, Svay Rieng, Takeo and Kratie. Figure 8.8 shows that not all districts fall easily into these four groups, but outliers are few in number.

152

J.-C. Diepart and C. Ngin 50 40

Rural upland districts

Net migraon rate (%)

30

Phnom Penh

Secondary cities

20 10 0 -10 Rural lowland districts

-20 -30

0.0

0.5

1.0

1.5

2.0 2.5 3.0 Log populaon density

3.5

4.0

4.5

5.0

Fig. 8.8  Net migration rates by log population density, districts of Cambodia, 2003–2008. (Source: 2008 Census. Data computation by the authors)

Combined with Fig. 8.7, Fig. 8.8 brings a much more differentiated view of internal migrations in Cambodia than would be suggested sole by reference to the urban transition model. Migration to urban areas (Phnom Penh and some secondary cities) is undeniably intense and represents movements that accompany the urbanisation and industrialisation of the country. It seems that this flow of migration is fed by flows from the central rural lowland districts. However, comparing Cambodia with other Asian countries only on that basis would be misleading because it only captures a fragment of the picture. Far less visible, though relatively more important, is migration to less densely populated rural upland regions. This migration flow is typically a movement to resource frontiers, as discussed in the next section. By connecting origin and destination districts, the main patterns of mobility become very explicit (Fig. 8.9). The flows that link rural lowland to rural upland districts, Phnom Penh and secondary cities clearly stand out. Migration flows from rural lowland and upland districts to secondary urban centres are also very distinctive, as are the moves from secondary cities to Phnom Penh. To illustrate the links between origin and destination regions, we focus on the cases of Phnom Penh (urban area) and Pailin (rural area), which are the two provinces with the highest migration rates in the country. Migrants to Phnom Penh come from every corner of the country, but this migration conforms to a basic ‘gravity’ model with the main source being neighbouring provinces with large populations, most notably in the southeast of the country and in the rural lowland districts identified above (Figs. 8.9 and 8.10). Figure 8.9 also depicts an urban-to-urban flow with a significant number of migrants moving from secondary urban centres. In a similar way, migrants to Pailin were drawn from all corners of the country. Figure  8.11 reveals at least three major migration flows: (i) movements between

8  Internal Migration in Cambodia

153

Fig. 8.9  Bilateral migration flows, rural and urban areas of Cambodia, 2003–2008. (Source: Calculated from the 2008 Census (IMAGE-Asia Project)

districts within Pailin, (ii) a flow from neighbouring districts; and (iii) a flow of long-distance migrants from the southwest and central regions of the country. The last of these flows, from rural lowlands to rural uplands, represents a key feature of internal migration in Cambodia (Fig. 8.8).

8.8  Understanding Internal Migration in Cambodia Cambodia is one of Asia’s poorest countries, but it has witnessed sustained growth over the past two decades. Amidst a challenging global economic environment, the annual growth in Gross Domestic Product (GDP) between 2006 and 2016 was 6.9%. Although its relative share of GDP has declined, agriculture is still a central

154

J.-C. Diepart and C. Ngin

Fig. 8.10  Origin of migrants living in Phnom Penh, 2003–2008. (Source: 2008 population census, data computation and mapping by the authors)

Fig. 8.11  Origin of migrants living in Pailin, 2003–2008. (Source: 2008 population census, data computation and mapping by the authors)

8  Internal Migration in Cambodia In search of employment

Female

29%

Male

30%

0%

10%

Move of family

155 Transfer of work place

Marriage

44%

5%

29%

20%

30%

40%

12%

50%

Education

60%

7%

12%

70%

80%

Visiting

Other

4%

4%

7%

7%

5%

90%

5%

100%

Fig. 8.12  Reason for last move by sex, Cambodia, 2008. (Source: NIS 2010)

pillar of the economy. It represents 26.7% of the country’s GDP (World Bank 2017) and continues to provide the main employment for 50% of the population aged over 18  years (68.8% if both primary and secondary occupations are considered). However, labour productivity is low in agriculture and the search for employment is cited by most migrants as the main reason for their mobility (Fig. 8.12). Inundated rice cultivation is the cropping system best suited to the agro-­ecological conditions of the central lowland plain around the Tonle Sap Lake and in the Mekong Delta, a region that corresponds to the rural lowland districts identified in Figs. 8.8 and 8.9. Historically, this is where the Cambodian population has been concentrated and this area is the most densely populated rural region in the country. A factor underlying low labour productivity in the agricultural sector is the decline in the size of landholding per household, which is associated with demographic pressure in this lowland central region. In fact, the expansion of agricultural landholdings at the periphery of this region is no longer possible even though this has been a historical trend in the life of Cambodian peasants. This process is now constrained by law and by the privatisation of the commons. Land access is further complicated by a process of land commoditisation that neo-liberal land reform has exacerbated through the promotion of land titling, micro-credit and land markets, which are substantially wealth-biased and exacerbate land concentration (Diepart and Sem 2018b). A land squeeze is affecting a growing number of agricultural households in this region. In 2011, agricultural landlessness affected 29% (Phann et al. 2015), and 47% of households had less than 1 ha to cultivate, which is far from sufficient to sustain a household’s livelihood. In reaction to this land squeeze, a significant number of people have migrated to the cities in search of employment. This movement has been particularly pronounced among young women who have been recruited to work in the booming garment sector in Phnom Penh and its outskirts; this sector employed 23% of 2003–2008 migrants to urban areas. The other main categories of migrant occupation are street vendors (16%), construction workers (7%), and transport workers (5%). The dominant feature of migrant employment in urban areas is the predominance of wage labour in occupations that require relatively low skills and levels of education (13%).

156

J.-C. Diepart and C. Ngin

A recent study suggests that by 2030 the annual increase in the rural labour force will be approximately 140,000 (Diepart 2016), which is lower than the annual increase of 221,000 that occurred between 1998 and 2004, (Lundström and Ronnas 2006). Nevertheless, the transfer of unskilled labour from agriculture to industry and the tertiary sector will lag behind this increase in the active rural population because total job creation in the non-agricultural sectors remains limited. Diepart (2016) estimated that 40,000 unskilled jobs were created each year between 2008 and 2014, including jobs in both the industry and service sectors. Another study, commissioned by the International Labour Organisation, indicates that between 2004 and 2009 the industrial sector created 162,736 jobs (27,122 jobs per year) while the number of unskilled jobs in the service sector did not increase significantly during the same period (Kang and Liv 2013). For many farmers, the strategy to cope with this job shortage has been to migrate to upland regions in search of economic opportunities. As Figs. 8.6 and 8.7 show, this migration stream involved a large number of farmers and their families. In fact, 65% of those who migrated to rural areas between 2003 and 2008 were employed in agriculture. Most (91%) were working on their own farms, which clearly shows that rural-to-rural migration is primarily driven by the search for agricultural land. Wage labour is significant however (and the number of migrants who are primarily involved in farm and non-farm wage labour) has become proportionally more important since 2003, suggesting that the migrants have shifted to wage employment because less land is available for agricultural expansion. This feature echoes findings from the Northwest frontier area, which have shown that migrations are fluid and often involve multiple migration events. The arrival of migrants is motivated by the search for land, but migrants also turn to agricultural wages when the opportunity arises. To maximise family labour use in the lean season and cope with more constrained access to land, an increasing number of families rely on cross-­border migration to look for jobs in Thailand (Diepart and Sem 2018a; Pilgrim et al. 2012). Movements to rural uplands are voluntary and are not organised by the government. To a large extent, they can be seen as an expression of the agency of peasant households in responding to rural poverty. They are also the expression of an on-­ going response on the part of Cambodian peasants to use the principle of appropriation ‘by the plough’ as a legitimate mode of land acquisition, which has been a consistent trend throughout Cambodian agrarian history. Central State authorities have been aware of this situation and have not impeded it. They were probably happy to see spontaneous migration taking place because these movements were helping to solve poverty issues in the central plains, which the government was unable or unwilling to tackle. It also seems that migrant smallholder farmers have acted as the territorial spearhead of the State in helping to stabilise the peripheral margins of the country and consolidate the sovereignty of the State (De Koninck 1996). The State has also relied heavily on smallholder farmers to manage agrarian expansion across the country and to endorse their role in the production of cash crops that are heavily influenced by global markets (cassava, corn, soybean, etc.).

8  Internal Migration in Cambodia

157

8.9  Impacts and Implications To a large extent, the agricultural land appropriated by migrants in the upland regions comprises degraded forests classified as State land under the 2001 Land Law. The authorities have not publicised these movements as they have been inconsistent with land policy. Indeed, insofar as the Land Law forbade the acquisition of State land after 2001, these lowland-upland migrations are completely at odds with the legal framework for land that authorities were supposed to implement. This has resulted in a huge population living on land that they appropriated after 2001, in respect of which they have virtually no land tenure security under the 2001 Land Law. In a parallel and uncoordinated process, the government has granted large tracts of land and forests as agro-industrial concessions of up to 10,000 ha in the form of Economic Land Concessions (ELCs), and other agro-industrial concessions or mining concessions. The large-scale agricultural development model was expected to result in new types of investment and job creation in rural Cambodia, to stimulate agro-industrial activities requiring capital investment that the State could not provide, and to develop so-called ‘under-utilised’ land. Recent data show that following a comprehensive evaluation of all agro-industrial concessions in Cambodia, the number of projects at the time of writing was 255, covering a total area of 1.8 million ha. As far as mining concessions are concerned, there were 61 projects involving a total area of 0.8  million ha (Diepart and Sem 2018b). Adding to these large-scale agricultural land concessions, Protected Areas (7.5 million ha) complete the picture of State land management in the peripheral upland regions (Fig. 8.13). In fact, agricultural or mining concessions often occupy land that is already cultivated or used by smallholder farmers, resulting in an encroachment on farmland or pooled resources, thus having an adverse impact on farmers’ livelihoods. This has resulted in considerable conflict and tensions around questions of land access and control that occupy a central place in the political and social life of Cambodia. The figures released in relation to land conflicts differ because the methodologies and criteria used to calculate them are based on different definitions of conflict and rely on different sources of information. However, they all suggest that the magnitude of the problem is not small. During the period 2000–2013, land conflicts and resulting evictions affected 770,000 people (ADHOC 2014). According to data collected by LICADHO (2014), more than half a million people were affected by State-involved land conflicts between 2000 and 2014. Based on their monitoring of media sources and reports from network members, the NGO Forum on Cambodia (2015) reported that a total of 352 land disputes broke out between 1990 and 2014, of which 77% remained unresolved. A fundamental cause of land conflicts is that the genuine need for land by smallholder farmers has led to country-wide migration that is not adequately addressed in the current land reform. Another impact of labour migration is the shortage of labour at the peak of the cropping seasons (planting and harvest). To address this labour shortage, the main adaptation has been the mechanisation of agriculture, which further incentivises the increase in agricultural landholding and further exacerbates the problems of job creation.

158

J.-C. Diepart and C. Ngin

Fig. 8.13  Agro-industrial concessions, Protected Areas, hydropower dams, mining concessions and Special Economic Zones, Cambodia. (Source: Data computation and mapping by the authors)

8.10  Conclusions We have examined the intensity of migration, the demographic characteristics of migrants, and the spatial impacts of internal migration in Cambodia. We have also discussed the contribution of internal migration to development, particularly in addressing the contemporary labour conundrum. Internal migration has been related to socio-political events in the country, particularly wars, post-war reconstruction, and economic transformation. Lifetime migration intensity is higher than recent migration intensity. However, the ratio of lifetime migration to recent intensity is lower than that of other Asian countries. This suggests that the effect of historical migration events is still prominent in the country’s migration profile, but contemporary internal migration intensity in Cambodia is still moderately high. The propensity to migrate rises quickly and peaks in young adulthood, then declines steadily, rising again at the oldest ages. This ‘early and concentrated’ migration age profile is a common pattern among Asian nations. The age structure of migration also varies between the sexes and between rural and urban areas. The recent (five-year) migration rate is higher for men in the core working ages but higher for women of lower secondary or high school age. The migration rate into urban areas is substantially higher than that into rural areas and the peak is higher for women than for men.

8  Internal Migration in Cambodia

159

On the development front, internal migrations in Cambodia offer insights beyond the urban transition model. Migration to urban areas, mainly from the central rural lowland districts is strong, attracted by urbanisation and industrialisation. Job-­ seeking is the main reason for this migration flow. This is because agricultural labour productivity is low due to the fall in landholding per household, which is under pressure from population growth and wealth-biased land market dynamics in the lowland central region. To cope with the scarcity of land, a substantial number of people have migrated to the cities to find jobs, particularly young women who find employment in the garment sector in Phnom Penh and its outskirts. Comparing Cambodia with other Asian countries solely by reference to the urban transition would be incomplete because migration to less-populated, resource-rich rural upland regions is perhaps more strategically important. Over half of migrants to rural areas were employed in agriculture, and most of these were working on their own farms. This suggests that rural-to-rural migration is chiefly driven by the search for agricultural land. However, land for agricultural expansion has become scarcer, as is evident from the rising number of migrants involved in agricultural wage labour and non-farm employment. This is exacerbated by the increasing need for land sought by unskilled labourers; these labourers have been marginalised by the imbalance between growth in the labour force and the absorptive capacity of the non-agricultural sectors. This rural-to-rural migration is managed entirely by households. The State has allowed these spontaneous movements to occur because they were instrumental in reducing poverty in the central plains, in stabilising the peripheral margins of the country and managing the agrarian expansion in the uplands, especially the formation of new agrarian systems driven by global commodity markets. However, these movements have remained uncoordinated with the granting of large-scale land and mining concessions on upland territories. The overlay of land claims between migrant farmers and concessions has resulted in numerous land conflicts across the country. In order to promote more balanced development between rural and urban areas, a central challenge is to focus the discussion on internal migration when revising the legal framework, policies and practices of current land reform.

References ADHOC. (2014). Land situation in Cambodia 2013. Phnom Penh. Bell, M. and Charles-Edwards, E. (2014). Measuring internal migration around the globe: A comparative analysis. KNOMAD Working Paper 3/2014, The World Bank https://www.knomad. org/publication/measuring-internal-migration-around-globe-comparative-analysis Bell, M., Blake, M., Boyle, P., Duke-Williams, O., Rees, P., Stillwell, J., & Hugo, G. (2002). Cross-national comparison of internal migration: Issues and measures. Journal of the Royal Statistical Society, 165(3), 435–464. Bylander, M. (2013a). Depending on the sky: Environmental distress, migration, and coping in rural Cambodia. International Migration, 53, 135–147. https://doi.org/10.1111/imig.12087.

160

J.-C. Diepart and C. Ngin

Bylander, M. (2013b). The growing linkages between migration and microfinance. Migration Policy Institute. http://www.migrationinformation.org/Feature/print.cfm?ID=955 Bylander, M. (2014). Borrowing across borders: Migration and microcredit in rural Cambodia. Development and Change, 45(2), 284–307. https://doi.org/10.1111/dech.12080. Bylander, M. (2017). Migration disruption: Crisis and continuity in the Cambodian mass returns. International Migration Review, 52, 1130–1161. https://doi.org/10.1111/imre.12342. Bylander, M., & Hamilton, E.  R. (2015). Loans and leaving: Migration and the expansion of microcredit in Cambodia. Population Research and Policy Review, 34(5), 687–708. https://doi. org/10.1007/s11113-015-9367-8. Chan, S. (2009). Costs and benefits of cross-country labour migration in the GMS. Working Paper Series No. 44. Phnom Penh: CDRI. Charles-Edwards, E., Bell, M., Bernard, A., & Zhu, Y. (2019). Internal migration in the countries of Asia: Levels, ages and spatial impacts. Asian Population Studies, 15(2), 150–171. https://doi. org/10.1080/17441730.2019.1619256. Chheang, V. and Dulioust, J. (2012). Rural-rural migrations in Cambodia. Policy paper. Supreme National Economic Council and Agence Française de Développement. Courgeau, D., Muhidin, S. & Bell, M. (2012). Estimating changes of residence for cross-national comparison. Population-E, 67(4), 631–652. https://doi.org/10.3917/pope.1204.0631. Also published as Estimer les changements de résidence pour permettre les comparaisons internationales. Population-F, 67(4), 747–770. 10317/popu.1204.0747 Cuyvers, L., Reth, S., & Van de Bulke, D. (2009). The competitive position of a developing economy: The role of foreign direct investment in Cambodia. In D. Van de Bulke, A. Verbeke, & W. Yuan (Eds.), Handbook on small nations in the global economy: The contribution of multinational enterprises to national economic success. Cheltenham: Edward Elgar. Czymoniewicz-Klippel, M. T. (2013). Bad boys, big trouble. Subcultural formation and resistance in a Cambodian village. Youth & Society, 45(4), 480–499. De Koninck, R. (1996). The peasantry as the territorial spearhead of the state in Southeast Asia: The case of Vietnam. Sojourn: Journal of Social Issues in Southeast Asia, 11(2), 231–258. Deelen, L. & Vasuprasat, P. (2010). Migrant workers’ remittances from Thailand to Cambodia, Lao PDR and Myanmar: Synthesis report on survey findings in three countries and good practices. ILO/Japan Project on Managing Cross-border Movement of Labour. Derks, A. (2008). Khmer women on the move: Exploring work and life in urban Cambodia. Honolulu: University of Hawaii Press. Desbarats, J. (1995). Prolific survivors. In Population change in Cambodia, 1975–1993. Tempe: Arizona State University. Diepart, J.-C. (2016). They will need land! The current land tenure situation and future land allocation needs of smallholder farmers in Cambodia. In MRLG thematic study series #1. Vientiane: MRLG. https://doi.org/10.13140/RG.2.1.2877.2083. Diepart, J.-C., & Sem, T. (2018a). Fragmented territories: Incomplete enclosures and agrarian change on the agricultural frontier of Samlaut District, north-West Cambodia. Journal of Agrarian Change, 18(1), 156–177. https://doi.org/10.1111/joac.12155. Diepart, J.-C., & Sem, T. (2018b). The Cambodian peasantry and the formalisation of land rights: Historical overview and current issues (2nd ed.). Paris: Technical Committee on Land Tenure and Development. Diepart, J.-C., Pilgrim, J., & Dulioust, J. (2014). Migrations. In S.  C.’s. Wildlife (Ed.), Atlas of Cambodia: Maps on socio-economic development and environment (pp.  89–96). Phnom Penh: SCW. Heuveline, P. (2017). Households and family processes. In K.  Brickell & S.  Springer (Eds.), The handbook of contemporary Cambodia (pp.  336–345). London/New York: Routledge Handbooks. Kang, C., & Liv, D. (2013). Rural development and employment opportunities in Cambodia: How can a national employment policy contribute towards realization of decent work in rural areas?

8  Internal Migration in Cambodia

161

Bangkok: Thailand: International Labour Organization (ILO) Country Office for Thailand/ Cambodia and Lao People’s Democratic Republic. Kheam, T., & Treleaven, E. (n.d.). Women and migration in Cambodia: A further analysis of the Cambodian Rural-Urban Migration Project (CRUMP). Phnom Penh: UNFPA and National Institute of Statistics. Kiernan, B. (1996). The Pol Pot Regime: Race, power and genocide in Cambodia under the Khmer Rouge, 1975–1979. Bangkok: Silkworms. LICADHO. (2014). 2014 brings a new wave of Cambodian land conflicts. Phnom Penh: LICADHO licadho-cambodia.org Lim, S. (2007). Youth migration and urbanisation in Cambodia (Working Paper No. 36). Phnom Penh: Cambodian Development Resources Institute. Lundström, S., & Ronnas, P. (2006). Employment and growth in Cambodia – An integrated economic analysis (Country Economic Report). Stockholm: Swedish International Development Cooperation. Maltoni, B. (2010). Analyzing the impact of remittances from Cambodian migrant workers in Thailand on local communities in Cambodia. Phnom Penh: International Organization for Migration. MoP. (2012). Migration in Cambodia: Report of the Cambodian Rural Urban Migration Project (CRUMP). Phnom Penh: Ministry of Planning http://countryoffice.unfpa.org/cambodia/drive/ Rural-urbanMigrationinCambodiaReport2012_EngVersion.pdf NCDD. (2017). Commune database – 2016. Phnom Penh: National Committee for Democratic Development http://db.ncdd.gov.kh/cdbonline/ NGO Forum. (2015). Statistical analysis of land disputes in Cambodia, 2014. NIS. (2002). General population census of Cambodia 1998. Final census results (2nd ed.). Phnom Penh: National Institute of Statistics. NIS. (2009). General population census of Cambodia 2008: National report on final census results. Phnom Penh: Ministry of Planning – Royal Government of Cambodia. NIS. (2010). Cambodian demographic census. In Report 6 on migration. Phnom Penh: National Institute of Statistics, Ministry of Planning. NIS. (2012). Reclassification of urban areas in Cambodia, 2011. Phnom Penh: National Institute of Statistics and UNFPA. Oeur, I., Ang, S., & McAndrew, J. (2012). Understanding social capital in response to flood and drought: A study of five villages in two ecological zones in Kampong Thom Province. Engaging for the environment. In A.  Pellini (Ed.), Engaging for the environment. The contribution of social capital to community-based natural resource management in Cambodia (pp. 60–83). The Learning Institute: Phnom Penh. Parsons, L. (2017). Multi-scalar inequality: Structured mobility and the narrative construction of scale in translocal Cambodia. Geoforum, 85, 187–196. https://doi.org/10.1016/j. geoforum.2017.07.027. Parsons, L., & Lawreniuk, S. (2016). The village of the damned? Myths and realities of structured begging behaviour in and around Phnom Penh. The Journal of Development Studies, 52(1), 36–52. https://doi.org/10.1080/00220388.2015.1056787. Phann, D., Phay, S., Tong, K., & Pon, D. (2015). Landlessness and child labour in Cambodia. Phnom Penh: Cambodia Development Resource Institute. Pilgrim, J., Ngin, C., & Diepart, J.-C. (2012). Multiple migrations, displacements and land transfers at Ta Kream in Northwest Cambodia. In S. B. Hecht, S. Kandel, & A. Morales (Eds.), Migration, rural livelihoods and natural resource management (pp.  33–56). El Salvador: International Development Research Centre (IDRC) of Canada, Ford Foundation, Fundación PRISMA. Rees, P., Bell, M., Kupiszewski, M., Kupiszewska, D., Ueffing, P., Bernard, A., Charles-Edwards, E., & Stillwell, J. (2017). The impact of internal migration on population redistribution: An international comparison. Population, Space and Place, 23(6), 1–22. https://doi.org/10.1002/ psp.2036.

162

J.-C. Diepart and C. Ngin

Royal Government of Cambodia. (2008a). Law on administrative management of the capital, provinces, municipalities, districts and khans. Phnom Penh. Royal Government of Cambodia. (2008b). Sub-decree on the number of councilors of Phnom Penh municipality, provinces, municipalities, districts and khans for the 1st mandate. 18 ANK/ BK. Phnom Penh. Vickery, M. (1984). Cambodia: 1975–1982. Chiang Mai: Silkworms Books. World Bank. (2017). World Bank Open Data: Statistics on Cambodia. https://data.worldbank.org/ country/cambodia?view=chart. Accessed 5 March 2018.

Chapter 9

Internal Migration in Myanmar Maxime Boutry

9.1  Introduction This chapter explores internal migration in the Union of the Republic of Myanmar. Myanmar gained its independence from Britain in 1948. Since 1962, Myanmar has been under direct or indirect military control, most notably during the period 1962–2010 when it was ruled by military junta. Since 2010, the country has undergone a political transition towards liberal democracy, although armed conflict persists in parts of the country. Like many countries in Asia, Myanmar has made rapid progress through the demographic transition. Life expectancy at birth increased from 36 years between 1950 and 1955 to 65 years between 2010 and 2015 (UN 2019). The total fertility rate (TFR) declined from a high of six during the 1960s to just above replacement level, with a TFR of 2.25 by 2015 (UN 2019). Myanmar shows great regional diversity both in terms of ethnic composition and levels of socioeconomic development. There are 135 ethnic groups recognised by the government, each with distinct settlement patterns. The largest group are the Bamars comprising 68% of the total population. Other sizeable groups notably comprise the Shan, Rakhine, Kayin, Kayah, Kachin and Mon after which ethnic States are named. The highest levels of human development are found in the centre of the country – mainly around the cities of Yangon, Naypyidaw and Mandalay – while poverty is highest in eastern parts of Shan, Rakhine and Chin States (Burke et al. 2017). Research on internal migration in Myanmar has been growing in recent years, following a prolonged period of low activity due to poor data availability and political complexities. The 2014 population census was the first since 1983, and there were no large-scale data collections on internal migration in the intervening 30-year period. Between 1983 and 2014, population counts were estimated by the Central Statistical M. Boutry (*) PALOC − Patrimoines Locaux et Gouvernance, Institut de Recherche pour le Développement (IRD), Marseille, France e-mail: [email protected] © Springer Nature Switzerland AG 2020 M. Bell et al. (eds.), Internal Migration in the Countries of Asia, https://doi.org/10.1007/978-3-030-44010-7_9

163

164

M. Boutry

Organisation (CSO), drawing on data from the 1983 census and the 1991 Myanmar Population Changes and Fertility Survey. In 2012, official estimates put the national population at 61 million (Spoorenberg 2015). The 2014 census revealed the country had 51.5 million inhabitants, 15% fewer than the previous estimate, suggesting that the extent of fertility decline and emigration had been significantly underestimated. This chapter provides a detailed account of current internal migration levels and spatial patterns in Myanmar based on data from the 2014 Census. After reviewing the types of migration data and geographical framework used to collect data, the chapter summarises prior research on internal migration in Myanmar. In subsequent sections, the intensity, selectivity and spatial patterns of migration are analysed in turn. The chapter concludes by examining the impacts and implications of current and prospective migration trends.

9.2  Internal Migration Data The census is the principal source of data on internal migration in Myanmar, with the 2014 census of population being the first conducted for more than 30  years. According to the United Nations Population Fund (UNFPA), which served as the lead technical agency in support of the Myanmar Ministry of Immigration and Population (MOIP), nearly 98% of the population was counted by the 2014 census.1 In conflict-affected areas, the Ministry, now called the Ministry of Labour, Immigration and Population, reached informal agreements with various armed groups to allow these areas to be included.2 However, census enumeration did not cover some population groups in the northern parts of Rakhine State3 and in a few areas of Kachin State. In parts of Hpa Pun township in Kayin State, only the total number of households and population by sex were submitted to the Census Office. To minimise the effect of non-enumerated households, an estimate of 2,342,953 people was added to the overall census population (Department of Population and UNFPA 2016a, b). The 2014 census collected information on nine facets of the population: demographic characteristics, social characteristics, migration, educational characteristics, economic characteristics, births and childhood deaths, identity cards, disability, housing conditions and household amenities. The census included eight questions relating to internal migration: township of birth, township of usual residence, township of previous residence, and the classification of these locations as urban or rural. The time since migration is captured through a question on duration in place of  http://myanmar.unfpa.org/en/about-census. Accessed 17 April 2018.  http://myanmar.unfpa.org/en/mapping-country. Accessed 17 April 2018. 3  Households self-identifying as ‘Rohingya’  – a term that is not recognised by the government under the field ‘ethnicity’ – were not counted. It was estimated that a population of 1,090,000 was likely not to have been counted during the enumeration in Rakhine State alone (Department of Population 2015). 1 2

9  Internal Migration in Myanmar

165

usual residence; coupled with place of previous residence it has been shown that this provides a close surrogate for a fixed interval measure of migration. The 2014 census was designed to capture both permanent and semi-permanent changes of residence. A criterion of 6  months is used to establish the place of usual residence. People who move on a temporary basis, that is for less than 6  months, are not counted as migrants.

9.3  The Spatial Framework Myanmar’s administrative geography has a four-tier hierarchy. The first tier consists of 15 regions and states, including Naypyidaw Union Territory, created in 2005 when the capital relocated to the centre of the country. The second tier consists of 74 districts (Fig. 9.1). The third tier is comprised of townships (n = 413), which are further divided into urban wards (n = 3071) and rural village tracts (n = 13,620) (fourth tier). The 2014 census measured internal migration as inter-township movements (Department of Population 2015), however, data shared by the census board only allow manipulation at the regional (n = 15) and district (n = 74) levels. This poses some issues for capturing intra-district migration and reduces the capacity to understand rural-to-urban migration, as the latter is defined at the level of ward (urban) and village tract (rural) (Department of Population and UNFPA 2016a). Even in Yangon Region, the economic capital and largest town in Myanmar, only two districts (East and West Yangon) can be considered fully urbanised based on their component geography. Some migration to districts in Yangon is, therefore, classified as rural-to-rural migration, despite a majority of people moving into the region for off-farm employment.

9.4  Prior Research Research on the patterns of internal migration in Myanmar has been hindered by a multi-decadal absence of nationally representative data, with a 31-year gap between the two most recent censuses. As a consequence, internal migration studies before 2014 were mostly small-scale and qualitative. However, these studies did provide important insights into multiple facets of internal migration, including the impact of economic liberalisation, gender and the role of mobility in sustaining rural livelihoods. For example, Chaw (2003) examined the migration of rural women to garment factories in Yangon. This study followed the opening of Myanmar’s economy in 1988 to foreign trade and investment that led to the expansion of export-oriented industry, including garment production. The majority of migrants came from the Ayeyarwaddy Region and migration flows between Ayeyarwaddy and Yangon remain significant. Similarly, Kusakabe and Zin Mar Oo (2007) studied women migrating to the Kachin borderland in response to the re-opening of several

166

M. Boutry

Fig. 9.1  Regions, states and districts of Myanmar

international borders in the 1990s (here with China) leading to ‘growing business and livelihood opportunities in Tachilek [that] have since attracted rural ethnic Bamar, or Burman, migrants from all over Myanmar’ (Kusakabe and Zin Mar Oo 2007). With respect to the role of mobility in rural livelihoods, Takahashi (1997)

9  Internal Migration in Myanmar

167

Table 9.1  Migration questions, 2014 census, Myanmar Question Place of birth

Response Township Urban or rural Place of usual residence Township Urban or rural Duration in place of usual residence (in years) Years Reason for movement to this township (usual residence) Employment/in search of employment Education Marriage Followed family Conflict Did not move Other Place of previous usual residence Township Urban or rural

undertook a field-based study of migration in rural Myanmar that revealed a significant pool of agricultural labourers constituting a ‘floating’ population. Similarly, Okamoto (2009) examined floating labour groups in Rakhine fishing communities, which revealed the economic motivations of both short-and long-term intra-rural migrants (Table 9.1). The political transition toward a quasi-civilian government led to a relaxation of restrictions on spatial mobility (Kyan Htoo and Aye Myintzu 2016). Before 2010 overnight stays had to be registered with the local General Administration Department authorities and, in many border regions, movements between townships required travel authorisation from local authorities. The relaxation of controls on internal migration, coupled with the rapid industrialiation of Myanmar’s economy, led to growing interest in quantifying migration, with labour migration a particular focus. Griffiths and Kyaw Zaw Oo (2015) explored the internal migration of formal sector workers, who represent about 2% of the labour force, using a large sample survey (n = 15,000). Results showed that formal sector workers were highly mobile, with 39% moving between states/regions and 49% moving between and within states/ regions for work. The two largest sources of formal sector workers were the Ayeyarwady Region and Dry Zone (Sagaing and Magway Regions), while Yangon was a major destination. The International Labour Organization undertook a survey of labour migration (both formal and informal) in six regions of Myanmar (Rogovin and International Labour Organization 2015). This research aimed to identify ‘patterns in internal labour migration, exploitive labour practices and human trafficking among 7,295 internal labour migrants’ (Rogovin and International Labour Organization 2015). The majority of respondents (94%) were aged 18 or older and 72% originated from the main ethnic Bamar group. The study revealed that internal labour migrants were more likely to come from households with five members, and had moved for

168

M. Boutry

economic reasons, with debt an important motivator. Among respondents who migrated between states/regions, migration from a rural to an urban area was more common while those who migrated within the same state/region were more likely to migrate between two rural areas. Bamar migrants were more likely to migrate to another one of the 14 states/regions than within their region, while migrants belonging to a registered ethnic minority were more likely to move within their own state. The World Bank and Enlighten Myanmar Research (2016) undertook a migration study in two regions of Myanmar – Ayeyarwaddy and Magway – which examined both labour-related and other forms of migration, combining quantitative surveys (800 households per region) and qualitative methods. The impetus to the study was the ‘structural transformation away from a rural agricultural economy toward a more urban, industrial and service-based economy’ and the perception of increased rural-to-urban migration identified in previous longitudinal studies. Results showed that one in four households in the Ayeyarwady region was directly impacted by migration, compared with one in five in Magway. The study revealed an increased probability of migration for individuals originating from landless households. In a similar vein Kyan Htoo and Aye Myintzu (2016) studied rural-to-­ urban migration among 1102 households in four townships (Kayan, Maubin, Nyaungdon, and Twantay) located close to Myanmar’s primate city, Yangon. The study concluded that, overall, landless households were slightly less likely to have migrants than those with land (14% of households versus 19%). Migration was largely rural-to-urban (90% of migrants) and a large majority (70%) of migrants worked in the manufacturing sector (shared equally between women and men). The Thematic Report on Migration and Urbanisation based on the 2014 census (Department of Population and UNFPA 2016a) was the first nationwide analysis of internal migration. This report found that 19.3% of individuals moved at least once within the country during their lifetime. It also revealed that a large proportion of movement within Myanmar focussed on Yangon, either taking the form of movement into Yangon or movement between districts within the region. A high proportion of recent migrants in Yangon were employed in manufacturing, including almost half of female migrants to North Yangon. The report also revealed that states bordering Thailand and China were attracting many migrants, due to the positive economic impacts of cross-border trade. In addition, findings suggest that ‘rather than serving as the first step for international migration, internal migration largely operates in a different set of households to international migration’. Another striking finding was that half of migration over the preceding five years occurred between urban areas, with rural-to-rural migration the second most significant flow. Only 10% of recent flows were from rural to urban areas. This brief review of the literature highlights a growing volume of research on internal migration in Myanmar. For a fuller exposition, readers are directed to the work of Gupta (2016) and Griffiths and Ito (2016). While the studies discussed in this section cover numerous aspects of internal migration in Myanmar, there remain gaps in understanding, particularly with respect to the effectiveness of migration in redistributing population within the country. The following sections endeavour to address these gaps by presenting a systematic analysis of migration intensity and its spatial impacts.

9  Internal Migration in Myanmar

169

9.5  How Much Movement? Migration Intensity An estimated 7% of Myanmar citizens changed their township of residence at least once in the five years leading up to the 2014 Census (Table  9.2), while 5.1% of individuals changed their district of residence and 3.8% changed their state or region. For lifetime migration, an estimated 19.3% of individuals moved at least once between townships, while 14.6% moved between districts and 10% moved between states or regions. The lack of historical data means it is not possible to track intensities over time. However, calculating the ratio of lifetime against five-year migration intensity ‘…provides some … insight into temporal trends in a form that is directly comparable across countries’ (Charles-Edwards et al. 2019). The ratio of lifetime to recent migration ranges from 2.6 for states/regions to 2.9 for districts. This ratio is relatively low by Asian standards, similar to that observed in countries such as China, Mongolia, Kyrgyz Republic and Cambodia. A low ratio suggests relatively high levels of recent migration compared to the past (Charles-Edwards et al. 2019). The overall level or rate of population movement is commonly captured by the Crude Migration Intensity (CMI), calculated as the proportion of the population who changed residence in a given interval. However, differences in the statistical geography over which migration is measured complicate comparisons of migration intensity between countries. The Aggregate Crude Migration Intensity (ACMI) overcomes this by estimating all changes of address. Where these changes are not captured directly, as in the case of Myanmar, the ACMI can be estimated using the approach devised by Courgeau et al. (2012), and this is the approach used here (see also chapter three of this volume). The ACMI provides further evidence of relatively high levels of recent migration in Myanmar when compared to other parts of Asia. This suggests that, over the five years prior to the 2014 Census, 17.3% of the Myanmar population changed their place of usual residence, either between or within states, districts or townships. This is marginally above the Asian average of 16.5% (Charles-Edwards et al. 2019), suggesting that Myanmar sits around the midpoint of the league table of migration intensities for the countries in Asia.

Table 9.2  Crude migration intensity by type of migration and spatial scale, Myanmar, 2014 All moves (ACMI) 17.3

Type of migration Recent (duration