Fertility, Family Planning and Population Control in China (Routledge Studies in Asia's Transformations)

Fertility, Family Planning, and Population Policy in China China’s one-child population policy, first initiated in 1979, has had an enormous effect on the country’s development. By reducing its fertility in the past two decades to less than two children per woman, and developing a family planning program focused heavily on sterilization and abortion, China has undergone a significant transition in status to a demographically developed country. Fertility, Family Planning, and Population Policy in China brings together contributions from leading scholars on such developments as family planning policy and contraceptive use, biological and social determinants of fertility, patterns of family and marriage, and China’s future population trends. As such it will be essential reading for academics, researchers, policy-makers, and government officials with an interest in China’s demography, fertility, and population policy. Dudley L.Poston, Jr is Professor of Sociology at Texas A&M University, College Station, Texas. Che-Fu Lee (deceased) was Professor of Sociology at the Catholic University of America, Washington, DC. Chiung-Fang Chang is Assistant Professor of Sociology at Texas A&M University, Kingsville, Texas. Sherry L.McKibben is Assistant Professor of Sociology at the University of Texas of the Permian Basin, Odessa, Texas. Carol S.Walther is a PhD Candidate at Texas A&M University, College Station, Texas.

Routledge Studies in Asia’s Transformations Edited by Mark Selden Binghamton and Cornell Universities, USA

The books in this series explore the political, social, economic, and cultural consequences of Asia’s transformations in the twentieth and twenty-first centuries. The series emphasizes the tumultuous interplay of local, national, regional, and global forces as Asia bids to become the hub of the world economy. While focusing on the contemporary, it also looks back to analyze the antecedents of Asia’s contested rise. This series comprises several strands: Asia’s Transformations aims to address the needs of students and teachers, and the titles will be published in hardback and paperback. Titles include: China in War and Revolution, 1895–1949 Peter Zarrow Confronting the Bush Doctrine Critical views from the Asia-Pacific Edited by Mel Gurtov and Peter Van Ness Japan’s Quiet Transformation Social change and civil society in the 21st century Jeff Kingston State and Society in 21st Century China Edited by Peter Hays Gries and Stanley Rosen The Battle for Asia From decolonization to globalization Mark T.Berger Ethnicity in Asia Edited by Colin Mackerras Chinese Society, 2nd edition Change, conflict and resistance Edited by Elizabeth J.Perry and Mark Selden

The Resurgence of East Asia 500, 150 and 50 year perspectives Edited by Giovanni Arrighi, Takeshi Hamashita, and Mark Selden The Making of Modern Korea Adrian Buzo Korean Society Civil society, democracy and the state Edited by Charles K.Armstrong Remaking the Chinese State Strategies, society and security Edited by Chien-Min Chao and Bruce J.Dickson Mao’s Children in the New China Voices from the Red Guard Generation Yarong Jiang and David Ashley Chinese Society Change, conflict and resistance Edited by Elizabeth J.Perry and Mark Selden Opium, Empire and the Global Political Economy Carl A.Trocki Japan’s Comfort Women Sexual slavery and prostitution during World War II and the US occupation Yuki Tanaka Hong Kong’s History State and society under colonial rule Edited by Tak-Wing Ngo Debating Human Rights Critical essays from the United States and Asia Edited by Peter Van Ness Asia’s Great Cities aims to capture the heartbeat of the contemporary city from multiple perspectives emblematic of the authors’ own deep familiarity with the distinctive faces of the city, its history, society, culture, politics, and economics, and its evolving position in national, regional, and global frameworks. While most volumes emphasize urban developments since the Second World War, some pay close attention to the legacy of the longue durée in shaping the contemporary. Thematic and comparative volumes address such themes as urbanization, economic and financial linkages,

architecture and space, wealth and power, gendered relationships, planning and anarchy, and ethnographies in national and regional perspective. Titles include: Singapore Wealth, power and the culture of control Carl A.Trocki Representing Calcutta Modernity, nationalism and the colonial uncanny Swati Chattopadhyay Hong Kong Global city Stephen Chiu and Tai-Lok Lui Shanghai Global city Jeff Wasserstrom Beijing in the Modern World David Strand and Madeline Yue Dong Bangkok Place, practice and representation Marc Askew Asia.com is a series which focuses on the ways in which new information and communication technologies are influencing politics, society, and culture in Asia. Titles include: The Internet in Indonesia’s New Democracy David T.Hill and Krishna Sen Asia.com Asia encounters the internet Edited by K.C.Ho, Randolph Kluver, and Kenneth C.C.Yang Japanese Cybercultures Edited by Mark McLelland and Nanette Gottlieb Literature and Society is a series that seeks to demonstrate the ways in which Asian Literature is influenced by the politics, society, and culture in which it is produced. Titles include:

Chinese Women Writers and the Feminist Imagination (1905–1945) Haiping Yan The Body in Postwar Japanese Fiction Edited by Douglas N.Slaymaker Routledge Studies in Asia’s Transformations is a forum for innovative new research intended for a high-level specialist readership, and the titles will be available in hardback only. Titles include: 1. Fertility, Family Planning, and Population Policy in China Edited by Dudley L Poston, Jr, Che-Fu Lee, Chiung-Fang Chang, Sherry L.McKibben, and Carol S.Walther 2. Genders, Transgenders and Sexualities in Japan Edited by Mark McLelland and Romit Dasgupta 3. Developmental Dilemmas Land reform and institutional change in China Edited by Peter Ho 4. Japanese Industrial Governance Protectionism and the licensing state Yul Sohn 5. Remaking Citizenship in Hong Kong Community, nation and the global city Edited by Agnes S.Ku and Ngai Pun 6. Chinese Media, Global Contexts Edited by Chin-Chuan Lee 7. Imperialism in South East Asia ‘A fleeting, passing phase’ Nicholas Tarling 8. Internationalizing the Pacific The United States, Japan and the Institute of Pacific Relations in War and Peace, 1919–1945 Tomoko Akami 9. Koreans in Japan Critical voices from the margin Edited by Sonia Ryang

10. The American Occupation of Japan and Okinawa* Literature and memory Michael Molasky * Now available in paperback. Critical Asian Scholarship is a series intended to showcase the most important individual contributions to scholarship in Asian Studies. Each of the volumes presents a leading Asian scholar addressing themes that are central to his or her most significant and lasting contribution to Asian studies. The series is committed to the rich variety of research and writing on Asia, and is not restricted to any particular discipline, theoretical approach, or geographical expertise. China’s Past, China’s Future Energy, food, environment Vaclay Smil China Unbound Evolving perspectives on the Chinese past Paul A.Cohen Women and the Family in Chinese History Patricia Buckley Ebrey Southeast Asia A testament George McT.Kahin

Fertility, Family Planning, and Population Policy in China Edited by

Dudley L.Poston, Jr, Che-Fu Lee, Chiung-Fang Chang, Sherry L.McKibben, and Carol S.Walther

LONDON AND NEW YORK

First published 2006 by Routledge 2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN Simultaneously published in the USA and Canada by Routledge 270 Madison Ave, New York, NY 10016 Routledge is an imprint of the Taylor & Francis Group This edition published in the Taylor & Francis e-Library, 2006. “ To purchase your own copy of this or any of Taylor & Francis or Routledge’s collection of thousands of eBooks please go to http://www.ebookstore.tandf.co.uk/.” © 2006 Dudley L.Poston, Jr, Che-Fu Lee, Chiung-Fang Chang, Sherry L.McKibben, and Carol S.Walther, selection and editorial matter; the contributors, their own chapters All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Cataloging in Publication Data A catalog record for this book has been requested ISBN 0-203-35644-6 Master e-book ISBN

ISBN 0-203-38955-7 (Adobe e-Reader Format) ISBN 0-415-32330-4 (Print Edition)

To our co-editor, Che-Fu Lee, who played a major role organizing, editing, and writing this book, but who did not live to see it published.

Requiescat in Pace!

Contents List of figures

xiii

List of tables

xv

List of contributors Preface

Prologue DUDLEY L.POSTON, JR AND CAROL S.WALTHER 1 Fertility and population policy: an overview QIUSHENG LIANG AND CHE-FU LEE PART I Family planning policy and contraceptive use 2 Patterns of induced abortion JUAN WU AND CAROL S.WALTHER 3 Patterns of sterilization CAN LIU AND CHIUNG-FANG CHANG PART II Family and marriage patterns 4 The impact of family structure on fertility FEINIAN CHEN 5 The impact of intermarriage on the fertility of minority women CHIUNG-FANG CHANG 6 Emerging patterns of premarital conception CAROL S.WALTHER 7 Changing patterns of desired fertility LI ZHANG, XIAOTIAN FENG, AND QINGSONG ZHANG PART III Biological and social determinants of fertility 8 Age at menarche and the timing of the first birth SHERRY L.MCKIBBEN

xviii xix

1 7 20

21 35 48

49 61 75 86 106

107

9 The effect of floating migration on fertility XIUHONG YOU AND DUDLEY L.POSTON, JR 10 The impact of language dialect on fertility XIAODONG WANG AND XIUHONG YOU PART IV Implications and the future 11 The managed fertility transition in rural China and implications for the future of China’s population CHE-FU LEE AND QIUSHENG LIANG 12 China’s demographic destiny: marriage market implications for the twenty-first century DUDLEY L.POSTON, JR AND KAREN S.GLOVER Index

122 140 152

153

167

182

Figures 1.1

Total fertility rate (average number of children per woman) for urban and rural places and all China, 1949–1999

12

1.2

Crude rates of birth, death, and natural increase in China, 1949– 14 2000

2.1

Probability of surviving the hazard of an abortion after the birth 27 of the first child, stratified by rural and urban residence, China, 1997

3.1

Kaplan-Meier estimates of Chinese women surviving the hazard 41 of sterilization, 1997

3.2

Kaplan-Meier estimates of Chinese women surviving the hazard 42 of sterilization, by residence, 1997

4.1

Cumulative distribution of first birth by duration months since date of marriage, CHNS

53

4.2

Residential patterns of parents-in-law and parents of married women, CHNS, 1993

54

5.1

Comparison of the observed distribution of CEB, with the univariate Poisson distribution with mean of 2.82

65

8.1

Kaplan-Meier survival estimates for the hazard of a first birth: Chinese Han women, 1997

113

8.2

Kaplan-Meier survival estimates for the hazard of a first birth: Chinese minority women, 1997

114

9.1

Frequency distribution of CEB variable, and univariate Poisson 127 distribution based on mean of 2.148

9.2

Generation of the migration status variable

129

11.1 Age-specific fertility rates by urban and rural places: all China

158

11.2 Age-specific fertility rates by urban and rural places: Hebei province

159

11.3 Population growth rates of China and India, 1950–2000

162

12.1 TFRs: China, 1950–2001

168

12.2 SRB: Mainland China and the United States, 1980–2001

171

12.3 Sex ratios by age: the United States, July 1, 1999

173

12.4 Males and females born in Mainland China, 1980–2001

174

12.5 Number of excess males at marriageable age of 22: Mainland China, 2000–2021

174

Tables 1.1

Population, birth, and death rates of China, 1949–2000

10

2.1

Descriptive data: 11,762 women with at least one child, China, 1997

27

2.2

Cox proportional hazards models for the hazard of having abortion for the pregnancy following a woman’s first child, China, 1997

31

3.1

Descriptive data of 10,406 currently married women with at least one child, China, 1997

40

3.2

Hazard ratios of female sterilization in China, 1997

43

4.1

Descriptive statistics of selected sample characteristics (N=218) 56

4.2

Woman month risk of first birth, 1989–1993, CHNS (N=12432) 57

5.1

Descriptive data for ever-married women aged 15–49, by ten ethnic groups, China, 1990

5.2

Mean numbers of CEB for exogamous women and endogamous 68 women, by ten ethnic groups, China, 1990

5.3

Reduced Poisson regression models predicting number of CEB: 69 minority ever-married women, aged 15–49, China, 1990

5.4

Poisson regression models predicting number of CEB, by rural and urban, minority ever-married women, ages 15–49, China, 1990

71

6.1

Dependent and independent variables: means and standard deviations, China, 1997

80

67

6.2

Logits and odd ratios of Chinese women having a premarital conception, 1997

82

7.1

Patterns of the desired number of children in rural areas of China

90

7.2

Pattern of the desired number of children in urban areas of China

94

7.3

Patterns of the desired gender of children in rural areas of China 95

7.4

Patterns of the desired gender of the next child and the number of existing children in rural China

97

7.5

Patterns of the desired gender of children in urban areas of China

99

8.1

Descriptive statistics of Chinese Han women for the hazard of a 112 first birth, 1997

8.2

Descriptive statistics of Chinese minority women for the hazard 112 of a first birth, 1997

8.3

Cox Proportional Hazard analysis of the hazard of a first birth: ever-married Han females, China, 1997

116

8.4

Cox Proportional Hazard analysis of the hazard of a first birth: ever-married minority females, China, 1997

117

9.1

Social demographic variables by migration status: ever-married 126 women in China, aged 15–49, 1990

9.2

Percentage distributions of education status, for nonmigrants 130 and groups of migrants: ever-married Chinese women, aged 15– 49, 1990

9.3

Poisson regressions of number of CEB: ever-married Chinese women, aged 15–49, 1990

131

9.4

Poisson regressions of number of CEB: ever-married Chinese

135

women, aged 15–49, 1990 (rural nonmigrants and rural-tourban floating migrants) 9.5

Poisson regressions of number of CEB: ever-married Chinese women, aged 15–49, 1990 (urban nonmigrants vs. rural-tourban floating migrants and urban permanent migrants)

136

10.1 Chinese dialects and their major locations in China

143

10.2 Lu’s quantified indices for selected cities and the dialects groups

144

10.3 The coding of Chinese dialects by province

145

10.4 Descriptive statistics of independent and dependent variables

147

10.5 Zero-order correlations of dependent and independent variables 147 10.6 Standardized regression coefficients of effects of independent variables on CBR

147

11.1 Crude rates of birth, death and natural increase for urban and rural China, selected years, 1957–1999

155

11.2 Development indices of China and India around the year 1950

161

Contributors Chiung-Fang Chang is Assistant Professor of Sociology at Texas A&M University, Kingsville, Texas. Feinian Chen is Assistant Professor of Sociology at North Carolina State University, Raleigh, North Carolina. Xiaotian Feng is Professor of Sociology at Nanjing University, Nanjing, China. Karen S.Glover is a Doctoral Candidate in Sociology at Texas A&M University, College Station, Texas. Che-Fu Lee (deceased) was Professor of Sociology at the Catholic University of America, Washington, DC. Qiusheng Liang is Research Associate at the Catholic University of America, Washington, DC. Can Lin is a JD Candidate at the School of Law at the University of Texas at Austin. Sherry L.McKibben is Assistant Professor of Sociology at the University of Texas of the Permian Basin, Odessa, Texas. Dudley L.Poston, Jr is Professor of Sociology, and Holder of the George T. and Gladys H.Abell Endowed Professorship in Liberal Arts, at Texas A&M University, College Station, Texas. Carol S.Walther is a Doctoral Candidate in Sociology at Texas A&M University, College Station, Texas. Xiaodong Wang is a Doctoral Candidate in Sociology at Texas A&M University, College Station, Texas. Jnan Wu is a Doctoral Candidate in Finance at Texas A&M University, College Station, Texas. Xiuhong You is a Doctoral Candidate in Sociology at the University of Texas at Austin; and a Research Associate in the Office of the State Demographer. Austin, Texas. Li Zhang is a Doctoral Candidate in Sociology at Texas A&M University, College Station, Texas. Qingsong Zhang is a Human Resources Officer at the Huawei Company, Shenzhen, China.

Preface The idea of a book on fertility, family planning, and policy in China was introduced at the meetings of the North American Chinese Sociologists Association (NACSA) held in Anaheim, California in August of 2001. Earlier versions of several of the chapters in this book were presented at the NACSA conference. Che-Fu Lee of the Catholic University of America served as the NACSA conference organizer. Since the authors of several of the NACSA conference papers were at the time graduate students of Dudley Poston at Texas A&M University, Lee and Poston discussed developing a China fertility book using the NACSA papers as a base. They approached several other scholars who did not make presentations at the NACSA conference to write chapters for the book, to help fill voids and to flesh out theoretical and empirical aspects of Chinese fertility and family planning not covered in the conference presentations. Poston then asked Chiung-Fang Chang, Sherry McKibben, and Carol Walther to assist him and Lee in putting the book together. In the subsequent development of this book, we are very much indebted to Mark Selden of the State University of New York, Binghamton, who serves as a China book series editor for Routledge, for assisting us in the preparation of a book proposal; to Zoë Botterill of Routledge Publishers for working with us in developing the book contract; and to Helen Baker of Routledge Publishers for helping us get the book ready for production. As the book chapters were revised and/or written, the five editors each, in turn, read and edited them. The edited chapters were then returned to the authors for revision. Poston was responsible for the final reading and editing. All the chapters were then copy-edited at Texas A&M University by Chris Lewinski, who then discussed and reviewed the final changes and edits with Poston. We thank her for her dedicated and timely work, and all the others who assisted us. When the book was in the final stages of editing in February of 2005, we were saddened to lose our co-editor Che-Fu Lee, who died at the age of 64. Without Che-Fu’s encouragement and hard work, this book would never have been completed. As already noted, Che-Fu was the organizer of the 2001 annual meeting of the North American Chinese Sociologists Association where several of the chapters in this book were first presented. He was very instrumental in commissioning several of the other chapters, he provided valuable editorial assistance to the authors of all the chapters, and he wrote two of the chapters with his colleague Qiusheng Liang. We dedicate our book to him. Dudley L.Poston, Jr College Station, Texas Chiung-Fang Chang Kingsville, Texas Sherry L.McKibben Odessa, Texas Carol S.Walther College Station, Texas

Prologue Dudley L.Poston, Jr and Carol S.Walther This book focuses on fertility, family planning, and population policy in China—the most populated country in the world. In 2004 China had nearly 1.3 billion inhabitants. After India, the United States is the third most populated country, with a population in 2004 of 293 million. But China has a land mass slightly less than that of the United States (9.6 million square kilometers of surface area compared to 9.8 million of the United States) with a population that is 4.4 times larger. China has reduced its fertility in the past two decades to less than two children per woman, has a family planning program focusing heavily on sterilization and abortion, and a population policy based on one child, and no more than two children, per woman. The country’s fertility transition, family planning programs, and one-child policy have captured the attention of academicians, researchers, policy practitioners, government officials, and laypeople the world over. Indeed, some have seen China’s experiences as providing important lessons for the demographic transitions of many countries in the developing world. Others have denounced the policy’s system of quotas, and particularly forced abortions, as gross violations of human rights. What is certain is that the transition has made it possible for China to achieve in a relatively brief time the status of a demographically developed country. Its fertility, family planning, and policy dynamics that are analyzed in the chapters of this book need to be considered in this context. This Prologue first places these demographic issues in a historical perspective. China today is not demographically a country with an aged population. In 2000, only one-tenth of China’s population was over age 60, compared to 16 percent for the United States. By comparison, in 2000, 10 countries, all in Europe, had more than 20 percent of their populations over age 60 (United Nations 2003). But China is one of the oldest countries in existence. Statistics on the size of China’s population suggest that the country had a population of around 60 million people at the time of Christ (Durand 1960). Of course, Chinese civilization began much earlier than the time of Christ, with the Xia Dynasty, the first dynasty of China, lasting from about the twenty-first century BC to the sixteenth century BC. There are, however, no demographic records of the Chinese population in the centuries before Christ, other than an estimate of about 13 million at the start of the Xia Dynasty (Sun 1988:9), a figure whose accuracy is difficult to establish. During the Han Dynasty, China took a population count in the second century AD, and it showed a population size of just under 60 million people (Banister 1992). The population increases and decreases over the almost 20 centuries since the time of Christ have usually been associated with dynastic growth and decay. Typically, the beginning of a new dynasty was followed by a period of peace and order, cultural development, and population growth. As population density increased, it often exceeded the availability of food, and the Malthusian struggle for existence was intensified. Then there would come a period of pestilence and famine resulting in a reduction in the size of the population. Two thousand years of Chinese records and archives show that for all the centuries prior to the seventeenth century, China’s population size increased to around 50–60

Fertility, family planning, and population policy in china

2

million before declining. Indeed at the start of the Ming Dynasty (in 1368) the size of China’s population was only slightly larger than it was at the time of Christ. For all the dynasties up until China’s last dynasty, the Qing (1644–1911), China’s population swayed roughly with the rise and fall of a dynasty (most dynasties reigned for about 200– 300 years). The population grew in the initial years of the dynasty, then fell, so that onethird or sometimes one-half of the original population was decimated. Mortality then was too high to allow much of an increase in population. To illustrate, from 1400 to 1500, the size of the Chinese population grew by around 25 million. It grew by another 50 million from 1500 to 1600. But since the mid-1700s after the establishment of the Qing Dynasty, slight reductions in mortality enabled the population to continue growing. By 1850 there were roughly 420 million people in the country, 6–8 times the traditional level (of 60–80 million) that was the demographic norm 200 years or so previously. The Qing Dynasty was supremely successful at living up to the Chinese ideal of “numerous descendants.” It is indeed ironic that by achieving this ideal, not only was the Qing Dynasty wiped out, but China’s dynastic system of almost four thousand years was eradicated. Previously, declines in population resulted in the collapse of the dynasties. The Qing fell in 1911, among other reasons, because the population became too large. By the date of the birth of the People’s Republic of China (PRC) on October 1, 1949, the population exceeded 500 million. It is at this point in the country’s demographic narrative that we may turn to this book’s first substantive chapter “Fertility and population Policy” by Qiusheng Liang and Che-Fu Lee. The authors revisit the process of the formation of the Chinese government’s population policies over the past half-century. Their primary aim, through a systematic review of the evolving policies over a period of five decades, is to provide a comprehensive picture of the dynamics between policy decisions and their implementation at different points in time since the 1950s. The population policy considerations are reviewed in the context of the political economy and are demarcated into three stages over the past 50 years: (1) the harbinger of population planning in the 1950s, (2) the chaotic decade of the 1960s and the establishment of the birth-control institution in the 1970s, and (3) a policy experiment followed by a decentralization since the early 1980s. Liang and Lee’s policy reviews serve to frame the discussions of China’s current fertility by presenting and discussing year-to-year baseline fertility data since the early 1950s. They show how the “on-again and off-again” fertility control policies of the 1950s and 1960s interacted with non-fertility related policies and ideologies to keep fertility rates high, especially in the rural areas. Their discussions provide an important perspective for the contemporary analyses of fertility that follow in later chapters of this book. Two key features of China’s family planning policy are abortion and sterilization. These are the subjects of Chapters 2 and 3 and comprise Part I of the book. The chapter by Juan Wu and Carol Walther deals with induced abortion. In 1957, induced abortion was introduced by the Chinese government as part of the first birth control campaign. This method was to be used during the first ten weeks of pregnancy. In later decades, however, for some women, induced abortion became the primary form of birth control, although the government discourages this practice. In this chapter, the authors show that population policy factors play a significant role in affecting a woman’s chances of

Prologue

3

aborting the pregnancy that occurs following the birth of her first child. The authors also show that if the first child is a daughter, the woman has a lower chance of aborting her next pregnancy. Chapter 3 is by Can Liu and Chiung-Fang Chang and deals with patterns of sterilization. Sterilization is the world’s most widespread form of birth control and has contributed significantly to China’s fertility decline. This chapter analyzes women’s knowledge of contraception as a significant predictor of sterilization. Women with greater contraceptive knowledge are shown to be less likely to experience sterilization than women with less contraceptive knowledge, among other reasons. Liu and Chang suggest that improvements in the education of couples about contraception and improved communication between clients and providers can achieve “informed choice.” Part II deals with family and marriage patterns. Feinian Chen’s Chapter 4 on “The impact of family structure on fertility” introduces the concept of a “modified extended family” in her analysis of fertility. Instead of focusing only on co-residence, she extends the boundary of the household by also incorporating parents who live close by, a residential type that is sometimes referred to as “quasi-co-residence.” One of the key mechanisms of how family structure affects fertility is via the reduced opportunity costs of children. In contemporary China, while grandparents do not necessarily bear the direct economic costs of raising children, they act as important alternative childcare givers. Indeed grandmothers are often identified as the most important caregivers other than parents themselves. But caregiving is not just limited to co-residing grandparents. Grandparents who live nearby are often just as likely to help. Thus, with parents in the household or living nearby as future childcare helpers, couples may choose to start having children earlier than if no such choice was available. Moreover, young couples could be under stronger normative pressures from their parents to start the childbearing process earlier for the purpose of carrying on the family line. Chen’s results show that the extended family living arrangements have a clear impact on the transition to the first birth. Co-residence with the husband’s parents substantially increases the likelihood of an early first birth. Moreover, when Chen expands her definition of extended family to include co-residence and quasi-co-residence with parents, a very similar effect is found. Her findings clearly demonstrate the importance and relevance of the family context, and the fact that this is not bounded by the household. Chiung-Fang Chang (Chapter 5) deals with the effects of intermarriage on the fertility of minority women. She measures intermarriage in two ways by coding the ethnicities of the woman and her spouse; if the minority woman married a Han man; or if the minority woman married a man from another minority ethnic group. In her analyses the reference group is the endogamous couple, that is, minority women who marry in the same ethnic group. She shows that both intermarriage variables have negative effects on fertility irrespective of whether the intermarriage was between a minority and a Han or between a minority and another minority. And these effects are independent of an assortment of other socioeconomic, demographic, and policy variables. Exogamous women, that is, women who marry outside their ethnic group, have lower levels of fertility than endogamous women. In particular, for minority women who intermarried with Han, they appear to be more willing to accept the norms and values of the mainstream Han culture.

Fertility, family planning, and population policy in china

4

Chang (Chapter 5) writes that the issue of ethnic assimilation is particularly important when one considers the fertility rates of the minority populations. Since intermarriage is an important consequence of ethnic assimilation, she notes the need to move beyond the research reported in her chapter. For instance, one area requiring attention is the ethnic identity issue of the second generation of intermarried couples in relation to the dynamics of overall population growth and the future population policy. Attention also needs to be directed to the phenomenon of multiple ethnicities and ethnic inequalities, an area just now being addressed in the West, but not yet in China. Chapter 6 by Carol Walther analyzes emerging patterns in China of premarital conception. More than 20 years ago Rindfuss and Morgan (1983) claimed that a silent but profound sexual revolution was occurring throughout Asia in which couples were moving away from arranged marriages to romantic marriages, thus leading to increases in premarital conceptions. In this chapter, Walther finds that among other factors, having friends with more nontraditional attitudes toward premarital sex, and having had sex education, are positively correlated with premarital conceptions. This is a new area of inquiry in China, and further study is needed regarding the impact of sex education on romantic relationships, premarital conceptions, and contraceptive usage. Chapter 7 is by Li Zhang, Xiaotian Feng, and Qingsong Zhang and focuses on desired fertility in number and gender preference. Using a meta-analysis, the authors show that there is a declining pattern in the desired number of children and son preference from the 1970s to the 1990s in both the rural and urban areas of China. A family that consisted of two children was preferred by rural and urban citizens in the 1980s, but in the 1990s, more urban people preferred one child. Generally speaking, urban citizens have lower desired number of children compared to their rural counterparts. Also, Beijing and Shanghai residents desired a small number of children, and Fujian and Guangdong residents desired a higher number of children. Part III moves to a consideration of biological and social determinants of fertility. The biological effect that is analyzed is age at menarche, and the social effects studied are migration and language dialect. Sherry McKibben’s Chapter 8 investigates whether a biological factor such as age at menarche has an independent effect on the likelihood over time of having a first birth. Her central goal is to model the likelihood of giving birth to a first child during the period of time starting with the onset of a woman’s menarche. The analysis is conducted for Han women and for high fertility minority women. She finds that the effect of age at menarche is consistently significant and positive for both minority and Han Chinese women. As a woman’s age at menarche increases, her likelihood of having a first birth also increases. McKibben notes that among the Han women, as they get older when reaching menarche, their hazard of a first birth increases by 3.6 percent and for the minority women, by 2.4 percent. Biologically, women who reach menarche at a later age have a shorter period of subfecundity and are thus more likely to experience a first birth sooner after reaching menarche than those women with an early age at menarche. The findings reported in McKibben’s chapter support this hypothesis. A woman who reaches menarche early is likely to waste many of her most viable follicles before she is ready to conceive, and her ovulation cycles will be spaced further apart. Therefore, if menarche is postponed, the woman’s chances of conception increase and she will experience her first birth soon after reaching menarche.

Prologue

5

Xiuhong You and Dudley Poston (Chapter 9) investigate the effect of floating migration on fertility. Many in China believe that floating (or temporary) migrants have higher fertility than non-migrants. They hold that one of the main reasons why some floaters migrate is to escape family planning controls in order to have more children. You and Poston point out that such an expectation is contrary to the prediction of most migration theories that expect migrants to have lower, not higher, fertility than nonmigrants. To date there has been no comprehensive empirical examination of the fertility of floating migrants and the degree to which their fertility surpasses that of non-migrants and permanent migrants. You and Poston show that after controlling for relevant factors, rural to urban floating migrants have lower, not higher, fertility than those left behind. Floating migrants do not have fertility rates higher than most of the population and they should not be referred to as “Child-bearing Guerillas,” as they often are in the popular press. Chapter 10 in Part III by Xiaodong Wang and Xiuhong You inquires whether language dialect has a significant effect on fertility. Expanding on the diffusion perspective of Susan Watkins (1991), the authors show that speaking different dialects does indeed appear to affect fertility. Those provinces speaking Mandarin and those provinces speaking dialects similar to Mandarin have fewer births than those not speaking Mandarin. Another interesting finding is that among the provinces, there is a significant negative relationship between the percentage of households with television sets and fertility. The researchers suggest that as China becomes more dominated by Mandarin, and as it experiences increased access to mass media, especially television, the crude birth rates overall, and differences in birth rates among provinces should further decrease. Part IV deals with implications and the future. It consists of two chapters. Chapter 11 is by Che-Fu Lee and Qiusheng Liang and evaluates the effects of birth control in China. China’s population of 1.27 billion, as reported by the most recent 2000 census, serves as a baseline reference for their analysis. Various projections extrapolating China’s population and its growth rate in earlier eras, with explicit assumptions of the policy effects (or lack thereof), are compared with China’s population in 2000. Lee and Liang demonstrate that China’s population today is at least 500 million less than it would otherwise have been as a result of changes during the last three decades in government efforts to implement a compulsory population control policy. Of the more than 500 million births averted, the authors show that the administrative and close monitoring of citizens’ reproductive behavior since 1970 in both the urban and rural regions accounted for more than half of the reduced population growth. The spurt of the radical “one-child” policy in the early 1980s was soon adjusted in 1984 in the rural areas to allow for a second birth if the first was a girl. The immediate effect was seen in a rise in period fertility rates in the subsequent years of the 1980s. Allowing for the possibilities of the underreporting of unplanned births, especially in rural places, fertility declined further in rural China in the 1990s to nearly converge with urban China. Chapter 12 by Dudley Poston and Karen Glover inquires into the future marriage market implications of China’s unbalanced sex ratio at birth, another outcome related to the one-child and modified one-child policy. The authors first show that in China in every year, beginning in 1980 through to the year 2001, many more boys have been born than girls. Indeed in the year 2000, almost 120 boys were born for every 100 girls. The authors

Fertility, family planning, and population policy in china

6

estimate that when the number of boys, who have already been born in China, reach marriage age, they will all not be able to find Chinese women to marry. They estimate that there are more than 23 million boys already born, who will not be able to find brides when they reach marriage age. Poston and Glover (Chapter 12) consider a number of possibilities about what these numerous extra boys will do when they are unable to find brides. The migration of foreign mail-order brides, a solution available to some rich countries such as Japan, is deemed unlikely. So too is a major increase in levels of homosexuality. Polyandry too is not thought by the authors to be a likely solution. The likely possibility is that these Chinese bachelors will never marry and will have no choice but to develop their own lives and livelihoods in the absence of nuclear families. Many are likely to live with one another in “bachelor ghettos” either in the countryside or in cities where commercial sex outlets are available. There is thus the real possibility of major increases in China in the incidence of HIV and AIDS owing to their transmission mainly through commercial sex outlets. Research also suggests that these bachelors will be more prone to crime than if they were married. This possibility has alerted some to the potential increases in crime in China’s future, and perhaps political ramifications, resulting from these excess males. The authors conclude that no one knows for sure what this excess number of young Chinese males will do. The only known fact is that there have already been born in China over 23 million more Chinese boys than there will be Chinese girls for them to marry.

References Banister, J. 1992. “A Brief History of China’s Population” pp. 51–57 in Dudley L.Poston, Jr and David Yaukey (eds), The Population of Modern China. New York: Plenum. Durand, J.D. 1960. “The Population Statistics of China, A.D. 2–1953.” Population Studies, 13:209–249, 255–256. Rindfuss, R.R. and S.P.Morgan. 1983. “Marriage, Sex, and the First Birth Interval: The Quiet Revolution in Asia,” Population and Development Review, 9(2):259–278. Sun, J. 1988. “The Chinese Population: Its Size and Growth.” pp. 9–14 in China Financial and Economic Publishing House, New China’s Population. New York: Macmillan. United Nations. 2003. World Population Prospects: The 2002 Revision. Two volumes. New York: United Nations. Watkins, S.C. 1991. From Provinces into Nations: Demographic Integration in Western Europe, 1870–1960. Princeton, NJ: Princeton University Press.

1 Fertility and population policy An overview Qiusheng Liang and Che-Fu Lee China, the most populous nation in the world, reported a population of 1.27 billion in the year 2000. Compared with the population of the United States, 280 million, China had 1 billion more people at the turn of the new millennium. Given 6 billion people over the whole world, Chinese represented a little more than one-fifth of the human race on earth. In late 1979, when the Chinese Government launched an ambitious one-child-percouple population policy, the target was to limit the population of the country within 1.2 billion by the end of the twentieth century. To implement the “one-child” policy the government devised various “carrot and stick” measures to reward couples who pledged to have no more than one child and to penalize those who exceeded the one-child limitation. Anecdotal reports of forced abortions by local officials who were overzealous in meeting reproductive quotas produced international uproars over human rights violations and calls for sanctioning foreign aid to China’s family planning programs. Many were left with the impression that China had only one population policy, that is, a policy permitting only one child per couple. This fragmented picture of China’s attempts to control its population problem tended to obscure the historical processes of social construction of population growth and efforts to control it in the context of broader development strategies. This chapter revisits the process of the formation of the Chinese government’s population policies over the past half-century. The primary aim, through a systematic review of the evolving policies over a period of five decades, is to provide a comprehensive picture of the dynamics between policy decisions and their implementation at different points in time since the 1950s. Developments of population policy in the context of political economy, which are discussed here in some detail, can be demarcated into three stages over the past 50 years: (1) the harbinger of population planning in the 1950s, (2) the chaotic decade of the 1960s and the establishment of the birth-control institution in the 1970s, and (3) a policy experiment followed by a decentralization since the early 1980s. These policy reviews will serve to frame the discussions of China’s current fertility by showing and discussing year-to-year baseline fertility data for the country since the early 1950s. It will be shown how the “on-again and off-again” fertility control policies of the 1950s and 1960s interacted with non-fertility-related policies and ideologies to keep fertility rates high, especially in the rural areas. The discussions in this chapter will provide an important perspective for the more contemporary analyses of fertility that follow in later chapters of this book (also see Poston and Glover’s Chapter 12 in this book for more discussion).

Fertility, family planning, and population policy in china

8

The 1950s—harbinger of population planning The Korean war (1950–1953) delayed China’s post-civil-war reconstruction for a few years after the establishment of the new People’s Republic in 1949. On July 1, 1953, China conducted for the first time a national census of its population. The census was required for the preparation of the official First Five-Year Plan and for the apportioning of political representation, the same political purpose for which the United States took its first census in 1790 (Baumle and Poston 2004). The first Plenary Congress of the People was held in 1954. The results of the first population census revealed that China had a total population of 602 million in 1953, a figure that was one-third larger than 450 million, the number commonly believed to be the size of the population in China at that time. Many concerned scholars and government officials, especially those involved with the drafting of the First Five-year Plan, were alarmed by the unexpectedly large number of people that they had to deal with. The first communication at the upper echelon of the ruling party pertaining to the issue of population may be traced to Deng Xiaoping, the late supreme leader and architect of China’s economic reform from the late 1970s to the 1990s. He was then the Secretary General of the Chinese Communist Party (CCP). In his letter (1954) responding to Madam Zhou Enlai (Mr Zhou was the Premier), he stated, “I think it is necessary and beneficial to promote effectively the use of contraception” (Ca et al. 1999:112–113). At the first meeting of the People’s Congress in 1954, a delegate and a noted scholar in China, Shao Lize proposed, “in our country, we may set aside the issue of abortion, but health knowledge about contraception must be propagated. In addition, practical guides and necessary means and material supplies of contraception must be provided” (People’s Daily 1954). Apparently the top leaders were concerned about population pressure, and this was echoed in the public forum. Shao subsequently published a pamphlet entitled, “Basic Knowledge Regarding the Propagation of Contraception,” and continued to write a series of related articles in the public media. Shao would become the first person in the history of the new Republic to openly advocate the practice of family planning or birth control (Ca et al. 1999:113). Under the leadership of President Liu Shaoqi, a meeting was called at the ministry level in December 1954, to gather inputs on family planning from various government offices. At that meeting, Liu pronounced the support of the Party for planned fertility. In 1955, the Party’s Central Committee approved a report submitted by the Ministry of Health, in which it was said that regulation of birth is of concern with the living of people’s lives. It is an important policy matter. Given the conditions at this point of our history, for the benefits of the nation, family, and future generations, the Party supports appropriate regulation of births. The Party Secretariats at all levels of government (except in the autonomous regions of the ethnic minorities) must work with the masses to spread this policy of the Party. (She 1988:120–121) This became the first official document known as “the Directive on the Problem of Population Control.” This policy decision was then followed by the organization of a

Fertility and population policy: an overview

9

leading team for Planned Births, headed by Chen Muhua, who later became the first female vice-premier. Thereupon, the first official policy on population was announced, and the first government office in charge of implementing population policy was born. In subsequent years, there were other reiterations of the need for population control in relation to economic development. For example, an official document on agricultural development distributed in 1957 included population control as an intrinsic component of the policy. On the other hand, beginning with the 1955 “Directive” followed by other government initiatives in population control, a heated debate was unfolding in academic circles on the pros and cons of the perception of the so-called “population problem.” In the year 1957 alone, it is estimated that more than eighty scholars of all disciplines engaged themselves in writing hundreds of articles on this debate. The most memorable was by an economist and a prominent intellectual, Ma Yinchu, the then president of Beijing University. He treaded a thin line between Marxian ideology and Malthusian theory by publishing a booklet called, “The New Population Theory” (Ma 1997). His major argument was to negate the Marxist doctrine that a socialist society could not have a population problem. He maintained that in a socialist planned economy, population reproduction must be an integral part of the planning. The opposite camp, characterizing themselves as orthodox Marxists, came forth immediately to accuse Ma of adopting bourgeoisie Malthusian population theory. They claimed that the Malthusian view of the population problem blamed the victims, or the oppressed class, for the problem generated by a capitalist system, and hence was antithetical to Marxist doctrine. The misfortune of Ma’s new population theory was, ironically, aggravated by the Great Leap Forward that Chairman Mao launched in 1958 in the midst of the Party’s internal power struggle, an Anti-Rightist political movement. Mao led this political movement to eradicate the socalled capitalist revisionists within the Party. Under fire from this political movement, Professor Ma was labeled as a political rightist and in 1960 was forced to resign the presidency of the most prestigious university in China. The discussions of the population problem and ways to deal with it were muted not only by political pressures but by the economic calamity during the years 1959–1962.

The 1960s—the haunting concern of population in a chaotic decade If the idea of a planned population and regulated reproduction dawned on the horizon of China in the 1950s, its debate and controversy were limited to the elite circle of the Party leaders and concerned academicians. For the general masses, controlling the number of births contradicted the age-old cultural tradition which deemed a large family as a blessing of good fortune. To the Marxist ideologues, this was antithetical to the building of a Utopian socialist society. Aside from cultural and ideological resistance, the Chinese economy and medical technology in the 1950s were not equipped with a public health system capable of implementing a policy of birth control. Even if the camp advocating control of population growth were to prevail, there was no realistic means to provide the public with the necessary knowledge and effective means of contraception on any scale. As if adding salt to the wound of the idea of population policy, there was an economic calamity caused by misguided policy and aggravated by radical natural disasters in 1959– 1961. During that period, China experienced a sharp decline in the birth rate together

Fertility, family planning, and population policy in china

10

with an unprecedented spike in the death rate. An estimated 30 million or more deaths from starvation or diseases related to under-nourishment occurred during the turn of the decade of the 1960s. In this period, China actually witnessed a negative rate of population growth. Obviously, this historical accident or incident did not help in resolving the paradox of population control in China. Immediately following the years of the economic fiasco, a “baby boom” occurred, beginning in 1962. During the years 1962–1966, the crude birth rate hovered at around 40 per thousand, representing total fertility rates between 5 and 7 (see Chapter 12 by Poston and Glover in this book for more discussion). The return to traditional high-level uncontrolled fertility was especially true in the rural areas. The total fertility rate was as high as 6 and 7 during most of the 1960s (Table 1.1 and Figure 1.1). The rise in fertility was due mainly to compensatory births delayed during the prior bad years and a recovery of the normal rate of marriage. The demographic impact of this baby boom in the early 1960s would have an echo in the 1980s, as will be discussed in a later section. Alarmed by the bumper baby crop, the central government returned to the consideration of population control. In 1962, the State Council, the government central executive office, announced a new “Directive.” This time, it was to “seriously” confront the problem of planned reproduction. It commanded the government at all levels not only to promote birth control but also to review the results of its birth-control campaign. To clear up any ideological ambivalence, the new policy document pronounced that planned reproduction was necessary for the construction of a socialist society and was in no way to be confused with reactionary Malthusian thoughts. The seriousness of this official installation of a population policy was indicated by the institution of a Committee on Planned Births at the central government, and similar administrations in charge of family planning at all subordinate levels down to the local authorities. Government health departments and hospitals began to conduct medical research in contraceptive devices, induced abortion, and sterilization. In addition to the dissemination of contraceptive knowledge, the public health agencies provided contraceptives free of charge. Health workers were recruited for training in order to meet the expanded need for birth control. Finally, the central government set a goal to reduce the population growth to

Table 1.1 Population, birth, and death rates of China, 1949–2000 Year

Population (‘000)

Crude birth Crude death rate (0/00) rate (0/00)

Natural increase (0/00)

Total fertility rate National Rural Urban

1949

541,670

36.00

20.00

16.00

6.139

—

—

1950

551,960

37.00

18.00

19.00

5.813 5.963

5.001

1951

563,000

37.80

17.80

20.00

5.699 5.904

4.719

1952

574,820

37.00

17.00

20.00

6.472 6.670

5.521

1953

587,960

37.00

14.00

23.00

6.049 6.183

5.402

1954

602,660

37.97

13.18

24.78

6.278 6.390

5.723

Fertility and population policy: an overview

11

1955

614,650

32.60

12.28

20.32

6.261 6.391

5.665

1956

628,280

31.90

11.40

20.50

5.854 5.974

5.333

1957

646,530

34.07

10.80

23.23

6.405 6.504

5.943

1958

659,940

29.22

11.98

17.24

6.679 5.775

5.253

1959

672,020

24.78

14.59

10.19

4.303 4.323

4.172

1960

662,070

20.86

25.43

−4.57

4.015 3.996

4.057

1961

658,590

18.20

14.24

3.78

3.287 3.349

2.982

1962

672,950

37.01

10.02

26.99

6.023 6.303

4.789

1963

691,720

43.37

10.04

33.33

7.502 7.784

6.207

1964

704,990

39.14

11.50

27.64

6.176 6.567

4.395

1965

725,380

37.88

9.50

28.38

6.076 6.597

3.749

1966

745,420

35.05

8.83

26.22

6.259 6.958

3.104

1967

763,680

33.96

8.43

25.53

5.313 5.847

2.905

1968

785,340

35.59

8.21

27.38

6.448 7.025

3.872

1969

806,710

34.11

8.03

26.08

5.732 6.263

3.299

1970

829,920

33.43

7.60

25.83

5.812 6.379

3.267

1971

852,290

30.65

7.32

23.33

5.442 6.001

2.882

1972

871,770

29.77

7.61

22.16

4.984 5.503

2.637

1973

892,110

27.93

7.04

20.89

4.539 5.008

2.387

1974

908,590

24.82

7.34

17.48

4.170 4.642

1.982

1975

924,200

23.01

7.32

15.69

3.571 3.951

1.782

1976

937,170

19.91

7.25

12.66

3.235 3.582

1.608

1977

949,740

18.93

6.87

12.06

2.844 3.116

1.574

1978

962,590

18.25

6.25

12.00

2.716 2.968

1.551

1979

975,420

17.82

6.21

11.61

2.745 3.045

1.370

1980

987,050

18.21

6.34

11.87

2.238 2.480

1.147

1981

1,000,720

20.91

6.36

14.55

2.631 2.910

1.390

1982

1,016,540

22.28

6.60

15.68

2.860 3.230

1.580

1983

1,030,080

20.19

6.90

13.29

2.420 2.780

1.340

1984

1,043,570

19.90

6.82

13.08

2.350 2.700

1.220

1985

1,058,510

21.04

6.78

14.26

2.200 2.480

1.210

1986

1,075,070

22.43

6.86

15.57

2.420 2.770

1.240

Fertility, family planning, and population policy in china

12

1987

1,093,000

23.33

6.72

16.61

2.590 2.940

1.360

1988

1,110,260

22.37

6.64

15.73

2.520 2.740

1.460

1989

1,127,040

21.58

6.64

15.04

2.350 2.540

1.550

1990

1,143,330

21.06

6.67

14.39

2.310 2.400

1.490

1991

1,158,230

19.68

6.70

12.98

2.200 2.250

1.420

1992

1,171,710

18.24

6.64

11.60

2.000 2.110

1.360

1993

1,178,440

18.09

6.64

11.45

1.810 1.960

1.300

1994

1,192,840

17.70

6.49

1121

1.810 1.890

1.290

1995

1,204,860

17.12

6.57

10.55

1.780 1.810

1.280

1996

1,217,550

16.98

6.56

10.42

1.700 1.700

1.210

1997

1,230,080

16.57

6.51

10.06

1.620 1.610

1.170

1998

1,242,210

16.52

6.55

9.97

1.490 1.520

1.130

1999

1,259,090

15.23

6.46

8.77

—

—

—

2000

1,268,000

—

—

6.94

—

—

—

Figure 1.1 Total fertility rate (average number of children per woman) for urban and rural places and all China, 1949–1999. Source: Table 1.1. 1 percent or less by the end of the twentieth century, and each of the provinces and municipalities was required to project short- and long-term targets for the birth-control campaign. Hebei province, for instance, in its 1963 “ten-year plan of fertility control,” aimed at a reduction of its crude birth rate to 13 per thousand by 1975; and the Shanghai municipality proposed to decrease its birth rate to 15 per thousand by the year 1967.

Fertility and population policy: an overview

13

It was unclear whether these targets of fertility control were supported by any scientific demographic analysis. It was undoubtedly, however, advancement from a stage merely advocating a planned reproduction by the government to another at which the government engages actively in the implementation of birth control. This policy imprint on the precipitous decline during the years 1962–1966 was visible in terms of the period’s total fertility rate (TFR), especially in urbanized areas, as shown in Figure 1.1 in which the rural and urban TFRs are displayed separately. Led by Shanghai, a campaign slogan, “one is not too few, two, just right, and three, too many” (Ca et al. 1999:157–158), was spread quickly throughout the nation. And the effective measure to achieve a small family size was put in three words, “wan, xi, shao,” or “late, sparse, and few.” “Late” referred to a delay in marriage and the first birth to a later age, “Sparse” was to space the birth interval by 4–5 years, and “few” to limit the number of births to 2 or no more than 3 per couple. To the extent that urban areas and the rural regions in the vicinity of major cities were regimented into work-unit organizations under a socialist system, the planned-birth policy was effectively carried out. The neighborhood committee in the city and production brigade in the countryside were formed as the grass-roots unit in which a women’s association was put in charge of regulating the childbearing behavior of all families in their jurisdiction. Technically, more than a dozen contraceptive paraphernalia were distributed free of charge by these local offices. There were clinics providing sterilization without charging a fee, and the patients were given leave from work with pay. The Great Cultural Revolution of 1966 in which Chairman Mao mobilized the masses to eradicate his political foes within the Party, brought the administrative system literally to a halt. The government’s role in the birth-control campaign dissolved quickly and was not revived until 1970. The demographic mark of this political turmoil in the late 1960s was a spike in the TFR in both the urban and rural areas (see Figure 1.1). The Cultural Revolution was not ended politically until the arrest in 1976 of the socalled “gang of four” led by Madam Mao, a group allegedly plotting to usurp the control of the Party central power. Premier Zhou Enlai, who oversaw the economic deterioration while handling a difficult act of balancing the political struggles, was determined to put China’s economic house in order. In 1970 he assembled the various ministries under his command to plan an economic revival. He addressed the issue of planned births and criticized the top officials, saying that “planned population reproduction is not limited to the jurisdiction of the health department, it is an integral part of overall planning. If you cannot plan for population growth, it is futile to talk about national planning” (Sun 1990:158). In 1971 the State Council issued a referendum on the nationwide campaign of fertility planning. The short-lived movement of birth control in the early 1960s was thereby resuscitated and even more vigorously enforced. This time the reinstitution of a population policy was also blessed by Chairman Mao’s endorsement. In approving a government document in 1974 in his later years, he uttered that “population must be controlled by all means” (Ca et al. 1999:166). At that time Mao was still a god-like supreme center of power in China. With his stature, he commanded the eager following of the planning policy not only by party officials but also widely among the masses, especially the city folks who were mostly employed by the state. Unlike the one-child policy in the 1980s, which had compulsory penalties for violators, the vigorous administrative means of watching closely over women of

Fertility, family planning, and population policy in china

14

reproductive ages and regulating their birth behavior in the 1970s was leaning more toward the prevention of excess births. Moreover, the allowance of two children per couple, spaced by 4–5 years, was easily acceptable by city residents, and it did not meet with strong resistance from the people in most of the densely populated rural areas. The results of the 1970s’ planned-birth campaign may be seen in the decline of the crude birth rate nationwide from 33 per thousand in 1970 to 18 per thousand in 1980 (see Figure 1.2). The TFR was reduced from 5.8 in 1970 to 2.2 in 1980 nationwide; and dropped to slightly less than 1.2 in the urban areas (Figure 1.1). This policy-induced fertility transition in a matter of one decade for a nation of 1 billion population was unprecedented in the demographic history of mankind. It is little wonder that the United Nations and other international organizations involved in population planning, in the era of its active promotion in the 1970s, picked China as a model for other developing countries.

Figure 1.2 Crude rates of birth, death, and natural increase in China, 1949– 2000. Source: Table 1.1.

1980 onward: a policy experiment followed by decentralization In 1979, China adopted, at its third plenary meeting of the fifth People’s Congress (or Parliament), an economic reform and open policy. Along with the restoration of the previously abolished social sciences, population research institutes were established in a number of major universities and in the Academy of Social Sciences. The new Premier, Zhao Ziyang, in consultation with experts’ projections of China’s population for the future 20–30 years, proposed a population limit of 1.2 billion by the end of the twentieth century. In order to achieve this goal of population control, the State Council issued an ambitious “one-child policy” in late 1979. This was to limit family size to one child for all newly wed couples except in the autonomous regions where the ethnic population was

Fertility and population policy: an overview

15

under 10 million. In fact, the goal of this new compulsory birth-control program to limit to one child per family, had already been nearly accomplished in urban areas, as reflected in the TFR of less than 1.2 in 1980. For the rural populace it was a different and much more complicated story. The rural economic reforms reverted from collective farming to family farming, for which family laborers, especially male ones, were minimally required assets. Moreover, for all the years of socialist construction, China could barely afford a prototype social security system, including the provision of a retirement pension in the cities in which most workers were state employees. The rural Chinese were left to fend on their own for old age support, depending on their male posterity. It was nearly impossible to convince the majority of Chinese in rural areas that having only one child, especially when the first birth was a girl, meant an advantage to their own families. The local enforcement of the one-child policy by the village Party officials was inevitably translated into forced abortion, the under-registration of female first births, and even bribery for the permission to have a second birth. The tension between the government’s demand on local officials and the wish of the rural people boiled. Reports on persecutions of women and families who failed to conform to the one-child policy spread outside the country, which led to an outcry regarding violation of human rights from the international community, especially in the United States during the Reagan administration of the 1980s (Hao 2003:17–19). Under these pressures inside and outside the country, the Chinese authorities reconsidered their radical one-child policy. In 1984, the central government approved a progress report submitted by the State Committee on Fertility Planning on the “Situations of Implementing Planned Births,” and announced a policy adjustment. In its official language, “the policy of planned births must be built upon the people’s support. It must be sensible and reasonable to the people, and doable to the responsible officials” (National Committee on Planned Births 1992:35– 43). To mitigate the tensions among the people and corruption of the officials thereof, a second birth may be allowed in the rural areas and a third one permitted under special conditions, but no more beyond that. This signaled the end of the “one-child policy” in rural China. In the province of Shandong, a policy allowing “only-girl families” to have a second child, after a minimum interval of four year, was initiated. By 1986, similar policies were soon adopted in other provinces for rural families (see Scharping 2003, for more discussion). By a rough estimate, this meant that one-half of the rural families would proceed to have two children instead of one. Notable in this policy adjustment in the late 1980s was the decentralization of the planning and implementation of population regulations from the central authorities to the provincial and county levels of the government. Regional differences and local variations in economic development and socio-cultural environment dictated realistic targets and appropriate measures of population control tailored by the local governments. In general, the one-child policy continued to be enforced among the state employees and residents in the cities and rural townships, although exceptions were granted case by case. In the wider countryside and regions of ethnic minorities, the population control policy continued to be enforced, but in a more flexible fashion. To correct over-zealous local officials and, at times, erratic compromises of laws, the new policy guideline emphasized “the three keys” to family planning work: the key of prevention (of unplanned pregnancies), the key of education, and the key of consistency (in policy implementation). Together these meant, in other words, that planned births must be

Fertility, family planning, and population policy in china

16

integrated with assistance to the peasant families in economic development, in raising the civility of rural culture, and in guaranteeing the reproductive health of the women. The policy adjustment in the mid-1980s contributed inevitably to a “baby-boomlet” as may be seen in the crude birth rates in the period of 1984–1990 (Figure 1.2). There were other factors besides birth-planning concerns that added to this period increase in the birth rate. As mentioned earlier, one was the “demographic echo,” that is, those who were born during the early 1960s’ baby boom were entering the marriage and reproductive ages by the mid-1980s. Moreover, in connection with the government’s reforms beginning in 1980, a new marriage law was legislated in the People’s Congress, which stipulated the minimum age of marriage at 20 for females and 22 for males. These legal ages for marriage were lower than the practice of 25 for females and 27 for males during the “late” policy of the 1970s, which was an executive order but never made a law. The new marriage law was oblivious of the population control policy. It nonetheless added many younger married couples in the 1980s to increase the momentum of child births in that period (Figure 1.2). The government, in view of the rising fertility rate in the later part of the 1980s, adjusted in the early 1990s its projected target number for the total population from 1.2 billion to nearly 1.3 billion. This latter figure proved to be more realistic and would soon be verified by China’s latest census conducted in November 2000. It is advisable, in order not to be overly alarmed by the upturn in fertility as a consequence of relaxing the onechild policy, to look at the TFR in addition to the crude birth rate, during the late 1980s (Figure 1.1). In spite of a discernible rise in the crude birth rate (Figure 1.2), the TFR was relatively stable over the same period. The TFR in urban China continued to decrease from 1.5 to about 1.2 per woman throughout the 1980s, and that of the rural population remained below 3.0. The breakdown separately for the urban and rural TFR was not yet available for the 1990s. The national average TFR was, however, brought down from 2.3 in 1990 to around 1.8 in the late 1990s. This is below the TFR level of 2.1, generally regarded as required to replace a population. The statistics cited so far suggest that the population control policy may have been stabilized or institutionalized, and the norm of a small family size had taken root in the 1990s in China.

Summary China was among the earliest countries in the world to realize population pressures in planning for economic development. The government’s intent to control population can be traced back to the 1955 document, “Directive on the Issue of Population Control.” Constrained by the postwar conditions of the economy and the lack of modern technology, the government then was not equipped to actually deliver any effective measures to reduce the traditionally high fertility rate. The innovative idea of birth control was even more hampered by the age-old Chinese value that a large family is equated with prosperity, on the one hand, and by the ideological contradiction to Marxist doctrine, on the other. Following the economic fiasco at the turn of the 1960s and the subsequent bumper crop of babies due to compensatory births for the delay during the bad years, the government resolved in 1962 to slow the rapid rate of population growth. This time, the official pronouncement of a population policy was to clear away the

Fertility and population policy: an overview

17

ideological ambivalence regarding whether population planning was contradictory to the Marxist economic planning. The government’s action then was to establish an office in charge of birth planning at the central government and related offices at the provincial and local levels. Research in contraceptive knowledge and technology was widely engaged in by hospitals and health agencies, and public health workers were recruited for special training in methods of averting births, including the use of contraceptive devices, induced abortion, and sterilization. This system of government machinery in birth control campaigns, however, was short lived, although it did make a clear mark on bringing down fertility from its peak in 1962. It was, unfortunately, disrupted by the political upheaval of the Great Cultural Revolution, beginning in 1966. Systematic birth control was, nonetheless, restored and put into more vigorous use in the early 1970s. Throughout the decade of the 1970s, the control policy of “late, sparse, and few” was effectively implemented to decrease precipitously the fertility rates in rural as well as in urban areas. Perhaps, encouraged by the 1970s’ success in carrying out the birth control program, the reform-minded leadership of the central government decided in 1979 to adopt a radical one-child policy to achieve a projected goal of limiting China’s population to a magic number of 1.2 billion in the year 2000. The rigid demand of one child per couple ran into strong resistance, especially in the rural areas. In 1984, the central government was forced to revise the one-child policy, and allow provincial and local authorities the option of permitting a second child if the first child was a girl. Although there appeared to be an upturn in the birth rate associated with this relaxation of the one-child policy in the 1980s, much of that short-term increase may also be attributed to the demographic echo-effect. The baby crop in the early 1960s had come of age in the 1980s to marry and have children. The relatively stable TFR, less than 2.0 per woman through the years of the 1990s, however, suggests that fertility regulation behavior and the norm of a small family have become institutionalized and taken root in the population of China. The imprints of the hesitation in putting a government’s population program in place in the 1950s and early 1960s and the vigorous implementation of birth control throughout the 1970s, on the changing fertility in China can be seen in the year-to-year statistics displayed in Table 1.1. For a quick overview, Figure 1.2 shows in graphic form the fertility transition over the past five decades, and, together with the declining death rate, shows also the declining rate of population growth in China. Figure 1.1 depicts the parallel transition of fertility in terms of total fertility rate between urban and rural China, though the latter remains at a higher level than the former. In a white paper on population control and development, released in December 19, 2000, China’s government set a target to limit its population to 1.33 billion by 2005, with an annual growth rate lower than 1 percent. For the long term, China projects its population to peak at 1.6 billion by the mid-twenty-first century and to gradually decline thereafter (People’s Daily, December 20, 2000). With a population pushing toward 1.2 billion in 1998, the latest available estimate of China’s GDP per capita was around US$ 800. This average level of living was higher than other highly populated developing countries like India, Indonesia, and Pakistan, which had a per capita GDP below US$ 500, but was in no way comparable to other Chinese in Hong Kong and Taiwan where the GDP per capita was between US$ 20,000

Fertility, family planning, and population policy in china

18

and 25,000. Although since the 1980s China has seen an annual rate of GDP growth averaging about 8 percent, the economic gains in the post-reform era were limited to the more developed urban regions on the east coast of China. Three-quarters of China’s population live in rural areas, and an estimated 50 million or more in remote pockets live in abject poverty. China is aiming, as stated in the most recent white paper on population and development, for a universal nine-year education for all its populace by 2005 and a popularized secondary and post-secondary education nationwide by 2010. Basic medical health care and social security are also expected to become generally accessible. In view of the fast-growing proportion and number of older people in China, because of the drastically reduced rate of fertility in the most recent three decades, a secure system providing basic living and health care for the retired elderly will be critical in the near future. Control of population, in view of all these goals of development, is of vital national interest for China. Just as national security is a vital national interest, China will continue to monitor and intervene as necessary in people’s reproductive behavior in the coming decades. The difference, if any, is seen in China’s turn to openness to multilateral international cooperation in defining and solving problems of national interest. This increased openness to international dialogue is even more foreseeable in the post-Cold War era, and especially since China has joined the World Trade Organization with imported intelligence from outside, together with the increasingly advanced social sciences, including demographic studies inside, China will likely help design future population policy in a more comprehensive and less intrusive way than the draconian one-child policy.

References Baumle, A.K. and D.L.Poston, Jr. 2004. “Apportioning the House of Representatives in 2000: The Effects of Alternative Policy Scenarios.” Social Science Quarterly, 85: 578–603. Ca, R., W.Hu, and J.Zei. 1999. Demography of A Century. Beijing: Beijing Arts and Literature Publishers. Hao, L. 2003. “Regarding the United States’ Refusal to Contribute to the U.N. Fund for Population Activities.” Population Research, 1(27):17–19. Ma, Y. 1997. The New Population Theory. Changchun, China: Jilin People’s Publishers. National Committee on Planned Birth. 1992. Collection of Planned Birth Documents, 1981–1991. Beijing: China’s Democratic-Law Publishers. People’s Daily. 1954, September 18. People’s Daily (overseas edition). 2000, December 20. “China’s Population and Development in 21st Century.” A White Paper by the Press Office of the State Council. Scharping, T. 2003. Birth Control in China, 1949–2000. London: RoutledgeCurzon. She, C. 1988. A History of Planned Birth Activities in China. Urumqi, China: Xinjiang People’s Publishers. Sun, M. 1990. A History of China’s Planned Birth Programs. Beijing: Beijing Publishers of Women and Children.

Part I Family planning policy and contraceptive use

2 Patterns of induced abortion Juan Wu and Carol S.Walther

Introduction China has experienced one of the most successful fertility declines in demographic history, an observation also made in other chapters of this book. Its total fertility rate (TFR) fell from about 6.5 in the 1960s to 1.8 in the year 2000 (United Nations Economic and Social Commission for Asia and the Pacific: http://www.unescap.org/). While socioeconomic development has undoubtedly played an important role in this process, the precipitous decline in fertility is more likely due to the implementation of the “one couple, one child” population policy initiated in 1979 (Poston and Gu 1987; Poston 1988). Induced abortion, among other proximate causes, is recognized as having greatly contributed to fertility decline in many developing countries (Bongaarts 1978; Frejka 1985; Knodel et al. 1987). There are three main objectives in this chapter. First, a general discussion about abortion will help the readers’ understanding of abortion issues in China. There has been much publicity about the controversial practice in China of using abortion to end unplanned pregnancies. And some international journalists have focused on the extreme examples rather than on the typical situation in China. Second, although there have been a few studies of abortion in China, the findings have not been conveyed effectively to the general public, especially to the international audience. Research reported in this chapter, based on a nationally representative sample, will present very detailed information about this issue, and will contribute to a better understanding of abortion. Third, the research findings reported here may indicate ways in which the Chinese Government can make family planning programs more effective and may also help understand various other social issues. For instance, an important social issue is China’s increasingly abnormal sex ratio at birth since the mid-1980s and its serious social consequences in years to come (see Poston and Glover’s Chapter 12 in this volume). Zeng and his associates (1993) argue that the increase in the sex ratios at birth is due in part to sex-selective abortion. Examining the effects of son preference on the hazard of having an abortion, as will be done in this chapter, will hopefully shed more light on the issue of rising sex ratios at birth. In fact, the analyses in this chapter will not only be important for China, but for other countries which may encounter similar problems at a later time.

Fertility, family planning, and population policy in china

22

History of abortion in China In China, abortion law may be traced back to the Tang dynasty (618–906) where it was a crime to induce an abortion by assaulting a pregnant woman (Rigdon 1996). In the following dynasties up to the Qing (1644–1911), the legal codes stipulated that the fetus had to have assumed a human form for the abortion-in-assault to apply. During the Qing dynasty, abortion was illegal, though there were some exceptions permitting a pregnant woman to obtain one (Rigdon 1996). In the first four years after the founding of the People’s Republic of China (PRC) in 1949, abortion was still illegal. Even though family planning as an official policy was initiated as early as 1953, abortion was available only when there was a risk to the mother’s health, or in cases when the birth interval since the last child was too short and the mother experienced difficulty in breastfeeding the previous child. Abortions were conducted in hospitals with the consent of the doctor, parents, and their work units (Li et al. 1990; Rigdon 1996). The regulations also stated that doctors were not authorized to perform abortions unless the couple seeking it already had four children (Rigdon 1996). In 1957, China legalized induced abortion in the first 10 weeks of pregnancy as part of the government’s first birth control campaign. However, due to a lack of medical facilities and personnel, and the unwillingness of many physicians to accept the new law, there was not a large increase in the number of abortions performed (Tu and Smith 1995; Rigdon 1996). Family planning was not a priority during the mobilization for the Great Leap Forward that began in 1958 when programs were initiated for the rapid communalization of agriculture and industry. In the following years of widespread famine in China, fertility control was not a priority. Only after the recovery from the consequences of the Great Leap Forward and the severe famine of the early 1960s did birth control become a priority. However it was not until the late 1970s when the onechild policy was initiated that abortion became a significant complement to the family planning program (Tien 1987; Henshaw 1990; Li et al. 1990; Tu and Smith 1995). The increases in the number of abortions in the 1970s were probably caused by both increasingly rapid socioeconomic development, especially in urban areas, and the implementation of the wan, xi, shao campaign (Tien 1987; see Chapter 1 of this volume for more discussion). As mentioned in the previous paragraph, with the implementation of the one-child policy in 1979, the number of abortions performed has risen rapidly. And the number has varied with changes in population policies. Generally speaking, China’s population policies have changed from a strict and centrally enforced policy implementation in the late 1970s and early 1980s to a more decentralized policy implementation characterized by local family planning regulations since the mid-1980s. These changes have affected the incidence of abortion in China (Greenhalgh 1986). For instance, according to HardeeCleveland and Banister (1988), in 1979, 7.8 million abortions were performed in China. In 1982, it was 12 million, and in 1983 when the policy was the strictest, it reached a peak of 14.4 million; the number dropped to about 9 million in 1984 when rural couples with one daughter were allowed to have a second birth; the number rebounded to about 12 million in 1986. Over the 1970s and 1980s, 200 million abortions are estimated to have been performed in China (Zeng 1991).

Patterns of induced abortion

23

Reasons for and patterns of abortion in China This section of the chapter focuses on abortion reasons that are specific for Chinese women in the special context of China’s population policy. One major reason for a Chinese woman to have an abortion is compliance with China’s family planning policy. The one-child population policy in China regulates not only the family size that a couple can have but also the spacing between children (in most cases, a four-year period between the first and the second child is required). When couples fail to meet those regulations, they are often encouraged to seek an abortion (Tu and Smith 1995). For some women, abortion has been used as a primary means of birth control, even though the state strongly discouraged the reliance on abortion because of the risks to women’s health and its high costs relative to contraception (Greenhalgh 1986). The state preferred couples to use sterilization or reliable contraception, but women with unauthorized pregnancies were sometimes encouraged to have abortions, especially during those periods when the enforcement of family planning policy was very strict. Cadres in some provinces encouraged women to have abortions because their work units would be fined if the birth quota assigned to them was exceeded (Greenhalgh 1986; Hardee-Cleveland and Banister 1988). As Tu and Smith note, under the family planning responsibility system, abortion serves as the main alternative that families can use to avoid penalties for an unplanned birth in cases of contraceptive failure, a last protection for local authorities against penalties for the failure to meet their preset demographic target, and a sensitive indicator of shifts in official family planning policy and its implementation at national and local levels. (1995:279) But Rigdon (1996) points out that the extent to which coercion was used varied widely from place to place; some villagers said that they never experienced it while others have reported a variety of coercive practices. Son preference in China is another major reason for ending a pregnancy with an abortion. Son preference is a product of the Confucian tradition that emphasizes the value of sons. There are a variety of historical, moral and ethical, and economic factors underlying the existence of son preference in China (Arnold and Liu 1986; Zeng et al. 1993; Poston et al. 1997; Zeng and George 2000). In the past, family structure in China was patriarchal. This patriarchal structure and the resulting preference for sons were so strong that it was even written in the Book of Rites that “A woman is to obey her father before marriage, her husband during married life, and her son in widowhood” (Arnold and Liu 1986:226). Chinese traditions also stress the importance of carrying on the family line through male progeny. Sons are desired mainly because of their economic value to the family, old-age support to their parents, and the provision of labor in farming or family business (Arnold and Liu 1986). With the implementation of the one-child policy, many couples could not have as many children as they desired. However, the ideology of son preference and the economic value of sons urged many couples to have at least one son in the family, especially in rural areas (Arnold and Liu 1986; Poston et al. 1997; Zeng and George 2000).

Fertility, family planning, and population policy in china

24

The existence of sex preference by itself will not influence fertility unless technologies that translate the gender preference to fertility modification are applied. Modern technologies like ultrasound and amniocentesis have enabled many Chinese parents to determine the sex of the fetus. Consequently, there have been reports of sex-selective abortions in China, that is, couples would undergo an abortion if they found out the fetus is female (Hull 1990; Zeng et al. 1993). The practice of sex-selective abortion is thought to be a reason for the rising sex ratios at birth in China since the mid-1980s (Zeng et al. 1993). The true sex ratio at birth does not vary much in most human societies. Countries with reliable data show that the sex ratio at birth is around 104–107 male births for every 100 female births (Chahnazarian 1991). Nevertheless, prenatal sex identification followed by gender-specific induced abortion is a main reason for the abnormally high sex ratio at birth and will have serious social consequences (Zeng et al. 1993; Gu and Xu 1994; Gu and Roy 1995; see Chapter 12 in this volume by Poston and Glover for more discussion). The third reason for Chinese women to seek abortion is social norms. There are various social taboos against single or widowed women bearing children. Not only do social norms pressure those pregnant women to have an abortion, but they also are officially sanctioned because family planning regulations do not allow births to single women (Rigdon 1996). With increasing exposure to Western lifestyles, the loosening of restrictions on social behavior, especially in large urban areas, and inadequate sex education among young Chinese women have contributed to increases in the number of pregnancies outside marriage that often end in abortion. Although abortion has increased across China since the implementation of the onechild policy in 1979, some differences may be observed at the regional level. Specifically, the inland areas often have lower abortion rates than the wealthier coastal provinces. This is partly due to the large concentration of minorities in the inland areas for whom the Chinese government has relaxed its population control policies (Chhabra 1984). Also urban areas are found to have higher abortion rates than rural areas (Kang 1991; Tien et al. 1992). In urban areas, the reported percentage of all pregnancies terminated by abortion rose from less than 3 percent in 1960 to about 30 percent in 1987; in rural areas, it rose from virtually zero to more than 15 percent (Kang 1991). Some studies show that education, age, marriage duration, and residence have apparent effects on abortion order (Li et al. 1990). Based on data from the 1982 One-Per-Thousand Fertility Survey, Arnold and Liu (1986) studied pregnancy outcomes by family composition and found that couples without a son were much less likely to have an abortion than couples with at least one son (Tu and Smith 1995).

Data and methods In this chapter, data are used from the 1997 Sample Survey of Population and Reproductive Health in China conducted by the China Population Information and Research Center and the State Family Planning Commission of China. The survey randomly interviewed 15,213 women between the ages of 15 and 49. Basic demographic data such as age, ethnicity, residence, marital status, and education were collected for each woman. In addition, each woman’s pregnancy histories were recorded in great

Patterns of induced abortion

25

detail, including information on the ending time of each pregnancy, the outcome of each pregnancy, the health status of each live birth, and birth parity. Every pregnancy was coded with six outcome categories: son, daughter, miscarriage, stillbirth, abortion, and currently pregnant. This survey also includes information about a woman’s attitude towards the ideal number of children and the sex composition of this desired number of children. In this chapter, for ever-married Chinese women with at least one child, their hazard of having an abortion is modeled for the pregnancy following their first child. Several restrictions have been imposed on the sample. Specifically, the woman’s hazard of having an abortion is examined for the pregnancy following her first child. Women who had only one child and were pregnant at the time of the survey were excluded from the study. Furthermore, the number of months between the birth of the first child and the second birth should usually not equal zero. When this occurred, it is assumed that the woman had multiple births, and they too are excluded from the analysis. The last restriction is that women under the age of 20 are excluded. This constraint was introduced because on one hand, women’s educational status, one of the independent variables, is not settled until they reach 20 years of age (reasons are specified later in the discussion of the education independent variable), and on the other hand, China’s Marriage Law requires that the minimum legal marriage age is 20 for women (China Population Information and Research Center: http://www.cpirc.org.cn/). The effects of individual characteristics and population policy on a woman’s hazard of having an abortion are analyzed. The dataset consists of ever-married women with at least one child, a total of 11,762 married women. Cox proportional hazard models are estimated to predict the hazard of having an abortion for the pregnancy following the first child.

Hazard models The most satisfactory way statistically to model either the occurrence or non-occurrence of an event, and its timing, is via hazard analysis. The term hazard rate, which refers to the rate at which the event of interest occurs, comes from the biomedical sciences’ use of survival analysis, that is, the surviving the hazard of a death (Yamaguchi 1991; Kleinbaum 1996). Hazard models are concerned with the patterns and correlates of the occurrence of events. Death, disease incidence, or any designated experience of interest that may happen to an individual, such as an abortion, would be an event. And “the occurrence of an event assumes a preceding time interval that represents its nonoccurrence” (Yamaguchi 1991:1). Specifically, there must exist a certain time period or duration of nonoccurrence in order for an occurrence to be recognized as an “event” (Yamaguchi 1991:1). Hazard analysis is used to examine duration data, which represent the nonoccurrence of a given event. Yamaguchi (1991:2) points out that “in conceptualizing the duration of the nonoccurrence of a given event, another important concept is the risk period.” Generally, the time period that represents the nonoccurrence of a given event can be divided into two parts: the period at risk and the period not at risk of experiencing the event. So specifically, hazard analysis is the analysis of the duration for the nonoccurrence of an

Fertility, family planning, and population policy in china

26

event during the risk period, or the analysis of the rates of the occurrence of the event during the risk period. One reason why Cox models are so widely used is that the Cox proportional hazard model has advantages over other hazard models. The most significant advantage is that Cox models can assume time dependence without having to specify its form (Yamaguchi 1991). In the situation when there are only time-independent variables, that is, when the values of the independent variables do not change over time, the form that the Cox model assumes is represented as follows: log h(t)=log h0(t)+b1X1+b2X2+…+BiXi where log h(t) is the hazard rate (or hazard function) which is an unobserved value gauging the instantaneous probability that a subject will have the event of interest during the interval. It is a rate, rather than a probability. Thus the values of the hazard function range between 0 and infinity; h0(t) is any function of time t; and X1 to Xi are independent variables. This function of time (h0(t)) does not have to be specified, and is the characteristic that makes the Cox model so attractive. Another advantage of Cox’s method is that it has the ability to estimate stratified models, which permit the investigator to control for a categorical independent variable that may have a complicated form of interaction effects with time, without specifying the form of the interaction effects (Yamaguchi 1991). One assumption of the Cox proportional hazards model is that the hazard rate for any one individual is proportional to the hazard for any other individual (Yamaguchi 1991; Kleinbaum 1996). However, the importance of meeting this assumption has been debated in the literature. Allison states that “as models go, the proportional hazards model is extraordinarily general and nonrestrictive—the main reason for its popularity. Even when the proportional hazards assumption is violated, it is often a satisfactory approximation” (Allison 1985:38). One way to describe the survival-time data for the subjects is to graph their Kaplan– Meier (K–M) survivor function curve. The K–M survivor curve is an empirical plot showing the probabilities of surviving for each unit of time. It starts with a survival probability of 1 and then steps down to the other survival probabilities, as one moves from one ordered failure time to another (in these analyses, from one month to the next). Let nj represent the number of observations that have not failed and are not censored at the beginning of time period t; and dj represent the number of failures that occur to the observations during the time period t. The following formula is the K–M estimator of surviving beyond time t, and is the product of survival probabilities in t and the preceding periods:

Figure 2.1 graphs S(t) against the number of months between the birth of the first child and the occurrence of the event of interest, abortion, in separate curves for all rural and urban women with at least one child. The K–M survivor curves show the probabilities of surviving the hazard of having an abortion following a woman’s first child for each group of women. For urban women, the curve steps down gradually from a probability of near 1.0 of surviving the hazard of having an abortion, to a probability of 0.6 by about the

Patterns of induced abortion

27

180th month, leveling off for the remainder of the risk period. The analysis to follow focuses on all women. Dependent variable The survival-time data for ever-married Chinese women consist of two variables. Abortion1 is a dummy variable indicating for each woman whether or not the event of interest, abortion for the pregnancy following the first child, occurred during the observation period. This is coded 1 if the pregnancy was aborted and 0 if otherwise. The data show that among women with at least one child, 22 percent

Figure 2.1 Probability of surviving the hazard of an abortion after the birth of the first child, stratified by rural and urban residence, China, 1997. Table 2.1 Descriptive data: 11,762 women with at least one child, China, 1997 Variable Abortion1

Mean

Standard deviation

Minimum value

Maximum value

0.22

0.41

0

1

Time1

37.40

40.15

0

351

Rural

0.78

0.42

0

1

Han

0.91

0.29

0

1

After80

0.57

0.50

0

1

Fertility, family planning, and population policy in china

28

Girl

0.49

0.50

0

1

Illiterate

0.25

0.43

0

1

Elementary

0 32

0.47

0

1

Junior

0.29

0.45

0

1

Senior

0.14

0.35

0

1

Number_ preference

1 86

0.74

0

1

35 10

7.79

20

49

North

0 14

0.35

0

1

Northeast

0.09

0.28

0

1

East

0.29

0.46

0

1

Mid-South

0.26

0.44

0

1

Southwest

0 15

0.36

0

1

Northwest

0.06

0.24

0

1

Age

Source: The 1997 sample survey of population and reproductive health in China.

had an abortion following the birth of their first child (see Table 2.1). The second variable, Time1, refers to the amount of time since the birth of the first child and the abortion or the censoring event. Independent variables The variable, Girl, represents the gender of the first child. This variable is intended to measure a woman’s sex preference for her children. If there is no sex preference in the society, there should be no association between the sex of an already born child and the likelihood of having another child. Since various studies point out the existence of son preference in China, the general hypothesis in this chapter is that controlling for several other independent variables, having a boy as the first child will increase the hazard of ending the following pregnancy with an abortion (see Chapter 3 by Liu and Chang in this volume). If there was no or only minimal sex preference in China, the sex of the first child should not be associated with the hazard of having an abortion. The Girl variable is a dummy variable coded 1 if the first child is a girl and 0 if the first child is a boy. The education level of a woman has been found to be closely associated with her fertility attitudes and behavior. Previous studies find that in most societies a woman’s educational attainment is positively associated with her probability of having an abortion (Powell-Griner and Trent 1987; Ahiadeke 2001). In this analysis, four dummy variables are used representing a woman’s educational attainment, namely, illiterate or semiilliterate (Illiterate), elementary (Elementary), junior high school (Junior), and senior high school (Senior). Each variable is coded 1 if yes, 0 if no, and Senior is the reference group. It is hypothesized that other things being equal, illiterate women, women with

Patterns of induced abortion

29

elementary education, and women with junior high school education are less likely to have an abortion, compared to the reference group, that is, women with senior high school education or above. Also, as already mentioned, the data contain detailed information about a woman’s attitudes toward the ideal number and sex composition of her children. One question asks about the number of children a woman deems to be ideal. The responses are categorized into no child, 1 boy, 1 girl, 1 child regardless of sex, 1 boy and 1 girl, 2 boys, 2 girls, 2 children regardless of sex, at least 1 boy, at least 1 girl, at least 1 boy and 1 girl, the more the better, do not care, others, and no idea. Using these 15 categories, one variable has been constructed to estimate a woman’s attitude about the ideal number of children, named Number_preference. It is coded 0 if a woman prefers no children, 1 if she prefers 1 child, 2 if she prefers 2 children, 3 if she feels the more, the better, and 4 if she does not care. It is hypothesized that other things being equal, the more children a woman desires, the lower the hazard of having an abortion. This is because if women desire more children or do not care, they are less likely to terminate a pregnancy with abortion, even if they are faced with fines or other punishments. Three independent variables are used that are related to China’s population policy. The first is a woman’s residence. It is termed Rural, a dummy variable, coded 1 if a woman lives in rural areas and 0 if she lives in urban areas. The hypothesis is that other things being equal, living in a rural area decreases the hazard of having an abortion for the pregnancy following the first child, compared to living in an urban area. Another policy-related variable is the ethnicity of a woman. In China, population policies for the most part have been relaxed for the minority nationalities. In most cases, minorities are allowed to have 2 children (China Population Information and Research center: http://www.cpirc.org.cn/). Minority couples may have 2 children and under special circumstances, they may even have 3 (Heberer 1989). Members of the Han nationality are subject to a stricter policy. A dummy variable, Han, is used that is coded 1 if a woman is Han and 0 if not. It is hypothesized that other things being equal, being a Han increases the hazard of having an abortion for the pregnancy following the first child, compared to being a non-Han woman. Another important policy-related factor is the time of the pregnancy following the first child. As mentioned earlier, China started its one-child policy in 1979; therefore pregnancies after 1979 were under the influence of the policy. Allowing some time for the policy to be diffused, the time of January 1980 is used. A dummy variable named After80 is introduced which is coded 1 if a woman’s first child was born after January 1980, and 0 if otherwise. It is hypothesized that other things being equal, a pregnancy following the birth of the first child that occurred after the policy is more likely to be terminated with an abortion than the pregnancy that occurred before the population policy. Women’s age is also one of the factors among the intermediate variables affecting fertility (Bongaarts 1978). It is known that in most societies, the age-specific fertility rate first rises and then falls, with the age of the mother. Therefore, age of the women at the time of the survey is also used as a control variable. Six regional dummy variables are also introduced as control variables. China has six very large regions, namely, Dongbei (Northeast) which includes Liaoning, Jilin, and Heilongjiang provinces, Huadong (East) which includes Shanghai city, Jiangsu, Zhejiang,

Fertility, family planning, and population policy in china

30

Anhui, Fujian, Jiangxi, and Shandong provinces, Huabei (North) which includes Beijing city, Tianjin city, Hebei, Shanxi, and Inner Mongolia provinces, Xinan (Southwest) which includes Sichuan, Guizhou, Yunnan, and Tibet provinces, Xibei (Northwest) which includes Shaanxi, Gansu, Qinghai, Ningxia and Xinjiang provinces, and Zhongnan (Mid-South) which includes Henan, Hubei, Hunan, Guangdong, Guangxi, and Hainan provinces. There are considerable differences in socioeconomic and demographic development among the six regions. The East is more developed, particularly along the east coast, and the Northwest and Southwest regions are less developed (Poston and Jia 1990). Also the eastern region is more rigorous in implementing the population policy. Each of the six region variables indicating where a woman resides is scored 1 if she resides in that region, and 0 if not. The East region will be used as the reference group. Table 2.1 lists the means, standard deviations, minimum, and maximum values of each variable in the dataset containing 11,762 women with at least one child. Rural women account for 78 percent of the sample. Also Han women dominate the sample. Among the first-born children, 49 percent were girls, and 57 percent of them were born after the one-child policy (After80). One quarter of the women are illiterate, 32 percent have an elementary education, 29 percent have a junior high school education, and 14 percent have a senior high school education and above. On average, most women prefer less than two children. Also it can be seen that the average age of women with at least one child is 35, and women are unevenly distributed among the six different regions of China.

Results Table 2.2 reports the results of eight nested hazard models. From Model 1, the most basic model, to Model 8, the complete model, are added increasing numbers of control variables. The results reported in the eight models are consistent and support the hypotheses. Other things being equal, having a daughter as a woman’s first child significantly lowers her likelihood of having the following pregnancy aborted; an illiterate woman has a significantly lower hazard of having an abortion for the pregnancy following a woman’s first child, compared to a woman with senior high school education or above; the more children a woman prefers to have, the lower her hazard of having an abortion for the pregnancy following her first child; living in rural areas reduces a woman’s hazard of having an abortion for the pregnancy following her first child; a Han woman is significantly more likely to end the pregnancy following her first child with abortion; and a pregnancy occurring after the implementation of the one-child policy has a much higher hazard of being terminated with abortion.

Patterns of induced abortion

31

Table 2.2 Cox proportional hazards models for the hazard of having abortion for the pregnancy following a woman’s first child, China, 1997 Model Model Model Model Model Model Model 7 Model 8 1 2 3 4 5 6 Girl

Illiterate

Elem entary Junior Senior Number_ prefe rence

Age

Rural

Han

After80

North

Northeast East

Mid-South Sout

−0.08* −0.10* −0.09* −0.09* −0.09* −0.09* −0.23*** −0.21*** (0.92) (0.90) (0.91) (0.91) (0.91) (0.91) (0.80) (0.81) −1.46 −1.39 −1.39 −0.99 −0.97 −0.90 −0.9 *** *** *** *** *** *** 9*** (0.23) (0.25) (0.25) (0.37) (0.38) (0.40) (0.37) −0.84 −0.79 −0.79 −0.44 −0.43 −0.38 −0.43 *** *** *** *** *** *** *** (0.65) (0.69) (0.43) (0.46) (0.46) (0.64) (0.65) −0.1 −0.17** −0.41 −0.40 −0.40 −0.19 −0.19 4** *** *** *** *** *** (0.66) (0.67) (0.67) (0.82) (0.82) (0.87) (0.84) Refer Refer Refer Refer Refe Refer Refere ence ence ence ence rence ence nce −0.19 −0.19 −0.15* −0.15 −0.29 −0.27 *** *** ** *** *** *** (0.82) (0.82) (0.86) (0.86) (0.75) (0.76) 0.00 −0.01* −0.01* −0.03 −0.03 *** *** (1.00) (0.99) (0.99) (0.97) (0.97) −0.53 −0.52 −0.85 −0.86 *** ** *** *** (0.59) (0.59) (0.43) (0.42) 0.33 0.45 0.37 *** *** *** (1.39) (1.57) (1.45) 2.82 2.82 *** *** (16.77) (16.86) −0.45 *** (0.64) −0.23 ** (0.79) Refer ence −0.38 *** (0.69) −0.09 (0.91)

Semistand ardized 0.90

0.65

0.82 0.92

0.82

0.79

0.69

1.11

4.11

0.86 0.94

0.85 0.97

Fertility, family planning, and population policy in china hwest North west Pseudo R2 Final log

32 −0.56* ** (0.57)

0.00 0.01 0.01 0.01 −22,26 −21,99 −21,97 −21,97 2.18 3.74 4.47 4.45

0.02 0.02 −21,9 −21,91 21.63 3.93

0.08 −20,54 4.03

0.87

0.08 −20,49 7.31

Note ***p<0.001, **p<0.01, *p<0.05.

The results from the given eight models dealing with the hazard of having an abortion for the pregnancy following a woman’s first child may be characterized as follows. First of all, population policy factors play a very important role in affecting a woman’s hazard of having an abortion for the pregnancy following her first child. This can be seen in three ways. First, all the population policy factors are statistically significant. Second, the relative importance of those factors is high, as viewed by their high semi-standardized hazard ratios (last column of Table 2.2). Third, the pseudo R-squared statistic in models with only the population policy factors is higher than the models with only individual factors. Compared to policy factors, individual factors appear to have less of an impact on a woman’s hazard of having an abortion for the pregnancy following her first child. These results indicate that although individual factors do affect the abortion hazard, women are indeed influenced as well by the one-child policy. This is particularly apparent with the changes in the hazard ratio for the Girl variable. When policy factors are added into models with only the individual factors, the significance level of the Girl variable increases. This indicates that the one-child policy intensifies the effect of the sex of the first child on the hazard of having an abortion for the pregnancy following a woman’s first child.

Conclusion The results of this chapter have several implications. First, fertility behavior among Chinese women is influenced by individual characteristics, and most obviously and importantly, by the government population policies. It can be seen that population policy in China has been instrumental in affecting couple’s reproductive behavior. Although individuals vary in terms of their gender preferences for children, levels of education, and preferences for an ideal number of children, their personal characteristics are less likely to be an obstacle to the implementation of the population policy. Furthermore, the findings suggest that the residence of a woman plays an important role in influencing her hazard of abortion. Specifically, rural women have a lower hazard of having an abortion, and this finding is consistent with previous studies (Henshaw and O’Reilly 1983; Powell-Griner and Trent 1987). Although the abortion patterns of rural and urban women in China are similar to those of women in the United States, the main reasons in China appear to be quite different from those in the United States. Due to the existence in China of extensive family planning service stations, there are almost no

Patterns of induced abortion

33

differences in the access to abortion for rural and urban women as there are for women in the United States. The reason why rural women have a lower hazard of having an abortion compared to urban women is mainly due to the fact that the one-child policy has been more strictly enforced in urban areas than in rural areas. This is in part due to the greater ease with which economic and social sanctions can be imposed in urban areas. For instance, urban couples may lose their jobs or be demoted if they violate the population policies. Among the individual characteristics, a woman’s gender preference for children exerts an important influence on the hazard of having an abortion. The findings in this chapter showed that son preference exists in both rural and urban areas, which seems to be inconsistent with You’s claim that “the traditional Chinese belief in ‘more sons, more happiness’ has apparently begun to fade somewhat among urban residents as more young couples have accepted the notion of having only one child with ‘more intelligence, more health’” (1993:58). In fact, the findings here demonstrate that son preference has actually been intensified for both rural and urban women under the one-child policy. As noted previously, the issue of abortion is closely related to and strongly affected by government policies. Another suggestion for the family planning policy is that the family planning program “should shift its policy directions from the current quantitative fertilityreduction oriented policies to qualitative health and welfare approach which emphasizes a balanced sex ratio and prevention of induced abortion” (Cho and Ahn 1993:67). Also education and the dissemination of information can “contribute significantly to a better understanding among policymakers and program officers of population issues and to a better balance between the national demographic objective and individual well-being” (Tu and Smith 1995:284). Family planning might be better designed to improve women’s status, health, and welfare. Family planning workers could invoke strategies to educate and motivate women to use contraception properly. Some encouraging indications suggest that such development has been taking place in China and might gain momentum in the future (Tu and Smith 1995).

References Ahiadeke, C. 2001. “Incidence of Induced Abortion in Southern Ghana.” International Family Planning Perspectives, 27:96–101, 108. Allison, P. 1985. Event History Analysis: Regression for Longitudinal Event Data. Beverly Hills, CA: Sage Publications. Arnold, F. and Z.Liu. 1986. “Sex Preference, Fertility, and Family Planning in China.” Population and Development Review, 12:221–246. Bongaarts, J. 1978. “A Framework for Analyzing the Proximate Determinants of Fertility.” Population and Development Review, 4:105–132. Chahnazarian, A. 1991. “Determinants of the Sex Ratio at Birth: Review of the Recent Literature.” Social Biology, 35:214–235. Chhabra, R. 1984. “China: Our Arduous Task” People, 11:16–17. China Population Information and Research Center. 1997. http://www.cpirc.org.cn/ Cho, N. and N.Ahn. 1993. “Changes in the Determinants of Induced Abortion in Korea.” Journal of Population, Health and Social Welfare, 13:67–79. Frejka,T. 1985. “Induced Abortion and Fertility” Family Planning Perspectives, 17:230–234.

Fertility, family planning, and population policy in china

34

Greenhalgh, S. 1986. “Shifts in China’s Population Policy, 1984–1986: Views from the Central, Provincial, and Local Levels.” Population and Development Review, 12:491–515. Gu, B. and K.Roy. 1995. “Sex Ratio at Birth in China, with Reference to Other Areas in East Asia: What We Know.” Asia-Pacific Population Journal, 10:17–42. Gu, B. and Y.Xu. 1994. “A Comprehensive Discussion of the Birth Gender Order Ratio in China.” Chinese Journal of Population Science, 6:417–431. Hardee-Cleveland, K. and J.Banister. 1988. “Fertility Policy and Implementation in China, 1986– 88.” Population and Development Review, 14:245–286. Heberer, T. 1989. China and Its National Minorities: Autonomy or Assimilation? Armonk, NY: M.E.Sharpe, Inc. Henshaw, S. 1990. “Induced Abortion: A World Review, 1990.” Family Planning Perspectives, 22:76–89. Henshaw, S. and K.O’Reilly. 1983. “Characteristics of Abortion Patients in the U.S., 1979 and 1980.” Family Planning Perspectives, 13:6–18. Hull, T. 1990. “Recent Trends in Sex Ratios at Birth in China.” Population and Development Review, 16:63–83. Kang, X. 1991. “Dynamics of Abortion Among Married Women in China and the Main Causes.” Chinese Journal of Population Science, 3:315–325. Kleinbaum, D. 1996. Survival Analysis: A Self-Learning Text. New York: Springer. Knodel, J., A.Chamratrithirong and N.Debavalya. 1987. Thailand’s Reproductive Revolution. Madison, WI: University of Wisconsin Press. Li, V., G.Wong, S.Qiu, F.Cao, P.Li, and J.Sun. 1990. “Characteristics of Women Having Abortion in China.” Social Science and Medicine, 31:445–453. Poston, D.L. Jr 1988. “Fertility and Family Planning in China: An Analysis of Provincial Patterns.” Journal of Biosocial Sciences, 20:379–391. Poston, D.L. Jr and B.Gu. 1987. “Socioeconomic Development, Family Planning, and Fertility in China.” Demography, 24:531–551. Poston, D.L. Jr and Z.Jia. 1990. “Socioeconomic Structure and Fertility in China: A County Level Investigation.” Journal of Biosocial Science, 22:507–515. Poston, D.L. Jr, B.Gu, P.Liu, and T.McDaniel. 1997. “Son Preference and the Sex Ratio at Birth in China: A Provincial Level Analysis.” Social Biology, 44:55–76. Powell-Griner, E. and K.Trent. 1987. “Sociodemographic Determinants of Abortion in the United States.” Demography, 24:543–560. Rigdon, S. 1996. “Abortion Law and Practice in China: An Overview with Comparisons to the United States.” Social Science and Medicine, 42:543–560. Tien, H. 1987. “Abortion in China: Incidence and Implications.” Modern China, 13: 441–468. Tu, P. and H.Smith. 1995. “Determinants of Induced Abortion and Their Policy Implications in Four Countries in North China.” Studies in Family Planning, 26:278–286. United Nations Economic and Social Commission for Asia and the Pacific. Available http://www.unescap.org/ Yamaguchi, K. 1991. Event History Analysis. Newbury Park, CA: Sage Publications. You, J. 1993. “Does the Gender of the Child Affect Acceptance of the One-child Certificate? The Case of Shaanxi Province, China.” Asia-Pacific Population Journal, 8:47–59. Zeng, Y. 1991. “A Note on the Numbers of Induced Abortions and Averted Births.” China Population Today, 3:9–11. Zeng, Y. and L.George. 2000. “Family Dynamics of 63 Million (1990) to More Than 330 Million (in 2050) Elders in China.” Demographic Research, 2, Article 5. May. http://www.demographic-research.org/ Zeng, Y., P.Tu, B.Gu, Y.Xu, B.Li, and Y.Li. 1993. “Causes and Implications of the Recent Increase in the Reported Sex Ratio at Birth in China.” Population and Development Review, 19:283–302.

3 Patterns of sterilization Can Liu and Chiung-Fang Chang In the last two decades, China has been a major focus of demographic study. This has occurred not only because China has the largest population of any country in the world, but also because of the dramatic demographic transition China has experienced since the late 1970s. As was noted in Chapter 2, China’s total fertility rate (TFR) dropped from 6.5 in the 1960s to 1.8 in the year 2000 (China Population Information and Research Center: http://www.cpirc.org.cn/). Although socioeconomic development has been a contributing factor, scholars generally agree that much of the credit is due to the central government’s family planning programs (Poston 1986; Wolf 1986; Feeney and Wang 1993). Since the founding of the People’s Republic of China (PRC) in 1949, the country has undergone many changes in arriving at its current population policy and birth control system (see Liang and Lee’s Chapter 1 in this volume). Research on China’s population dynamics has recognized that contraceptive prevalence, that is, the percentage of married women currently using contraception, is strongly related to the level of fertility. Local leaders, party officials, family planning workers, neighbors, and colleagues work together through ideology and persuasion to ensure that women of the reproductive ages adhere to local family planning policies. In factories, family planning workers carefully watch women to make sure they do not have unplanned pregnancies. In neighborhoods, elderly women known as the “Granny Police” make frequent home visits to ensure that women are using contraception, and sometimes even listen in on their private conversations. Women who use the intrauterine device (IUD) are requested to have regular checks to make sure the coil is properly placed. Family planning policies explicitly regulate not only the numbers, timing, and spacing of children that a couple may have, but local family planning workers frequently impose the contraceptive method a couple should use to prevent unplanned pregnancies and births (Tu 1995). As a result, sterilization is chosen by a large percentage of married Chinese couples and deserves much of the credit for China’s progress in fertility reduction. Starting in 1982, in order to effectively control its population, China adopted a birth control policy of “IUD insertion after the first birth and sterilization after the second birth” in the vast rural areas (State Family Planning Commission of China 2002). The major focus of this chapter is on sterilization, the most permanent form of contraception. It is of particular interest in China because it is strongly promoted by the government. The chapter also discusses some of the implications of the analyses for family planning programs, medical services, and other social development issues in China.

Fertility, family planning, and population policy in china

36

Literature review China’s birth control policies The mass production and distribution of contraceptives in China began as early as 1958. At the time, the only modern contraceptives available were diaphragms and condoms, which were often of poor quality. Even so, the mass produced supply of contraceptives was only enough to meet the need of about 2 percent of all couples in the childbearing ages. Nonetheless, both abortion and sterilization had been legalized in 1957 (Banister 1987). A strong national population control framework was developed in the early 1970s, specifically, the “wan-xi-shao” policy which stressed later marriage (wan), longer intervals between births (xi), and fewer children (shao) (Tien 1980; Feeney and Wang 1993). It recommended 2 children for urban couples and 3 for rural couples (see Liang and Lee’s Chapter 1 in this book for more discussion). At that time, a comprehensive health system was not available in most rural areas, so a “barefoot doctor” health care system was promoted. “Barefoot doctors” were not formal medical school graduates, but young people, usually women, with high school, or even junior high school educations, who had received some basic medical training (Hartmann 1987). It was targeted to have at least one barefoot doctor in every county health station. These women received specific training in contraceptive counseling and techniques. Typically, they were trained to insert and remove IUDs, perform abortions, deliver babies, provide contraceptive education, and maintain birth-planning records (Chen 1985). With the introduction of target population growth rates in different levels of the government in the mid-1970s, individual incentives for sterilization became part of the campaign routine. The decision of when to have a child became a public affair, and involved getting permission to become pregnant from local-birth planning units (Hartmann 1987:149–151). The one-child family policy, introduced in late 1979, urged women to cease childbearing after the birth of the first child (Feeney and Wang 1993:71). Theoretically, the nationwide birth control system made all contraceptive methods available to all married couples. In practice, the family planning programs promoted the IUD and sterilization as the primary contraceptive methods (Chen 1985:139). It is obvious that the control of population growth requires the use of effective and long-lasting contraceptive methods. The most effective contraceptive method is sterilization. According to survey data gathered in China in 1982, more than one-half of all contraceptive users relied on the IUD, and sterilized women and men accounted for more than one-third of China’s contraceptive users (Chen 1985:139). By 1987, national contraceptive prevalence levels among married women of childbearing age were reported to be as high as 77 percent, with sterilization and the IUD being the most used methods (State Family Planning Commission 1988). Since 1991, the central government has requested that chief officers at different levels of administrative units be fully responsible for maintaining the objectives of the family planning program. The officers’ chances of promotion and honor were determined by whether the target of the family planning

Patterns of sterilization

37

program was reached or not. This so called “one-vote-down” (yi-piao-fou-jue) campaign helped to increase the financial support of family planning from the different levels of governments, especially in less developed rural areas. With stronger material backup and more comprehensive family planning services, fertility decreased. This arbitrarily implemented system made some of the official workers try every avenue, including coercion, to meet the birth control assignments (Zeng 1996:29).

Sterilization in urban and rural China While it is true that many urban couples have pledged to have only one child, the onechild policy has not been strictly implemented in the rural areas (Kaufman et al. 1989). The government not only requires couples to practice birth control, but also often tells them what type of contraceptive method they must use. The policy became nationwide in scope when the government announced in 1982 that women with one child must use the IUD, and couples with two children must have one partner sterilized. In urban areas, the supervision of women of childbearing ages is accompanied by strong threats of penalties for attempted violations. Hence, near complete conformity has been achieved in urban areas (Hardee-Cleaveland and Banister 1988). Prior to the family planning campaign, sterilization was primarily accepted by older couples in rural areas who had completed their childbearing with the birth of a third or fourth child (Davin 1985:47). This changed during the 1980s when a second birth was allowed only for daughter-only couples, and having a third birth clearly constituted a violation of family planning policy in vast rural areas of the country. A study of 4 rural counties in the late 1980s showed that 76 percent of women with 1 child were using IUDs, and two-thirds of the women with two children were sterilized (Kaufman et al. 1989:725). Although the government continues to advocate one child per couple, “rural couples who have practical difficulties and who wish for a second child are encouraged to space their children properly” (Cooney and Li 2001:68). Among rural women with two children, those whose second birth was a birth-quota violation were more likely to be sterilized than those who had received permission to have a second child. This suggests that in rural areas, sterilization tends to be a sanction that the government imposes upon couples who violate the family planning policy. For women with three children, sterilization as a sanction for policy violation is frequently the result (Cooney and Li 2001:72–73). Consistent with these findings, Zhou’s study of Chinese women’s contraceptive patterns found that the general contraceptive rate of urban women was higher than that of rural women. The IUD was the most widely chosen contraceptive method among urban women, with sterilization being second; the reverse was the case for rural women (Zhou 1991). Differential contraceptive patterns for the majority and minority According to China’s 2000 census, the 55 minority nationalities consisted of 97 million people and accounted for 7 percent of the population, while the Han majority accounted for the remaining 93 percent (China Population Information and Research Center:

Fertility, family planning, and population policy in china

38

http://www.cpirc.org.cn/, see Chang’s Chapter 5 in this book for a more detailed discussion of minority fertility). There are considerable differences between Han and minority women in their use of contraceptive methods (Poston 1986). Minority women prefer to use the IUD, followed by sterilization and other contraceptives (Zhou 1991). Analysis based on the 1982 China One-Per-Thousand National Survey found that over 56 percent of minority women, and close to half of the Han women, used IUDs. Sterilization was used by almost 36 percent of the Han women, compared to slightly more than 21 percent of the minority women (Poston 1986). Quality of family planning service Family planning in China consists of more than simple propaganda and education campaigns using slogans and posters. Unremitting social pressure is often used to isolate the individual, leaving her little choice but to comply. Some critics of China’s family planning remind the public that women are pressured to use specific methods, mostly IUDs and sterilization, often without adequate counseling (Hardee-Cleaveland and Banister 1988). Field studies suggest that when women have no strong method preferences and information services are lacking, provider recommendations are the principal influence on contraceptive choices (Kaufman et al. 1992). Informed choices refer to the outcome of a dynamic process that is intended to ensure that clients are able to make decisions about their health care, particularly about family planning (AVSC International 1999). When individuals make informed choices about family planning, they are more likely to be satisfied with the chosen method. Informed choice is a crucial element of quality for all reproductive health and family planning services. Thus it is especially important for programs to be focused on providing information-based services. Since sterilization is a surgical and provider-dependent contraceptive method intended to be permanent and nonreversible, the importance of voluntary client choice is very important. Kaufman and her colleagues (1992) examined the quality of family planning services (the availability of contraceptive choice, information given to users, and provider knowledge about methods) in rural China. They found that improvements in quality, especially in mixed methods, provider’s level of knowledge, and the quality and quantity of information provided to users were all likely to improve contraceptive continuation, client satisfaction, and women’s health. Their findings suggest that in places where women were poorly educated about the characteristics and risks of various contraceptives, providers and clinics were influential regarding women’s choice of contraceptives. The poor information often offered to women included the fact that providers tend to use national guidelines instead of health considerations for prescribing methods. They tend to recommend the IUDs for birth spacing after the first birth, and sterilization after the second birth. Short and her colleagues (2000) note that one reason why sterilization may not be strongly related with other variables is the birth planning policy. As far as policy is concerned, after controlling for other factors, policy severity, as indicated by exceptions to the one-child policy, is negatively related to contraception.

Patterns of sterilization

39

Number and sex composition of living children Although the policy advocates sterilization as the preferred method for couples with a second birth, patterns of sterilization are not entirely parity dependent (Short et al. 2000). In fact, not all women with two children have undergone sterilization. China’s 1988 TwoPer-Thousand Fertility Survey revealed that almost 33 percent of the women who already had 2 babies choose other contraceptive methods over sterilization (Yang 1994:28). Cooney and Li (2001) found that among couples who already had 2 or even 3 children, the most important factor predicting sterilization was the number of living sons. Couples who already had 2 children with at least 1 son were twice more likely to be sterilized than were couples with no sons. Among couples who already had 3 children, almost twothirds of the women with no sons were not sterilized. These data suggest that in the rural areas local cadres work with couples to ensure that all families have a son. Short and her colleagues (2000) also found that over 75 percent of women with one child were using the IUD. The effect of education on contraceptive choice has been recognized for a long time. Zhong and his colleagues (2000) studied the contraceptive knowledge of married reproductive-age Chinese women. Their results showed that contraceptive knowledge was generally low, especially for women who were older, who resided in rural areas, who had low education, and who had a non-Han husband. Zhong suggested that the task of spreading contraceptive knowledge should be the future focus of the family planning program (see a related discussion in Chapter 10 in this book by Wang and You).

Data and methods The data used here are from the 1997 Sample Survey of Population and Reproductive Health in China conducted by the China Population Information and Research Center and the State Family Planning Commission of China (for more information see the discussion of this data-set by Wu and Walther in Chapter 2 of this volume.) In this data-set, each woman’s contraceptive histories were recorded in detail, including the currently used contraceptive methods, when sterilization occurred, and who made the contraceptive decisions. Questions regarding women’s contraceptive knowledge and their satisfaction toward family planning service were also asked. This chapter examines the hazard of becoming sterilized among currently first married women with at least one child. This risk population is chosen because sterilization is often encouraged by family planning officials once couples have met or exceeded their allowable number of children, usually one child per couple, with few exceptions. The total sample contains 10,406 currently first married women with at least one child. Divorced and widowed women are omitted because they are outside the risk period of sterilization, and remarried women are treated differently in the framework of the family planning policy. Hence, women who are remarried, divorced or widowed, and women who have not had any children, are excluded from the sample. To examine the hazard of being sterilized during a specific risk period, Cox proportional hazard model is estimated (see Wu and Walther’s discussion of Cox models in Chapter 2 of this book). The dependent variable is composed of two parts:

Fertility, family planning, and population policy in china

40

1 STERILIZE is a dummy variable, which measures whether or not the woman had a sterilization after her first baby’s birth. It is scored 1 if the woman experienced sterilization after the birth of her first baby by the time of survey, and 0 if not. About 38 percent of the women had been sterilized by the time of the survey. 2 TIME measures the number of years that elapsed between the birth of the first child and the event of sterilization, or a censoring event. The mean duration is 8.34 years, and ranges from 1 to 33 years (Table 3.1). The length of the duration is recorded from the time the woman entered the risk period of sterilization (at the birth of her first baby) until she experiences the event of interest (sterilization) or is censored. There are four ways of measuring time duration. (1) If by the time of the survey, the woman has not used any contraceptive methods after her first born

Table 3.1 Descriptive data of 10,406 currently married women with at least one child, China, 1997 Variable

Mean

Standard deviation

Minimum value

Maximum value

STERILIZE

0.38

0.49

0

1

TIME

8.34

6.27

1

33

AGE

35.51

7.59

15

50

EDUC

6.24

4.30

0

16

PLAN

0.78

0.42

0

1

MAGE

21.46

2.73

11

40

MINORITY

0.09

0.29

0

1

KNOW

1.13

1.21

0

4

CEB

1.91

0.97

1

9

SFCHO

0.43

0.50

0

1

WIFESTER

0.72

0.45

0

1

SATIS

0.72

0.45

0

1

BOYIN2

0.67

0.47

0

1

URBAN

0.21

0.40

0

1

baby but is still within her fecund age (<49), then duration is the number of years between the birth of her first baby and time of the survey (which is October 1997). (2) If the woman has chosen contraceptive methods other than female sterilization after her first-born baby but is still within her fecund age (<49), then duration is the number of years between the birth of her first baby and the time of the survey. (3) If the woman has not been sterilized but has already reached the age of 49 before the time of survey, then the time period is the number of years between the birth of her first baby and the year when she reached age 49. (4) If the woman chose female sterilization as her contraceptive

Patterns of sterilization

41

method after her first baby, but before reaching the age of 49, then the time period is the number of years between the birth of her first baby and the time of being sterilized. To better describe the survival time data for the subjects, Kaplan–Meier (K–M) survival curves are estimated. The K–M survival curve is an empirical plot showing the probabilities of surviving sterilization for each unit of time. In Figure 3.1, the K–M curve steps down gradually from a probability of near 1.0 of surviving the hazard of sterilization, to a probability of around 0.4 by about the fifteenth year, leveling off for the remainder of the risk period. This means that by 15 years after their first birth, 60 percent of the women have been sterilized. Figure 3.2 shows the K–M survival curves stratified for urban and rural women with at least one child. The rural women have a lower survivor curve than the urban women. The curves for rural women with at least one child drops much more quickly than that for the urban women, which indicates that rural women have a higher hazard of experiencing sterilization after their first baby birth than do urban women. Twelve independent variables are used to examine possible influences on the hazard of sterilization; these include 9 socio-demographic characteristics of the women and 3 features of family planning services.

Figure 3.1 Kaplan-Meier estimates of Chinese women surviving the hazard of sterilization, 1997.

Fertility, family planning, and population policy in china

42

Figure 3.2 Kaplan-Meier estimates of Chinese women surviving the hazard of sterilization, by residence, 1997. The nine socio-demographic characteristics are as follows: 1 AGE is the woman’s age at the time of the survey and ranges from 15 to 49; 2 EDUC measures the years of education the woman received and ranges from 0 to 16; 3 MAGE is the age at which the woman married. The mean marriage age is 21.5 years with a range from 11 to 40; 4 KNOW measures women’s knowledge of reproductive health. Four questions regarding reproductive health are used, including knowledge about the vulnerable period for women to get pregnant, the necessity and proper method of contraceptive in the breastfeeding period, and reasons for contraceptive failure. An accumulated score (from 0 to 4) on this variable is assigned to each woman depending on how many of four questions were answered correctly. Less than 4 percent of the women have a score of 4, while almost 42 percent of the women answered none of the 4 questions correctly; 5 WIFESTER is a dummy variable, which measures the woman’s personal view toward sterilization. It is coded 1 for women who think the wife should be sterilized, and coded 0 if they believe the husband should be sterilized or do not know. Almost 75 percent of women in the sample believe the wife should be sterilized; 6 CEB is the number of children ever born for each woman; 7 BOYIN2 is a dummy variable coded 1 if the woman already had a boy as either her first or second child, and 0 if not; 8 URBAN is a dummy variable indicating women’s residence, coded 1 is she resides in an urban area; and

Patterns of sterilization

43

9 MINORITY is a dummy variable indicating whether a woman is a member of a minority group, scored 1 if she is a minority. The three family planning quality variables are: 10 PLAN is a dummy variable measuring whether the woman’s first child was born after the implementation of the one-child policy. If a woman’s first birth occurred after 1979, it is coded 1, indicating that her contraceptive choice was likely influenced by the family planning policy; 11 SFCHO is a dummy variable indicating whether the woman’s contraceptive choice was made by herself or some other person; it is coded 1 if the woman chose the contraceptive method on her own will, and 0 if her choice was based on the advice of family planning workers, doctors, husbands, or other family members. Almost half of the women made their own choice of contraceptive methods; 12 SATIS is a dummy variable measuring the overall satisfaction of the family planning service the woman received, coded 1 if satisfied with her family planning services.

Results Table 3.2 presents the Cox proportional hazard estimates of the hazard of having sterilization, for those 10,406 currently married women with one live birth. Covariates are gradually added to the models to show their respective effects on a woman’s hazard of undergoing sterilization.

Table 3.2 Hazard ratios of female sterilization in China, 1997 Model 1

Model 2

Model 3

Model 4

Model 5

Standardized

AGE

0.99**

0.99**

0.97*

0.97**

0.98**

0.84

EDUC

0.95**

0.96**

0.98*

0.99**

0.99

0.98

MAGE

0.95**

0.97**

1.00

1.01

1.00

1.00

KNOW

0.84**

0.87*

0.90**

0.90*

0.88

WIFESTER

3.85**

3.82*

3.76**

3.73*

1.80

CEB

1.32*

1.31**

1.30*

1.28

BOYIN2

1.48*

1.45**

1.46*

1.20

URBAN

0.46**

0.48*

0.74

MINORITY

0.56**

0.57*

0.85

SFCHO

0.78*

0.88

PLAN

1.21*

1.08

SATIS

1.19*

1.08

Fertility, family planning, and population policy in china Overall R2

0.006

0.021

0.026

0.03

0.031

2

385.1

1,436.2

1,795.6

2,076.1

2,155.3

LR χ

44

Note **p<0.01

The findings of the full model (Model 5) suggest that women’s knowledge of contraceptive use (KNOW) and the attitude of who should undergo sterilization (WIFESTER) have significant effects on a woman’s hazard of experiencing sterilization. That is, women with greater contraceptive knowledge will have a lower hazard of experiencing sterilization than women with less contraceptive knowledge. Also, women who believe that the wife should be sterilized have a 273 percent greater risk of experiencing sterilization than women who believe that the husband should be sterilized or do not know. In addition, the number of children a woman has had ever born (CEB) and the sex of women’s first two children (BOYIN2) have significant and positive effects on the hazard of women’s sterilization. The more the children, the greater the risk of accepting sterilization. A woman who had a boy among her first two children has a 46 percent greater risk of experiencing sterilization than a woman who did not have a boy. The findings support previous literature demonstrating the significance of enforcement of family planning regulations and existence of son preference (Short et al. 2000; Cooney and Li 2001). Regarding the effects of family planning services, the three family planning services variables, SFCHO (women’s independent contraceptive choice), PLAN (whether a woman’s first child was born after the start of the one-child policy), and SATIS (satisfaction toward family planning service) all have significant effects on women’s hazard of sterilization (Table 3.2). The results show that women who made the contraceptive choice suggested by other people, whose first child was born after the implementation of the one-child policy, and who were satisfied with the family planning services, have a higher risk of being sterilized. The hazard ratio of education is significant in Model 1 through Model 4, but loses its significance in Model 5. The collinearity between the education variable and variables of family planning services may be the cause. Results from previous studies on the effects of education on fertility vary. Some suggest that highly educated women are more likely to comply with the policy (Cooney and Li 2001), while other studies find little effect of education and suggest that the fertility decline may be due more to women’s response to birth-planning policy (Freedman et al. 1988:54–55).

Discussion and implications Sterilization is the world’s most widespread form of birth control. In 1990, it was used by approximately 20 percent of all married couples of reproductive age. No one can deny the contribution of sterilization toward China’s fertility decline. Chinese women who chose sterilization account for about one-half of the world’s total number of women who are sterilized (Ross 1992). From a demographic perspective, the usual irreversibility of

Patterns of sterilization

45

sterilization most sharply distinguishes it from other methods of fertility control, and it is the most effective form of preventing pregnancy (Nortman 1980). Because of its finality, sterilization is considered to be the most effective means of preventing induced abortions and is a permanent cure for the extra-quota pregnancies assigned to China’s population policy-makers and family planning service (Banister 1987). With rapid economic development and the steady improvement of living standards, family planning, as a state policy, is becoming more widely accepted by the public. To cater to people’s increasing awareness of protecting their rights and interests, family planning policy-makers have begun to change their strategies by focusing more on voluntary choice. The findings reported in this chapter show that knowledge of contraception is a significant predictor of women resorting to sterilization as a method of contraception. Failure to provide sufficient contraceptive knowledge to women signifies that the quality of the family planning service needs some improvement. The education of couples about contraception and improved communication between clients and providers can be achieved. Thus the promise of “informed choice” can well be a reality instead of just an idea. There are three basic principles that can be generalized from “women-centered” approaches to reproductive health or contraceptive choice. First, women should have the ability and the authority to make decisions about reproduction. The results reported here reveal that inadequate contraceptive knowledge, and not making contraceptive choices independently, both contribute to female sterilization more than adequate contraceptive knowledge and independent contraceptive choices. Second, it is important to understand and address reproductive health from the perspective of women. In China’s specific case, instead of recommending contraceptives based solely on the interests of policy implementation, family planning workers should give more consideration to the women’s individual health conditions, personal preferences, and contraceptive histories. Third, reproductive health services should consider all the different levels of the environment (e.g. international, national, and local) in which the policies and programs are carried out. In China, although there is significant socioeconomic variation across the country, discrepancies between urban and rural remain after considering the differences of local policy strength and service quality. Given the limited delivery system of quality family planning information, and the predominance of son preference in rural areas, the family planning workers face greater difficulties when promoting and implementing the policy in the countryside. However, these difficulties can never be justified by adopting “highpressure” persuasion or “biased” recommendations of sterilization without comprehensive information to promote these long-term and provider-controlled ways of birth control. Derived from the general principles of the newly developed worldwide reproductive health strategy, the Chinese government set forth a framework for “information-based choice” services based on China’s national conditions. They stated that informed choice pertains to the process where regulations and administrative directives dictate what people should do. With this program, people can now make their own decisions as to whether to accept or reject services, what services to choose, what kind of contraceptives to use, and whether or not to use referral services. This new framework emphasizes three critical components of information-based choice: voluntarism, information, and choice.

Fertility, family planning, and population policy in china

46

These three factors not only constitute the “information-based choice” program, but also are important signs of the quality of family planning service. Given China’s low average education levels, especially among rural women, the tasks of family planning programs to enhance women’s contraceptive knowledge should include their awareness of their own reproductive rights. The results of the analysis presented in this chapter suggest that quality service and an “information-based choice” program should be introduced into the rural areas. For the long term, China should try to narrow the gap between the official goals to limit the expansion of the population and the rural families’ desire for more children. Improving universal and compulsory education, prohibiting child labor, improving women’s social and financial status, endeavoring to eliminate the traditional preference for sons, and providing an insurance system for the elderly, especially in the rural areas, will all contribute to achieving this aim. China passed a new population and family planning law that went into effect on September 1, 2002. The new law continues the former policy of birth control and family planning, but it also states the intention to improve social security arrangements so that basic old-age insurance, basic medical insurance, and other benefits can be provided to those who practice family planning. This law is an important first step.

References AVSC International. 1999. “Informed Choice and Sterilization: Special Challenges.” p. 3 in Insights Into Reproductive Health. New York: AVSC International. Banister, J. 1987. China’s Changing Population. Stanford, CA: Stanford University Press. Chen, P. 1985. “Birth Control Methods and Organization in China.” pp. 135–148 in E.Croll, D.Davin, and P.Kane (eds), China’s One-Child Family Policy. New York: St. Martin’s Press. China Population Information and Research Center. Available http://www.cpirc.org.cn/ Cooney, R.S. and J.Li. 2001. “Sterilization and Financial Penalties Imposed on Registered Peasant Couples, Hebei Province, China.” Studies in Family Planning, 32(1):67–78. Davin, D. 1985. “The Single-Child Family Policy in the Countryside.” pp. 1–36 in E.Croll, D.Davin, and P.Kane (eds), China’s One-Child Family Policy. New York: St. Martin’s Press. Feeney, G. and F.Wang. 1993. “Parity Progression and Birth Interval in China: The Influence of Policy in Hastening Fertility Decline.” Population and Development Review, 19(1):61–101. Freedman, R., Z.Xiao, B.Li, and W.Lavely. 1988. “Local Area Variations in Reproductive Behavior in the People’s Republic of China, 1973–1982.” Population Studies, 42(1): 39–57. Hardee-Cleaveland, K. and J.Banister. 1988. “Fertility Policy and Implementation in China, 1986– 88.” Population and Development Review, 14(2):245–286. Hartmann, B. 1987. Reproductive Rights and Wrongs: The Global Politics of Population Control and Contraceptive Choice. New York: Harper & Row Publishers. Kaufman, J., Z.Zhang, X.Qiao, and Y.Zhang. 1989. “Family Planning Policy and Practice in China: A Study of Four Rural Counties.” Population and Development Review, 15(4): 707–729. Kaufman, J., Z.Zhang, X.Qiao, and Y.Zhang. 1992. “The Quality of Family Planning Services in Rural China.” Studies in Family Planning, 23(2):73–84, Nortman, D.L. 1980. “Sterilization and the Birth Rate.” Studies in Family Planning, 11(9/10): 286– 300. Poston, D.L. Jr 1986. “Patterns of Contraceptive Use in China.” Studies in Family Planning, 17(5) 217–227. Ross, J.A. 1992. “Sterilization: Past, Present, Future.” Studies in Family Planning, 23(3): 187–198.

Patterns of sterilization

47

Short, S.E., L.Tao, and W.Yu. 2000. “Birth Planning and Sterilization in China.” Population Studies, 54:279–291. State Family Planning Commission of China. 1988. “Yearly Report of Family Planning Program.” Beijing, China: Chinese Population Press. State Family Planning Commission of China. 2002. “Family Planning Reform Focuses on Informed Choice and Clients’ Needs.” http://www.sfpc.gov.cn/en/enews20020306–1.htm Tien, H.Y. 1980. “Wan, Xi, Shao: The China Solution,” International Family Planning Perspectives, 6(2):65–73. Tu, P. 1995. “IUD Discontinuation Patterns and Correlates in Four Counties in North China.” Studies in Family Planning, 26(3):169–179. Wolf, A.P. 1986. “The Preeminent Role of Government Intervention in China’s Family Revolution.” Population and Development Review, 12(1):101–116. Yang, Q. 1994. “Provincial Patterns of Contraceptive Use in China.” Asia-Pacific Population Journal, 9(4):23–42. Zeng, Y. 1996. “Is Fertility in China in 1991–1992 Far Below Replacement Level?” Population Studies, 50(1):27–34. Zhong, Y Z., L.F.Ding, and E.S Gao. 2000. “Analysis of Contraceptive Knowledge Among Married Reproductive Women in China.” Reproduction and Contraception, 11(1/2):79–85. Zhou, F. 1991. “Zhong Guo Yi Hun Yu Lin Fu Nv De Bi Yun Xian Zhuang.” (The Current Contraceptive Information of Chinese Married Women) pp. 1–15 in Essays on China’s Sample Survey of Fertility and Contraceptives. Beijing, China: Chinese Population Press.

Part II Family and marriage patterns

4 The impact of family structure on fertility Feinian Chen

Introduction The influence of family structure on fertility behavior is a classic question in social demography. Theoretically, there are expectations that extended family structure provides support for early childbearing and consequently a high level of fertility (Davis 1955; Davis and Blake 1956; Goode 1970; Ryder 1983). Shared economic costs of children and child care, as well as an emphasis on family continuity and encouragement of early marriage, and early childbearing from parents, are often cited as some of the main reasons why extended family systems may induce high fertility (Davis 1955). To test empirically for the linkage between family structure and fertility behavior, a diverse range of dependent variables have been used by researchers. Among them are various fertility measures, including actual fertility, desired family size, and birth spacing, as well as other processes of family formation, such as age at marriage, and age at first birth (Morgan and Rindfuss 1984; Chamrathrithirong et al. 1988; Chi and Hsin 1996). In contrast, measurement of the key independent variable, that is, family structure, is much simpler. In the above mentioned studies, whether or not married children live in the same household as their parents is often used as a standard indicator of extended family living arrangements, with some studies further distinguishing parents from the husband’s or the wife’s side, and others focusing on the husband’s parents only, depending on specific cultural contexts. Nonetheless, household structure and family context do not always overlap. In studies of intergenerational exchange, it has been well recognized that family ties can extend well beyond the boundary of the household (Martin 1989; Tu et al. 1992; Bian et al. 1998; Ofstedal and Chayovan 1999). Non-coresident children and parents who live nearby often maintain a high level of contact and exchange with each other (Lee et al. 1994; Knodel and Chayovan 1997). While they may live in separate households, the shared responsibilities and activities allow them to function like one family. Hence, sociologists have created the term “modified extended family” or “network family” to characterize this situation (Greenhalgh 1984; Rossi and Rossi 1990; Unger 1993). In this chapter, the concept of a “modified extended family” is used in an analysis of the transition to first birth in China. Instead of focusing only on coresidence, the boundary of the household is extended by also incorporating parents who live close by, a residential type that is sometimes referred to as “quasi-coresidence” (Ofstedal and Chayovan 1999). As mentioned earlier, one of the key mechanisms of how family structure affects fertility is via the reduced opportunity costs of children. In contemporary

Fertility, family planning, and population policy in china

50

China, while grandparents do not necessarily bear the actual economic costs of raising children, they act as important alternative childcare givers. Grandmothers are often identified as the most important caregivers other than parents themselves (Short et al. 2002). And, care-giving is not just limited to coresiding grandparents. Grandparents who live nearby are often just as likely to help (Unger 1993). Chen and her colleagues (2000) have shown that in China, having grandparents living in the same household, or next door, substantially reduced the mother’s childcare hours. Thus, with parents in the household or living nearby as future childcare helpers, couples may choose to start having children earlier than if no such choice was available. In addition, the level of interaction between parents and children who live together or next door to each other is extremely high. Thus, young couples could be under stronger normative pressures from the parents to start the childbearing process earlier for the purpose of carrying on the family line. In this chapter, data are used from the China Health and Nutrition Surveys (CHNS) that were conducted in 1989, 1991, 1993, to model the transition to first birth in China. Due to data limitation, researchers often have to focus on household structure, despite strong substantive callings for the family context. Fortunately, the CHNS data provide detailed information on the proximity of parents and parents-in-law of married couples, which makes it possible to construct measures of coresidence and quasi-coresidence respectively, and to further separate matrilocal from patrilocal types of residence. Further, the longitudinal nature of the data allows one to use time-varying independent variables and to establish the right time order for the relationship between parental residence and fertility.

Background China is an interesting setting for studying the relationship between family structure and fertility. Historically, the extended family was the dominant family form in China. Despite falling family size and increasing proportions of nuclear families, the tradition of patrilineal and patrilocal extended families remains important in contemporary China (Zeng 1986). According to the 1990 Census of China, around 27 percent of the households in China contained 3 or more generations, and about 73 percent of the population aged 65 and above live in households made up of 2, or 3, or more generations (Guo 2000). In a study of 2 of the largest cities in China, Logan and his colleagues (1998) found that 67 percent of the parents lived with at least 1 adult child. Moreover, the significance of the ties between parents and children extends beyond extended family living arrangements. Traditionally, the Confucius ideology places a strong emphasis on filial responsibilities and intergenerational connections. It is no exaggeration to say that the intergenerational relationship, that is, the link between parents and children, often took precedence over the conjugal relationship between the husband and wife (Thornton and Lin 1994). While industrialization and urbanization inevitably shift the intergenerational flow of exchange from one that is parent-centered to one that is children-centered, the bond between parents and adult children remains strong in China. Rapid declines in fertility have accelerated the nuclearization process in China; however, living in separate households does not necessarily lead to weakened

The impact of family structure on fertility

51

relationships between parents and children. Indeed, it is not uncommon for parents and married children to be very much involved in each others’ lives, even if they do not live in the same household. The interaction between them is at a level that is much higher than the norm in most Western countries (Unger 1993). For example, Bian and his associates (1998) showed that non-coresident sons and daughters living in urban China maintained high face-to-face contacts with parents and provided regular help to them. In addition, the relationships between them were unaffected by whether the parents coresided with another child. Indeed, living nearby has been identified as a possible substitute for living together (Unger 1993). In rural areas, improved economic conditions often prompt the son to move out of the parents’ household and to build a house right nearby (Chen 2003). While they do not live in the same household, they may not function as separate families. They may still share meals and coordinate their work activities together. Due to the fluidity of the boundary of the household, living in two separate houses does not mean complete separation in life. The strength of the intergenerational ties between parents and married children is also expressed in terms of childcare provided by grandparents. In the United States, where the norm of noninterference characterizes intergenerational exchanges, it is common for grandparents to provide occasional babysitting, day care, or short-term coparenting (Cherlin and Furstenberg 1992). In contrast, in China, it is normative for grandparents, particularly grandmothers, to play a central role in nurturing the grandchildren. In addition, caregiving involves not just the coresident grandparents but also those who live nearby (Chen et al. 2000). Besides the cultural support for caregiving by grandparents, the intense conflict between maternal work and childcare also leads to the prevalence of grandparent caregiving. Women’s labor force participation rates are extremely high, as a consequence of the Communist government’s active promotion programs (Zhu and Guang 1991). Further, work arrangements are often not flexible, especially with regard to wage jobs. Movement in and out of a wage job is difficult, as wage jobs are in high demand. In rural areas, agricultural fieldwork, animal husbandry, or gardening are potentially more compatible with childcare. However, women often combine different work activities and carry a heavy load of work, thus making less room for childcare (Entwisle and Chen 2002). A lack of daycare facilities poses additional constraints (Tobin et al. 1989; Jacka 1997). As a result, help from grandparents is essential for women to balance work and childcare. The extent of the help is undoubtedly affected by the proximity of grandparents (Chen et al. 2000). Given the strength of the intergenerational relationship and the need for potential “mother substitutes,” it is hypothesized that early fertility decisions may depend on the availability of extended family members. Parents living in the household or nearby may exert an upward pressure on reproduction because of their anticipated childcare involvement, as well as their traditional family attitudes. In this chapter, the effect of both coresidence and quasi-coresidence on the transition to first birth is examined. It is widely known that China’s birth planning program places severe constraints on the number of births, affecting reproductive goals as well as behavior (Banister 1987). As a result, the total fertility rate (TFR) dramatically declined from around 5.3 in the 1950s to around 2.0 in the 1990s (Tu 2000) (see Poston and Glover’s Chapter 12 in this volume for more discussion). Nonetheless, while there is less variation in completed fertility and even though having a first birth is almost universal, the timing of the first birth still varies

Fertility, family planning, and population policy in china

52

considerably from one woman to the other (Tu 1991). This analysis focuses on the influence of the extended family and “modified extended family” structure on the risk of a first birth. It further distinguishes the effect of matrilocal from patrilocal types of residence because of the patriarchal and patrilineal tradition in China.

The China Health and Nutrition Survey This analysis draws on data from the CHNS, a longitudinal survey that was conducted collaboratively by the Carolina Population Center and the China Academy of Preventive Medicine. The survey was initiated in 1989 and collected economic, demographic, nutrition, and health information from 3,800 households in 188 urban and rural communities. It was undertaken in eight provinces of China, namely, Guangxi, Guizhou, Henan, Hubei, Hunan, Jiangsu, Liaoning, and Shandong. Although the provinces were not selected according to a probability design, they are geographically diverse and vary substantially in terms of demographic and economic indicators.1 Two of the provinces are coastal; four are located in central China; and two are mountainous southern provinces. Together the sample covers a third of China’s population. The CHNS is composed of a household survey, a nutrition survey, a community survey, an ever-married women survey, and a series of physical examinations. The evermarried women survey in 1993 serves as the base for the research reported in this chapter. Detailed questions on the marriage and birth history of women under the age of 52 are used to select a sample of married women at the risk of a first birth between 1989 and 1993 and to construct the dependent variable, that is, the transition to first birth. Selected characteristics of women, households, and communities in the sample come from the household and community surveys of 1989, 1991, and 1993. The sample is limited to married, fecund, Han women, who had not given birth prior to the interview date in 1989. Women who are at risk of a first birth prior to the start date are retained using a conditional likelihood approach (Guo 1993). Hazard models are used, and the observation window is specified as starting from the survey date in 1989 and ending at the survey date in 1993. Exposure begins nine months after the date of marriage and ends on the birth of the first child. Women are censored at the date of the 1993 survey if no birth has occurred. The final sample includes 218 women.

Transition to first birth Figure 4.1 shows the cumulative distribution of first birth by duration months since the date of marriage. Theoretically speaking, the exposure of a first birth should start nine months after the date of marriage. As shown in Figure 4.1, a sizable portion of the women in the sample gave birth before nine months after the date of marriage. This is likely due to premarital conception, or perhaps reporting mistakes in either the date of birth or date of marriage, or premature births. For those women who had a birth during the observation window, most of them occurred within 45 months. About 60 percent of the women in the sample had a first birth. This number may seem low, as it is known that first births are almost universal in China and occur relatively soon after marriage. The

The impact of family structure on fertility

53

relatively low number of births reported here is likely due to the way the sample is constructed. As the CHNS is a longitudinal survey that follows households over time, at each wave, new women can enter the sample by marrying into the family. Indeed, about one third of the women in this sample entered the survey after 1991. Hence their exposure to having a first birth has not been long, and a lot of them are censored by 1993. This is not a trivial issue because it is not possible to observe the full window of exposure. Cox proportional hazard models are used to address the issue of censoring (see further discussion in Wu and Walther’s Chapter (2) in this volume). Figure 4.1 clearly shows that there is significant variation in the timing of the first birth for individual women. What could have contributed to such individual differences? With parents or parents-in-law living in the same household or close by, will they exert an upward pressure on early childbearing? This question may be answered in the multivariate analysis.

Figure 4.1 Cumulative distribution of first birth by duration months since date of marriage, CHNS. Residential patterns of parents and parents-in-law Before moving to the multivariate analysis, the measurement of the residential patterns of the parents, which is central to the analysis, will be presented. Questions regarding the health status, care, and proximity of the parents and parents-in-law of the woman were asked in 1989, 1991, and 1993.2 As a result, one can determine whether her mother, father, mother-in-law, and father-in-law were living in the household, as neighbors, in the same village or neighborhood, in the same county or city, or elsewhere. Because of the very low divorce rate in China, the residence of one parent implies that of the other

Fertility, family planning, and population policy in china

54

(when both are living). Thus in the analysis, residence of the set of parents is considered, rather than those of mothers and fathers separately. Figure 4.2 shows the residential patterns of parents and parents-in-law of married women in 1993. The residential patterns of parents and parents-in-law in 1989 and 1991 are similar. Although the living arrangements of parents and married children may certainly change over the life course (Chen 2001), the sample used here consists of recently married women, who have relatively stable residences within a four-year window. Figure 4.2 shows that there is a clear patrilocal pattern in the residence of married children and parents in 1993. This is reflected not only by statistics on coresidence, but also by those dealing with quasi-coresidence. About half of the married women in the sample live with their husbands’ parents, as compared to

Figure 4.2 Residential patterns of parents-in-law and parents of married women, CHNS, 1993. 8 percent living with the wives’ parents. An additional one-fifth of the couples have the husbands’ parents as neighbors. In contrast, 3 percent of the couples live next door to their wife’s parents. Combining the statistics on coresidence and quasi-coresidence, 71 percent of the women in the sample have a patrilocal type of residence, suggesting a higher level of patrilocality than would occur using the coresidence measure only. On the

The impact of family structure on fertility

55

other hand, about 11 percent of the women live with the wives’ parents, which is a significant departure from the traditional patrilineal culture. This may have implications for fertility decisions, as the maternal grandmothers could also act as “mother substitutes.”

Results As already noted, Cox proportional hazard models are used to estimate the risk of a first birth. The major advantages of Cox regression models include their ability to handle censored data and to accommodate time-varying covariates (Allison 1984). (See Wu and Walther’s Chapter (2) in this volume for a more detailed description of Cox regression.) Since the data were collected using a multistage cluster design, conventional estimates of standard errors may not be accurate, since they assume independence of observations. Therefore, the Huber/White/ sandwich estimator of robust standard errors is used to adjust for clustering in the data (Statacorp 2001). The sample of 218 women is converted to a sample of 12,432 women-months. For example, if a woman gives birth 11 months after the date of marriage, she contributes 11 records to the sample. The dependent variable in the model has two components: (1) whether a birth has occurred from 1989 to 1993, and (2) the time elapsed between the date of marriage and first birth, or the duration from the time of marriage to the interview date of 1993 if no birth has occurred; these are censored observations. The duration time ranges from 0 to 379 months. Less than 15 percent of the sample has a duration time above 10 years (120 months). To make sure that these extreme cases do not bias the sample, sensitivity tests were performed by deleting cases with duration times longer than 120 months. Since the results were essentially the same, they are retained in the sample. A range of covariates is used in the Cox regression models. Of central interest are measures of family structure. Using the residential patterns described in Figure 4.2, family structure is operationalized in four different ways: patrilocal extended family (coresidence with the husband’s parents), patrilocal “modified extended family” (coresidence and quasi-coresidence with the husband’s parents), matrilocal extended family (coresidence with the wife’s parents), and matrilocal “modified extended family” (coresidence and quasi-coresidence with the wife’s parents). These four measures are used in four separate models. The variables are used as time varying and are lagged in the model. For example, family structure in 1989 and 1991 is used to predict the hazard of a birth in a given month in 1991 and 1993 respectively. As Morgan and Rindfuss (1984) have pointed out, using current residence to predict past fertility could be methodologically and conceptually flawed. Moreover, the causal relationship between parental residence and fertility could be two-way. While parents living in the same household are likely to lead to early childbearing, it is also possible for married children to move to their parents’ house after childbirth. Although using lagged terms does not settle the issue of causality completely, it is at least helpful to establish the correct time order. For those women who joined the survey after 1991, they obviously were not included in the 1991 ever-married women survey. In those cases, the 1993 residence measures are used, assuming that there was no change in between the two times. To further test for the validity of the assumption, the residential patterns of parents and

Fertility, family planning, and population policy in china

56

parents-in-law were cross-tabulated in 1989, 1991, and 1993; little change during the four-year interval was found (results not shown). This is not surprising, since the sample is selective of young women who have recently married. Coresidence and quasicoresidence rates for this group of women are the highest, and then gradually decline over the life course (Chen 2001). Various demographic and socioeconomic variables are also used in the models. Two are time varying and lagged: logged household income and logged percentage of agricultural labor force in the community in 1989 and 1991. Also included are age, education (measured by completing some middle school or higher), and residence type (urban or rural). Descriptive statistics are presented in Table 4.1. Table 4.2 shows the results from the Cox proportional hazard models estimating the hazard of a first birth. The estimates are presented in the form of hazard ratios. Four models are included. The first two models look at patrilocal residence, and the last two examine matrilocal residence. The results clearly suggest that family structure has a significant impact on the risk of a first birth. As shown in Model 1, for women living with the husband’s parents, the risk of a first birth at a given month is almost twice that of those who do not live with them. This is consistent with the theoretical rationale that extended family structure is conducive to early childbearing. Moving to Model 2, the estimate suggests that the positive effect exists not only for coresidence with the husband’s parents, but also for those whose husbands’ parents live as neighbors. Expanding the definition of extended family, by including quasi-coresidence, the effect of family structure becomes stronger. With the husband’s parents either living in the household or next door, the woman is 2.23 times more likely to have a first child, than those who do not. The hazard ratio is 26 percent higher than that in Model 1. This clearly highlights the importance of the family context. Living in the same household with parents is not the only way for young couples to be influenced by their parents. With parents living nearby, and with a high level of interaction, the lives of the two generations can be as intertwined as those who live under the same roof.

Table 4.1 Descriptive statistics of selected sample characteristics (N=218) Variable Age at 1989

Mean Standard deviation 24.495

5.960

Some middle school education or more

0.449

0.499

Logged household income in 1989 (deflated)

8.174

0.979

Logged household income in 1991 (deflated)

8.434

0.808

Urban (1=urban 0=rural)

0.339

0.474

Logged percent of agricultural labor force in the community in 1989

2.929

1.841

Logged percent of agricultural labor force in the community in 1991

2.842

1.878

The impact of family structure on fertility

57

Table 4.2 Woman month risk of first birth, 1989– 1993, CHNS (N=12432) Hazard ratio Model 1 Model 2 Model Model 4 3 Age

0.867*** 0.858*** (0.023) (0.024)

0.840 0.842*** (0.023) (0.022)

Logged household income

1.051 (0.080)

1.060 (0.086)

1.081 (0.090)

1.061 (0.087)

Some middle school or more

0.729 (0.164)

0.742 (0.166)

0.790 (0.175)

0.789 (0.176)

Urban

1.501 (0.563)

1.457 (0.568)

1.512 (0.638)

1.486 (0.616)

Logged % of agricultural labor force

1.025 (0.095)

0.989 (0.094)

1.003 (0.102)

0.998 (0.099)

1.975*** (0.414)

—

—

—

Modified patrilocal extended family (coresidence+quasi-coresidence)

— 2.230*** (0.524)

—

—

Matrilocal extended family (coresidence)

—

—

0.644 (0.174)

—

Modified matrilocal extended family (coresidence+quasi-coresidence)

—

—

—

0.746 (0.170)

48.13

52.23

44.26

45.4

6

6

6

6

Patrilocal extended family (coresidence)

Wald χ2 Degree of freedom Notes 1 ***p<=0.001. 2 Standard errors in parentheses.

Having parents available as future childcare givers has the potential of reducing the role incompatibility between work and childcare for women, and may thus lead to an earlier transition to first birth. This is one of the mechanisms through which the extended family, or the “modified extended family,” operates to influence fertility decisions, but definitely not the only one. Models 3 and 4 show that the matrilocal extended family, or matrilocal modified extended family, has virtually no effect on the transition to first birth. With the wife’s parents living in the household or as neighbors, the risk of a first birth is not significantly affected. This is not just a matter of power in statistics, since relatively fewer couples live with, or next door to the wives’ parents. The hazard ratios are indeed below one, indicating a negative effect that is completely opposite to the direction of the effect of patrilocal residence. A previous study showed that maternal grandparents living in the same household were as likely to help with childcare as paternal grandparents

Fertility, family planning, and population policy in china

58

(Chen et al. 2000). However, the availability of potential childcare providers might be one of the factors that influences fertility decisions, but definitely is not the only one. It is interesting to see that the matrilocal extended family does not appear to motivate early childbearing in the same way the patrilocal extended family does, which undoubtedly demonstrates the implications of patrilineal and patrilocal tradition in China. Most of the control variables in the model are not significant except for age. Younger women have a higher risk of a first birth, as expected. While age and residential patterns of parents are correlated, controlling for age does not take away the effect of extended family living arrangements. The lack of statistical significance of the socioeconomic variables does not mean that they are irrelevant factors for fertility decisions. Previous aggregate studies consistently document the impact of economic development on fertility (Poston and Gu 1987; Poston 2000). However, since the focus of the chapter is on the transition to first birth in a relatively short period of time, instead of completed fertility, it is not surprising not to find any significant effect for these control variables.

Summary and conclusion The findings from this chapter are straightforward to summarize. Using data from the CHNS, the risk of a first birth is modeled for women during a four-year observation window from 1989 to 1993. Results show that extended family living arrangements have a clear impact on the transition to first birth. Consistent with theoretical expectations and previous research findings, coresidence with the husband’s parents substantially increases a woman’s risk of first birth. Further, the implication of extended family systems span beyond the household. Using an expanded definition of extended family by combining coresidence and quasicoresidence with parents, a very similar effect on the transition to first birth compared to that of coresidence is found. The findings clearly demonstrate the importance and relevance of the family context. Family ties are not bounded by the household Grandparents who live close by can easily provide childcare, in the same way as those who live together with their grandchildren. Parents and adult children who live next door to each other maintain daily interaction and do not function as separate families. As a result, children are easily influenced by the traditional family attitudes of parents. It is even possible that fertility decisions may not entirely lie in the hands of the married children, but may also involve their parents. The findings in this chapter also demonstrate the legacy of the patriarchal, patrilocal, and patrilineal traditions in China. Living with, or next door to, the wife’s parents apparently does not affect the process of family formation for young married couples in the same way patrilocal residence does. While it is often noted that extended family living arrangements mean shared economic costs of children and better resources for childcare, the lack of an effect for matrilocal types of residence suggests that ideological factors cannot be ignored. In a traditionally patrilineal culture, the pressure of continuing the family line lies with the son, not with the daughter. Thus, living in the vicinity of the husband’s parents may well exert an upward pressure on early childbearing. China is currently undergoing a dramatic socioeconomic restructuring. It is accompanied by transitions in norms and ideology, a process that is characterized by persistence and change at the same time. While coresidence with parents is declining, as a

The impact of family structure on fertility

59

result of fertility decline, changes in preferences, and improved economic conditions, married children and parents continue to prefer to live in close proximity to each other. There is yet no clear sign of weakened intergenerational ties. Grandparents help with childcare and housework, while children provide old-age support to them, reflecting a model of social exchange that is not only driven by the needs of each generation but also ingrained in Chinese culture. To better understand and monitor such patterns of exchange, researchers need to look beyond the household and to refocus the research lens on the family context.

Notes 1 Detailed description of the China Health and Nutrition Survey is available at the website: www.cpc.unc.edu/projects/china 2 This question was asked in the 1991 and 1993 ever-married women’s survey, but was included in the household survey in 1989. In addition, the 1989 question did not have “living as neighbor” as a choice. By using a question on the distance to parents/ parents-in-law’s house and another on the relationship to the household head, it was possible to construct measures of parental residence in 1989 and further cross-validate them by using the 1991 and 1993 measures.

References Allison, P.D. 1984. Event History Analysis. Beverly Hills, CA: Sage. Banister, J. 1987. China’s Changing Population. Stanford, CA: Stanford University Press. Bian, F., J.R.Logan, and Y.Bian. 1998. “Intergenerational Relations in Urban China: Proximity, Contact and Help to Parents.” Demography, 35:115–124. Chamrathrithirong, A., S.P.Morgan, and R.R.Rindfuss. 1988. “Living Arrangements and Family Formation.” Social Forces, 66:926–950. Chen, F. 2001. “Residential Patterns of Parents and their Married Children in Contemporary China: A Life Course Approach.” Paper presented at the Annual Meeting of Population Association of America at Washington DC. Chen, F. 2003. “The Impact of Economic Reforms on Family Division in China.” Paper presented at the Annual Meeting of Population Association of America at Minneapolis. Chen, F., S.E.Short, and B.Entwisle. 2000. “The Impact of Grandparental Proximity on Maternal Childcare in China.” Population Research and Policy Review, 19:571–590. Cherlin, A.J. and F.Furstenburg, Jr. 1992. The New American Grandparent: A Place in the Family, a Life Apart. Cambridge, MA: Harvard University Press. Chi, P.S.K. and P.Hsin. 1996. “Family Structure and Fertility Behavior in Taiwan.” Population Research and Policy Review, 15:327–339. Davis, K. 1955. “Institutional Patterns Favoring High Fertility in Underdeveloped Areas.” Eugenics Quarterly, 2:33–39. Davis, K. and J.Blake. 1956. “Social Structure and Fertility: An Analytic Framework” Economic Development and Cultural Change, 4:211–235. Entwisle, B. and F.Chen. 2002. “Work Patterns Following a Birth in Urban and Rural China: A Longitudinal Study.” European Journal of Population, 18:99–119. Goode, W.J. 1970. World Revolution and Family Patterns, 2nd Edn. New York: The Free Press. Greenhalgh, S. 1984. “Networks and Their Nodes: Urban Society on Taiwan.” The China Quarterly, 99:529–552.

Fertility, family planning, and population policy in china

60

Guo, G. 1993. “Event-history Analysis for Left-truncated Data.” pp. 217–243 in Sociological Methodology, San Francisco, CA: Blackwell Publishers. Guo, Z. 2000. “Family Patterns.” pp. 98–111 in Peng Xizhe and Zhigang Guo (eds), The Changing Population of China. Oxford: Blackwell Publishers. Jacka, T. 1997. Women’s Work in Rural China: Change and Continuity in an Era of Reform. Cambridge, MA: Cambridge University Press. Knodel, J. and N.Chayovan. 1997. “Family Support and Living Arrangements of Thai Elderly” Asia-Pacific Population Journal, 12:51–68. Lee, Y., W.L.Parish, and R.J.Willis. 1994. “Sons, Daughters, and Intergenerational Support in Taiwan.” American Journal of Sociology, 99:1010–1041. Logan, J.R., F.Bian, and Y.Bian. 1998. “Tradition and Change in the Urban Chinese Family: The Case of Living Arrangements.” Social Forces, 7:851–882. Martin, L.G. 1989. “Living Arrangements of the Elderly in Fiji, Korea, Malaysia, and the Philippines.” Demography, 26:627–643. Morgan, S.P. and R.R.Rindfuss. 1984. “Household Structure and the Tempo of Family Formation in Comparative Perspective.” Population Studies, 38:129–139. Ofstedal, M. and N.Chayovan. 1999. “Intergenerational Support and Gender: A Comparison of Four Asian Countries.” Paper presented at the Annual Meeting of Population Association of America, New York City, New York. Poston, D.L. Jr 2000. “Social and Economic Development and the Fertility Transitions in Mainland China and Taiwan.” Population and Development Review, 26(supplement): 40–60. Poston, D.L., Jr and B.Gu. 1987. “Socioeconomic Development, Family Planning, and Fertility in China.” Demography, 24:531–551. Rossi, A.S. and P.H.Rossi. 1990. Of Human Bonding: Parent-Child Relations Across the Life Course. New York: Aldine de Gruyter. Ryder, N.B. 1983. “Fertility and Family Structure.” Population Bulletin of the United Nations, 15:15–34. Short, S., F.Chen, B.Entwisle, and Z.Fengying. 2002. “Maternal Work and Time with Preschool Children: A Multi-Method Approach.” Population and Development Review, 28:38–57. Statacorp. 2001. Stata Version 7.0. College Station, TX. Thorton, A. and H.Lin. 1994. Social Change and the Family in Taiwan. Chicago, IL: University of Chicago Press. Tobin, J.J., D.Y.H.Wu, and D.H.Davidson. 1989. Preschool in Three Cultures. New Haven, CT: Yale University Press. Tu, E.J., V A.Freedman, and D.Wolf. 1992. “Kinship and Family Support in Taiwan: A Microsimulation Approach.” Paper presented at the Population Association of America Annual Meeting, Denver, CO. Tu, P. 1991. “Birth Spacing Patterns and Correlates in Shaanxi, China.” Studies in Family Planning, 22:255–263. Tu, P. 2000. “Trends and Regional Differentials in Fertility Transition.” pp. 22–29 in Peng Xizhe and Zhigang Guo (eds), The Changing Population of China. Oxford: Blackwell Publishers. Unger, J. 1993. “Urban Families in the Eighties: An Analysis of Chinese Surveys.” pp. 25–49 in Deborah Davis and Steven Harrell (eds), Chinese Families in the Post Mao Era. Berkeley, CA: University of California Press. Zeng, Y. 1986. “Changes in Family Structure in China: A Simulation Study.” Population and Development Review, 12:675–703. Zhu, C. and X.Guang. 1991. “An Analysis of the Life-Cycle of Chinese Women.” Chinese Journal of Population Science, 3:247–257.

5 The impact of intermarriage on the fertility of minority women Chiung-Fang Chang Intermarriage is believed to be the final stage of assimilation (Gordon 1964). Many western studies have found a significant contribution of intermarriage on the fertility patterns of minority groups (Beaujot et al. 1982; Axelrod 1990; Coleman 1994; Hout and Goldstein 1994). When minorities gradually overcome cultural and structural barriers and intermarry, they tend to broaden their perspectives and cultural patterns, and become less traditional. Their fertility thus becomes lower and more similar to that of the majority. Increasingly, scientists within China and abroad have begun to devote more and more attention to the demography of China’s minority populations. However, only one publication study (Zhang 2001) has discussed the relationship between interethnic marriage and population dynamics among the majority Han and the other ethnic nationalities. In sum, with the exception of the research reported here, there have been no investigations of the relationship between intermarriage and fertility among the different ethnic nationalities of China. One reason for this neglect has been the difficulty in obtaining systematic and detailed information about the patterns of intermarriage. Most of the official publications from censuses and surveys provide only crude data. More detailed information is generally not available. In this chapter, micro-data from the 1990 Census of China are used to study differences in fertility behavior among ten different ethnic groups in China. In particular, these data are used to assemble intermarriage data by matching the ethnicity of each minority woman with her husband in each household unit. Two major issues will be discussed: (1) Instead of studying ethnic fertility at the national level, this chapter attempts to provide a more informative analysis by examining fertility patterns among different ethnic groups; (2) Since the one-child family planning policy is relatively less restrictive for minority nationalities in China, learning about the variation in fertility among different ethnic groups should be meaningful. This chapter will focus on the effects of intermarriage as a major independent variable of fertility. The fertility patterns of 10 minority nationalities are examined, namely, those 9 with the largest populations in 1990, Zhuang (15.5 million), Manchu (9.8 million), Hui (8.6 million), Miao (7.4 million), Uygur (7.2 million), Yi (6.6 million), Tujia (5.7 million), Mongolian (4.8 million), and Tibetans (4.6 million); the tenth group is the Koreans, which is the thirteenth largest group with a population size of 1.9 million, but the most socioeconomically advanced group compared with all other minorities. In this chapter a brief review of ethnic relations in China is presented along with a discussion of population policy in relation to the fertility behaviors of China’s minority women. Next the data, methods, and results of the descriptive and multivariate analysis are presented. The chapter is concluded with a discussion of some of the implications of the results.

Fertility, family planning, and population policy in china

62

China’s ethnic minorities In China, there are 56 nationalities, the majority Han and 55 other officially identified minority nationalities. The Han people comprise China’s and the world’s largest ethnic group. Its formation and development were a continuous process of integration and assimilation of the earlier tribal groups over the centuries. It was in the Han Dynasty (206 BC to AD 220) that they adopted the name “Han” (Heberer 1989; Ma 1992). In 1990 the population of the 55 minority groups numbered 91.2 million persons, and represented over 8 percent of the total population. There is great variation in the size of these minority nationalities. Among the 55 groups, 18 of them have over a million people. The 10 largest groups are the Zhuang with more than 15 million, followed by the Manchu, Hui, Miao, Uygur, Yi, Tujia. Mongolian, Tibetan, and Bouyei. Many of the remaining 45 groups are small; 7 of them have less than 10,000 persons, and about 700,000 minority people have still not been classified (State Statistical Bureau 1991). Although small in number, the people of the various minorities inhabit 50–60 percent of the country. Geographically, they are concentrated in many regions of the country, including Inner Mongolia, Xinjiang, Tibet, Guangxi, Ningxia, Heilongjiang, Jilin, Liaoning, Gansu, Qinghai, Sichuan, Yunnan, Guizhou. Guangdong, Hunan, Hebei, Hubei, and Fujian. More than 60 percent of them are located in the southern and southwestern areas of China, and about 30 percent live in the north, northeast, and northwest areas (State Statistical Bureau 1991). In general, 90 percent of the minority populations are located in border and mountainous provinces of China. Unlike racial groups, China’s minority populations are not distinguished from one another on the basis of phenotypic differences. Their identification depends mainly on the cultural and linguistic differences that developed over many centuries (Poston and Shu 1987; Heberer 1989; Poston et al. 2006). Among the ten major ethnic groups, the Zhuang and Tujia reside in southern China. The Zhuang have their own language, which belongs to the Chinese-Tibetan language family (Ma 1992). Over 90 percent of the Zhuang live in their own autonomous region, the Guangxi Zhuang Autonomous Region. Because of its large minority population, their family planning policies are pretty much the same as the Han (Zou 1991). Most Tujias dwell in the Xiangxi Tujia-Miao Autonomous Prefecture, the Exi Tujia-Miao Autonomous Prefecture, and some counties in Hunan and Hebei provinces. The Tujia language is very similar to that spoken by the Yis and belongs to the Chinese-Tibetan language system, but the majority of Tujias have come to speak the Han and Miao languages. Both the Zhuangs and Tujias have been largely assimilated, their clothing and customs are very much like those of the Han (Ma 1992). The Miao and the Yi reside in southwestern China. They are mostly concentrated in Guizhou, Yunnan, Sichuan, Hunan, and Guangxi provinces. The Miao language partly belongs to the Chinese-Tibetan system; however, many Miao have adopted Chinese, Yao, or Dong languages as their spoken language (Poston et al. 2006). The Yi, in general, do not have a complete written language (Ma 1992). In general, both groups are mostly mountainous residents and are much less developed socioeconomically than the Han. Both the Hui and the Uygur are largely Moslem. The Hui employ an Arabic script. Their ancestors were actually the same as those of the Uygur. After a course of movement and development, today’s Uygur language belongs to the Turkic group of the

The impact of intermarriage on the fertility of minority women

63

Altaic language family (Ma 1992). Over 99 percent of the Uygur population resides in the Xinjiang Uygur Autonomous Region in far north-western China, while the Hui people are more dispersed throughout China. Generally, exogamous marriage is not encouraged among Moslems. Moslem women, in particular, are specifically prohibited from marrying non-Moslems. The Manchu and the Korean are the two most advanced of China’s minority groups, and they reside in northeastern China. The Manchu are the second largest minority group. Almost 75 percent of them live in Manchuria, that is, Liaoning, Jilin, and Heilongjiang provinces, located in northeastern China. The Manchu have their own language, which belongs to the Altaic linguistic family (Ma 1992). Similar to the Manchu, almost 95 percent of the Koreans live in Manchuria, especially the Yanbian Korean Autonomous Prefecture in Jilin Province (China Handbook Editorial Committee 1985). Both the Manchu and Korean people are very advanced, and their socioeconomic status and demographic characteristics are very close to those of the Hans. And their fertility levels are much lower than the fertility of the other minority groups. The Mongolians live mostly in the Inner Mongolia Autonomous Region in northern China. Mongolians have their own spoken and written language, which belongs to the Altaic language family (Ma 1992). Contemporary Mongolians have changed their traditional nomadic life styles to more modern ones. Compared with the other minority groups in 1990, Mongolians have relatively low percentages of the population in agriculture (Poston et al. 2006). Most Tibetans live in the Tibet Autonomous Region in southwestern China, with others residing in Qinghai, Gansu, Sichuan, and Yunnan provinces. The Tibetan language belongs to the Tibetan-Myanmese language branch of the Chinese-Tibetan language family. Tibetans have their own written script, and three different spoken dialects, Weizang, Kang, and Amdo (China Handbook Editorial Committee 1985). The areas where Tibetans live in compact communities are mostly highlands and mountainous areas studded with snow-capped peaks. The living standards of the Tibetan people today are lower than those of most minority groups. Tibetans mainly live in rural areas, have low levels of education, and high levels of fertility. There have been assimilation issues between the majority Han people and the other minority groups in China for thousands of years. Throughout history, the vicissitudes of time, war, migration, and seizure of lands have produced many shifts of population in the border areas. Various ethnic minorities have lived in both mixed and separate compact communities. Some minority groups have been widely scattered over the country. A permanent presence of several dozen million ethnic minority people can be found in the country’s big and small cities and towns. Many have thus formed close ties with the Han people. Assimilation is likely unavoidable for most minority group members interacting with the mainstream culture. The language, customs, and cultural differentiation of various minority groups from the Han indirectly affect the differentiation of the minority populations, their socioeconomic status, as well as their fertility behavior. So, minority women who are in interethnic marriages indicate that they have more opportunities to interact with Han or other ethnic groups, usually have higher education, and tend to be more assimilated with the Han and others. Therefore, it is expected that they will have fewer children than the women who marry within their own ethnic groups.

Fertility, family planning, and population policy in china

64

China’s family planning policies for the minorities The one-child policy was first announced by the Chinese government in 1979. The purpose of the one-child policy was to efficiently reduce the country’s dramatic population growth. The plan called for each couple of dominant Han nationality to have only one child, although there were exceptions (see Liang and Lee’s Chapter 1 in this volume for more discussion). Most ethnic minorities were exempted from the policy (Banister 1987; Heberer 1989; Zhang 2001). The population policy for minority nationalities favors an increase in the rate of growth while favoring a reduction for the general population. Prior to the one-child policy, ethnic minorities were by and large exempt from all family planning policies. This population policy was carried out among minority regions to feed the manpower needs. From 1971 to 1980, there was discussion among government officials that the prior minority exemption had led to high birth rates in minority regions and faster population growth than among the Han (Heberer 1989). There were different arguments on how family planning policy should be applied in the minority regions. The prevailing point of view was that population growth should be controlled in dense minority regions, but the small minority nationalities should have more flexible birth control regulations. Detailed rules were proclaimed in April, 1984 that the one-child family policy should be encouraged for minorities comprising more than 10 million members (e.g. the Zhuang), and that all other nationalities could have 2 children per couple, and under special circumstances they could have 3, while 4 children or more would not be permitted (Zhang 2001:295). Since the one-child policy is implemented by each province and autonomous region, the regulations vary widely from one area to another. For example, urban Tibetans are allowed to have two births in Lhasa and several other cities, but no family planning policy has been enforced in the agricultural and pastoral areas (Zhang and Zhang 1991:494–497). In the Inner Mongolia Autonomous Region, there appear to be no restrictions on the number of children for the Daur, Ewenki, Elunchun, and Hezhe, except that the births must be three years apart from each other. In the same region, the Mongolians, Hui, Manchu, Koreans, and the descendants of Russians with Chinese citizenship may have two children. In the Xinjiang Uygur Autonomous Region it is stipulated that if both husband and wife are ethnic minorities, couples in urban areas may have 2 births, and couples in rural areas may have 3; under special cases, a fourth is allowed (Zou 1991:8). However, some minority nationalities in urban areas are subjected to much stricter regulations (e.g. Manchus in Heilongjiang and Liaoning provinces, Zhuangs in Guanxi province, and most minorities in Qinghai province); their family planning policies are the same as those of the Han. In general, there is no unified family planning regulation for intermarried couples. Most Han who intermarry ethnic minorities, except the Zhuang, are subjected to the same family planning regulations as minorities in most provinces and autonomous regions. Minorities who intermarry other ethnic minorities may choose to abide by either of the regulations that apply to their two groups (Zhang 2001:300–303). According to the national policy, intermarried couples may choose to maintain their original identities. Their children are allowed to choose their ethnic identity when they reach 18 years of age; before that, their ethnic identities are determined by their parents (Zhang 2001:97).

The impact of intermarriage on the fertility of minority women

65

Data and methods In this chapter, data are used from the One Percent Sample of the 1990 Census of China (State Statistical Bureau 1992). The 1990 Census contains self-identified data on the majority Han and 55 recognized ethnic groups. In this chapter, the fertility patterns of minority women from ten minority groups are examined. Available for analysis is a sample of 102,215 ever-married women aged 15–49 years of age who belong to one of these 10 minority groups. The dependent variable is Children Ever Born (CEB), which measures a woman’s number of children ever born over her childbearing years to date as of July 1, 1990. Among the more than 102 thousand women in the sample, CEB has an average value of 2.82 children with a standard deviation of 1.9. Poisson Regression Models are used to predict the effects of intermarriage on fertility. CEB is a “nonnegative integer-valued random variable” (Cameron and Trivedi 1998:1). Sometimes, count variables are treated statistically as if they are continuous and unbounded, and ordinary least squares (OLS) models are then used to estimate the effects of the independent variables on their occurrence. However, OLS models for count outcomes often produce biased, inefficient, and inconsistent estimates if the OLS regression assumptions are not met. In Figure 5.1 the observed CEB data for all 102,215 Chinese women are compared with a univariate Poisson distribution with a mean of 2.82. In the

Figure 5.1 Comparison of the observed distribution of CEB, with the univariate Poisson distribution with mean of 2.82.

Fertility, family planning, and population policy in china

66

figure, the solid line shows the frequency distribution of the CEB variable. The graph indicates that the distribution of the CEB variable closely resembles the univariate Poisson distribution for a mean of 2.82 (the dotted line). The effect of intermarriage on the number of children ever born of over 100,000 minority women will be examined using the following independent variables: Intermarriage. Intermarriage data were constructed by matching each minority woman’s ethnicity with that of her spouse. Two dummy variables were constructed to measure two types of exogamy: (1) out-marry to Han majority; and (2) out-marry to other minority group. The reference group is the endogamous variable, that is, minority women who marry in the same ethnic group. Ethnicity. To control for differences among the different ethnic groups, dummy variables were entered for the various groups. The Zhuang group is used as the reference because it is the largest of the ten, and it is subjected to a birth policy very similar to that of the Han. Policy index. This is an index that endeavors to measure the effect of the one-child policy. It represents the percentage of childbearing years (15–49) that the woman experienced before the year of 1979 (the year the one-child policy officially started). Years of schooling. This variable is the number of completed years of school. Professional/leadership job. This is a dummy variable measuring whether a minority woman is in a professional or leader job category (professional/leader job=1). Employment status. This is a dummy variable measuring whether a minority woman is employed (employed=1). Migration status. Each woman’s migration status is reflected in four dummy variables, namely, permanent migrants, long-term floaters, short-term floaters, and non-migrants. Permanent migrants are persons who migrate from other regions with legal household registration. Long-term floaters are those who migrated since or before 1985 but did not obtain a legal household registration permit in their destination. Short-term floaters are those who migrated after 1985 and did not have a legal household registration permit in their destination. Both long-term floaters and short-term floaters are temporary migrants. Nonmigrants is the reference category. The empirical analysis is undertaken in several ways. Descriptive data are first presented so to compare fertility differences between exogamous women and endogamous women for each group. Then, Poisson regressions are estimated to examine the effect of intermarriage on minority women’s fertility, controlling for the variables just presented. Next, additional Poisson regression analyses will be presented to compare how the intermarriage effect varied by residence.

Results Table 5.1 presents descriptive data for ever-married women, aged 15–49, for each of the 10 minority groups in 1990. Among the ten groups, the Koreans and the Manchu both have mean education scores of over 9 and over 7 years, respectively, and have mean fertility scores of less than 2 children. These 2 groups have the highest education and the lowest fertility. The Manchu have over 44 percent of married women out-marrying to majority Hans. Tibetans, on the other hand, are the group with low education, a low

The impact of intermarriage on the fertility of minority women

67

exogamous rate, and high fertility, an average of over 4 children per woman. Of the remaining 7 minority groups, the mean fertility scores ranged from 2.5 for Mongolians to 3.8 for Uygurs. Mean education

Table 5.1 Descriptive data for ever-married women aged 15–49, by ten ethnic groups, China, 1990 Ethnic group

Sample size

Mean of children ever born

Mean years of schooling

Proportion living in rural areas

Proportion outmarrying Han

Proportion out-marrying other ethnic groups

Mongolian 7,649

2.46

7.23

0.669

0.358

0.034

Hui

12,796

2.61

5.32

0.647

0.118

0.012

Tibetan

4,382

4.08

1.03

0.925

0.047

0.011

Uygur

10,105

3.84

5.86

0.781

0.002

0.005

Miao

9,600

3.05

3.06

0.939

0.146

0.108

Yi

7,877

3.33

2.55

0.973

0.148

0.020

Zhuang

21,266

3.02

5.66

0.949

0.150

0.016

Korean

3,146

1.86

9.28

0.680

0.080

0.008

Manchu

16,345

1.92

7.61

0.773

0.447

0.018

Tujia

9,049

2.53

4.96

0.950

0.247

0.079

ranged from 2.6 years for the Yi to 7.2 years for Mongolian. Exogamous rate to marriage to Han ranged from 0.2 percent for Uygurs to 36 percent for Mongolians. To farther examine the effect of intermarriage on ethnic fertility, mean differences in CEB between exogamous women and endogamous women are presented for each of the ten groups (Table 5.2). The means of endogamous women are subtracted from the means of exogamous women; all the values in the mean difference column are negative. The average CEB for exogamous women is lower than that for endogamous women for all ten groups. In particular, exogamous women in the Hui, Tibetan, and Uygur groups have average fertility of more than one child less than the respective endogamous women. But these differences do not control for other factors. Therefore, multivariate Poisson regressions are next estimated that examine the effect of intermarriage controlling for family planning policy and other socioeconomic factors. Table 5.3 presents the results of a series of Poisson regressions. Three models are shown in Table 5.3 examining how the effect of intermarriage changes after adding more control variables. Each model shows the results of Poisson regression coefficients and incidence rate ratios (IRR). CEB is first regressed on intermarriage controlling only for differences among the minority groups. The Poisson regression coefficients in the first column of the table represent the gross effects of intermarriage on fertility taking into account the women’s ethnicities. Both

Fertility, family planning, and population policy in china

68

Table 5.2 Mean numbers of CEB for exogamous women and endogamous women, by ten ethnic groups, China, 1990 Ethnic group

Mean number of CEB (N) Exogamous women

Endogamous women

Mean difference

Mongolian

2.00 (2,736)

2.72 (4,913)

−0.72

Hui

1.62 (1,510)

2.74 (11,286)

−1.12

Tibetan

2.65 (207)

4.15 (4,175)

−1.15

Uygur

2.00 (25)

3.85 (10,080)

−1.85

Miao

2.58 (1,405)

3.13 (8,195)

−0.55

Yi

2.60 (1,166)

3.45 (6,711)

−0.85

Zhuang

2.60 (3,187)

3.09 (18,079)

−0.49

Korean

1.68 (252)

1.88 (2,894)

−0.20

Manchu

1.82 (7,314)

2.00 (9,031)

−0.18

Tujia

2.37 (2,234)

2.58 (6,815)

−0.21

coefficients of intermarriage in the gross model are statistically significant (p<0.05). The exponentiated values, namely, the IRRs indicate how much higher, or lower, exogamous women’s average CEB is compared to that of endogamous women. So, the CEB of women who out-marry to Han is 19 percent lower than that of endogamous women; and the CEB of women who out-marry to other minority group members is 11 percent lower than that of endogamous women. The second and third columns of Table 5.3 show the results of two more Poisson regression equations that progressively add policy and other socioeconomic variables to the equations. In the second column of Table 5.3, the policy variable is added to control for the effect of the family planning policy. The effects of the two intermarriage variables drop a little, but remain statistically significant. The last model adds the socioeconomic variables, the effect of intermarriage on women’s fertility drop even more, but still maintains its significance. The fertility of women intermarried to Han is 10 percent less

The impact of intermarriage on the fertility of minority women

69

Table 5.3 Reduced Poisson regression models predicting number of CEB: minority ever-married women, aged 15–49, China, 1990 Independent variables

Gross model coef. (IRR)

Model 1 coef. (IRR)

Model 2 coef. (IRR)

Intermarriage W/Han

−0.210 (0.811)

−0.164 (0.849)

−0.104 (0.901)

W/Other Groups

−0.116 (0.891)

−0.063 (0.939)

−0.032 (0.969)

—

—

—

Mongolian

−0.159 (0.853)

−0.115 (0.891)

−0.084 (0.920)

Hui

−0.152 (0.859)

−0.161 (0.851)

−0.155 (0.857)

Tibetan

0.282 (1.326)

0.239 (1.270)

0.134 (1.144)

Uygur

0.213 (1.236)

0.250 (1.284)

0.266 (1.304)

Miao

0.020 (1.020)

0.040 (1.041)

−0.026 (0.975)

Yi

0.099 (1.104)

0.153 (1.165)

0.077 (1.080)

Korean

−0.497 (0.608)

−0.504 (0.604)

−0.382 (0.683)

Manchu

−0.391 (0.676)

−0.343 (0.710)

−0.307 (0.736)

Tujia

−0.150 (0.861)

-0.131 (0.878)

−0.159 (0.853)

—

—

—

1.789 (5.982)

1.682 (5.378)

Endogamous Ethnicity

Zhuang Policy Percent of childbearing year before 1979 Socioeconomic characteristics Year of schooling

−0.024 (0.976)

Prof/leaders

−0.259 (0.772)

Employed

−0.041 (0.960)

Permanent migrant

−0.213 (0.808)

Long-term floater

−0.119 (0.888)

Short-term floater

−0.193 (0.825)

Non-migrant Constant

— 1.134

0.430

0.639

Fertility, family planning, and population policy in china Pseudo R2 Goodness of Fit χ

2

70

0.039

0.145

0.154

110676.2

68053.6

64267.0

Note All coefficients are significant at p<0.05.

than that of endogamous women, and that of women intermarried to other ethnic groups 3 percent less. The results in Table 5.3 show that even after adding a series of control variables, the effects of intermarriage on fertility are maintained. In the bottom panel of Table 5.3 are the Poisson Goodness of Fit χ2 test statistics to compare the observed empirical distribution with the distribution predicted by the Poisson regression model. The null Hypothesis (H0) is that there is no difference between the observed data and the modeled data, indicating that the Poisson model fits the data. A small χ2 value (p>0.05) means that one cannot reject the null hypothesis that the observed CEB data are Poisson distributed (Long 1997; Poston 2002). In all three models in Table 5.3, the values of the Poisson Goodness of Fit χ2 statistics indicate that using Poisson regression to model the CEB is appropriate. In Table 5.4, two more Poisson regression models are estimated to compare the intermarriage effects on fertility for rural and urban minority women. The results show that rural minority women who intermarried with Han have 5.7 percent less children than rural minority women who married to their own groups; and rural minority women who out-married to other minority group members have 2.9 percent less children than endogamous rural women. Among urban women, the effect of intermarriage to Han is even stronger. Urban minority women who intermarried to Han have 12.3 percent less children than other endogamous urban women. However, among urban women the effect of intermarriage to other minority members is not statistically significant.

Implications and conclusion In this chapter, the effects of intermarriage on the fertility of ever-married women from ten ethnic minority groups are examined. This kind of quantitative analysis has not previously been undertaken, mainly because appropriate intermarriage data have not been available. In the descriptive analyses, clear patterns were shown for intermarried women and their levels of fertility. For example, the Manchu minority group is highly assimilated, has a high intermarriage rate, and low fertility. This is a very different pattern compared to that of Tibetan women with a very low intermarriage rate and high fertility. Presented next were three Poisson regression models that first introduced the intermarriage variables and then progressively controlled for a variable representing family planning policy and other socioeconomic variables. The most important finding was that the effect of intermarriage was maintained even after adding the control variables. This suggests that the effect of intermarriage is important and this effect of assimilation should not be neglected when examining ethnic fertility differentials.

The impact of intermarriage on the fertility of minority women

71

Table 5.4 Poisson regression models predicting number of CEB, by rural and urban, minority evermarried women, ages 15–49, China, 1990 Children ever born

Rural model

Urban model

coef

IRR

coef

IRR

W/Han

−0.058

0.943

−0.130

0.877

W/Other Groups

−0.030

0.971

−0.029*

0.972

—

—

—

—

Mongolian

−0.066

0.936

0.166

1.181

Hui

−0.025

0.975

−0.058

0.944

Tibetan

0.180

1.198

0.215

1.240

Uygur

0.269

1.308

0.553

1.738

Miao

0.002*

1.002

0.061*

1.063

Yi

0.103

1.108

0.020*

1.020

Korean

−0.400

0.670

−0.101

0.904

Manchu

−0.325

0.723

−0.039*

0.962

Tujia

−0.163

0.850

0.066*

1.069

—

—

—

—

1.728

5.627

1.766

5.846

Year of schooling

−0.013

0.987

−0.024

0.976

Professional/leader

−0.179

0.836

0.002*

1.002

Employed

−0.052

0.949

−0.204

0.816

Permanent migrant

−0.239

0.787

−0.069

0.933

Long-term floater

−0.148

0.862

0.033*

1.033

Short-term floater

−0.237

0.789

−0.045*

0.956

—

—

—

—

Intermarriage

Endogamous Ethnicity

Zhuang Policy Percent of childbearing Years before 1979 Socioeconomic characteristics

Non-migrant Constant

0.587

0.294

Fertility, family planning, and population policy in china Pseudo R2 Goodness of Fit χ

2

Number of observations

72

0.150

0.150

54,083.3

7405.5

85,494

16,721

Note * coefficients are not significant at p<0.05.

In addition, all three models in Table 5.3 showed that both intermarriage variables have negative effects on fertility irrespective of whether the intermarriage is between a minority and a Han or between a minority and another minority. Thus exogamous women have lower levels of fertility than endogamous women. In particular, for minority women who intermarried to Han, they appear to be more willing to accept the mainstream culture. As suggested by Bean and Tienda (1987) and other scholars (Goldscheider and Uhlenberg 1969; Halli 1987), the numerical smallness, low socioeconomic status, and disadvantages minority members experience in the society may well lead to their having feelings of marginality and insecurity. And “these feelings are experienced most acutely by those who aspire to greater mobility and who are therefore more sensitive to the obstacles placed in their paths by patterns of discrimination” (Bean and Swicegood 1985:7). Accordingly, intermarried women, especially those who married Han men, may be more likely to look down on their ethnic cultures and compensate for greater mobility by lowering their fertility. Finally, two Poisson regression models were estimated that compare the effects of intermarriage on fertility for rural and urban women. It was shown that the different fertility patterns may results also from lower socioeconomic status of rural women, and from the fact that the policy is more strictly enforced among urban women. But the effect of intermarriage, while stronger among urban women than among rural women, was nevertheless maintained. China’s minority policy protects minority members who reside in rural areas, allowing them legally to have more than one child, and sometimes more than two children. Urban minority members usually are allowed only to have one child, as is the rule for the majority Han. Many scholars have noted the success of China’s one-child family planning policy in successfully reducing China’s population growth since the late 1980s. However, compared with the greater fertility reduction of the Han population in general, China’s minority nationalities have increased their population by 35.5 percent from 1982–1990, which is 23 percentage points higher than the Han nationality (Park and Han 1990; Zhang 1991). This growth in the minority populations has been partially due to a change in many individuals’ declaration of ethnic background (Banister 1987), but also to the more lenient enforcement of the policy. Therefore, the various family planning regulations for minorities have resulted in increases in their numbers and has enlarged the variation in fertility among the minority groups. This has increased the importance of other social, economic, and cultural variables on their fertility, including ethnic assimilation and intermarriage. The issue of ethnic assimilation is particularly important when the fertility rates of the minority populations are examined. And since intermarriage is an important consequence of ethnic assimilation, it will be particularly noteworthy to move beyond the research

The impact of intermarriage on the fertility of minority women

73

reported in this chapter. There are several steps that may be taken to further explore these issues and relationships. For instance, one area requiring attention is the ethnic identity issue of the second generation of intermarried couples in relation to the dynamics of overall population growth and the future population policy. In addition, attention needs to be directed to the phenomenon of multiple ethnicities and ethnic inequalities, an area just now being developed in Western analyses, but not yet in China.

References Axelrod, P. 1990. “Cultural and Historical Factors in the Population Decline of the Parsis of India.” Population Studies, 44(3), November: 401–419. Banister, J. 1987. China’s Changing Population. Stanford, CA: Stanford University Press. Bean, F.D. and G.Swicegood. 1985. Mexican American Fertility Patterns. Austin, TX: University of Texas Press. Bean, F.D. and M.Tienda. 1987. The Hispanic Population of the United States. New York: Russell Sage Foundation. Beaujot, R.P., K.J.Krotki, and P.Krishnan. 1982. “Analysis of Ethnic Fertility Differentials through the Consideration of Assimilation.” International Journal of Comparative Sociology, 23(1–2), March-June: 62–70. Cameron, A.C. and P.K.Trivedi. 1988. Regression Analysis of Count Data. Cambridge, UK: Cambridge University Press. China Handbook Editorial Committee. 1985. China Handbook Series: Life and Lifestyles. Beijing, China: Foreign Languages Press. Coleman, D.A. 1994. “Trends in Fertility and Intermarriage among Immigrant Populations in Western Europe as Measures of Integration.” Journal of Biosocial Science, 26(1), January: 107– 136. Goldscheider, C. and P.R.Uhlenberg. 1969. “Minority Group Status and Fertility.” American Journal of Sociology, 74:361–372. Gordon, M.M. (1964). Assimilation in American Life: The Role of Race, Religion, and National Origins. New York: Oxford University Press. Halli, S.S. 1987. How Minority Status Affects Fertility: Asian Groups in Canada. NY: Greenwood Press. Heberer, T. (1989). China & Its National Minorities: Autonomy or Assimilation? Armonk, NY: M.E.Sharpe, Inc. Hout, M. and J.R.Goldstein. 1994. “How 4.5 Million Irish Immigrants Became 40 Million Irish Americans: Demographic and Subjective Aspects of the Ethnic Composition of White Americans.” American Sociological Review, 59(1), February: 64–82. Long, J.S. 1997. Regression Models for Categorical and Limited Dependent Variables. Thousand Oaks, CA: Sage Publications. Ma, Y. 1992. China’s Minority Nationalities. Beijing: Foreign Languages Press. Park, C.B. and J.Q.Han. 1990. “A Minority Group and China’s One-Child Policy: The Case of the Koreans.” Studies in Family Planning, 21(3), May/June: 161–170. Poston, D.L., Jr (2002). “The Statistical Modeling of the Fertility of Chinese Women.” Journal of Modern Applied Statistical Methods, 1:387–396. Poston, D.L., Jr and J.Shu. 1987. “The demographic and socioeconomic composition of China’s ethnic minorities.” Population and Development Review, 13:703–722. Poston, D.L., Jr, C.F.Chang, and H.Dan. 2006. “Fertility Differences Between the Majority and Minority Nationalities in China.” Population Research and Policy Review, 26(2), April 2007.

Fertility, family planning, and population policy in china

74

State Statistical Bureau. 1991. Zhong Guo 1990 Nian Renkou Pucha 10% Chouyang Ziliao (10 Percent Sampling Tabulation on the 1990 Population Census of the People’s Republic of China). Beijing: China Statistical Publishing House. State Statistical Bureau. 1992. One Percent Public Use Sample of the 1990 Census of China, Data Tapes. Beijing: State Statistical Bureau. Zhang, T. 1991. “Rapid Population Growth is Unfavorable to prosperity of Minority Areas.” Renkou Xuekan (Population Report), 2:25–27. Zhang, T. 2001. Ethnic Demography. Beijing: China Population Press. Zhang, T. and X.Zhang. 1991. Contemporary China’s Tibet. Beijing: Contemporary China Press. Zou, D. 1991. “China’s ethnic population and relevant policies.” China Population Today, (February): 8–9.

6 Emerging patterns of premarital conception Carol S.Walther

Introduction Many have heard the saying that the first child can come anytime, but the second one takes nine months. This is a reflection of the phenomenon of premarital conception. Once the conception has occurred, couples will often tend to marry for the benefit of the child. In the United States, these marriages have been referred to as “shotgun” weddings, in which it was believed that the father of the pregnant daughter brandished a shotgun to ensure that the son-in-law marries his daughter. However, with the sexual revolution in the United States, fewer couples decide to marry just because of a premarital conception. For instance, in the United States, with increases in cohabitation and nonmarital fertility, Bianchi and Spain (1996:11) estimate that “thirty-five percent of cohabiting households include children under the age of 15 in 1994, compared with 27 percent in 1980.” This indicates a steady rise in the United States of having children outside marriage (Graefe and Lichter 1999). Rindfuss and Morgan (1983) suggested more than 20 years ago that Asian countries were having a “silent sexual revolution.” At the time China was not included among these countries. This chapter considers nonmarital conceptions in China. It examines the prevalence of premarital conceptions in China and seeks to determine the effects of premarital conceptions on various independent variables, including sex education. Data are used from the 1997 Sample Survey of Population and Reproductive Health conducted by the China Population Information and Research Center and the State Family Planning Commission of China to examine the likelihood of a Chinese woman having a premarital conception. A premarital conception is said to have occurred if the woman’s marriage date occurred 8 months or less before a birth, 6 months or less before a stillbirth, or 3 months or less before an induced abortion or miscarriage. Recent studies from other Asian countries have demonstrated increases in premarital conception (Wolf 1972). For example in South Korea, the marriage and birth interval have decreased from over 27 to 11 months from the time of marriage to the time of birth (Lee 1987). In Taiwan, 76 percent of women had first sexual intercourse before marriage, while 58 percent of women reported engaging in sexual intercourse before engagement (Lee 1988). Additionally, 32 percent of currently married Taiwanese women reported being pregnant before marriage (Cernada et al. 1986). Given these new changes in Asian societies related to sexuality, it will be of interest to analyze the likelihood of Chinese women having premarital conceptions. It is hypothesized that attitudes toward premarital coitus, along with the woman’s education, residence, and ethnicity should affect the

Fertility, family planning, and population policy in china

76

likelihood of having a premarital conception. The phenomenon of premarital conceptions may indeed reflect a “silent sexual revolution” in China, for which the Chinese government may be unprepared.

Prior literature Due to presumptions of marriage and social activity only occurring within marriage, few researchers have studied premarital conception in China. Norms in China about premarital sex have been extremely rigid until recently (Kaufman et al. 1995:141; Reiss 1967; Wang and Yang 1996). In fact, the 1997 China data-set used here was designed under the assumption that single women do not have premarital sex. In the survey’s design single women were never asked questions such as the numbers of pregnancies, abortions, stillbirths, or miscarriages. The China Marriage Law of 1981 requires that the minimum age for women be 20 to marry, while men must be 22 (Kaufman et al 1995:141–142). Cheng (1995) contends that in rural areas, women often marry before reaching the age of 20. Furthermore, 20 years ago it was illegal for couples to cohabit, which reduced the opportunities and normative behavior of premarital sex. Given these legal and social constraints, researchers have suggested that premarital conception infrequently occurred in China. Rindfuss and Morgan have asserted that a “silent” but “profound sexual revolution has been taking place in Asia,” in the past 20 years because of changes in marriage arrangements, and this has increased the number of premarital conceptions (1983:273; Bo and Geng 1992; Wang and Yang 1996). It is likely that marriages in China, as well as in other Confucian societies, have changed from arranged marriages to romantic marriages (Wang and Yang 1996). Thus, increases in premarital conception should result from increased opportunities of social interaction between couples. Rindfuss and Morgan argued that [A]lthough a premarital birth is considered quite unacceptable in Confucian-heritage groups, the ethnographic literature repeatedly asserts that there is no shame attached to a premarital conception; in fact, it is often taken as a propitious sign that the woman is fertile. (1983:273) Therefore, the stigma of a premarital conception may not be as prevalent in China as it is in the United States; thus couples who are planning on marrying may well engage in premarital coitus. Not only are romantic marriages becoming more prevalent than arranged marriages, Kaufman and her colleagues (1995) assert that attitudes toward premarital sex have changed. First, increasingly, college students report having experienced intercourse, while over half did not disapprove of couples engaging in premarital coitus (Jankowiak 1993; Kaufman et al. 1995:142). Moreover, Kaufman and her colleagues found that among teenagers in Sichuan province, 70–80 percent did not view premarital sex as morally or socially wrong, and only approximately 20 percent of the teenagers reported

Emerging patterns of premarital conception

77

that a man would lose respect for his girlfriend if she agreed to have sexual relations with him (1995:147). Also, Fan and Pau (1998), using data from the 1996 Youth Sexuality Survey, reported that secondary school students were becoming more accepting of cohabitation, homosexuality, and premarital sex. In this sample, 17 percent of the students reported having been pregnant, six percent had premarital pregnancies, and 80 percent of the students’ pregnancies resulted in abortion (Fan and Pau 1998). Wang and Yang write that “sexuality is no longer taboo in public discussion and is not presented as a matter solely of reproduction. Instead, it has been openly treated as a source of individual pleasure and satisfaction” (1996:314). In addition to increasing permissive attitudes toward premarital sex, the intervals between marriage and first birth have significantly decreased, suggesting the impact of premarital coitus. W.She (1998) finds that premarital pregnancy has increased steadily over time, especially between 1987 through 1991. Furthermore, Cheng argues that the shortening of the interval suggests that it is a direct result from early conception (1995:10). Wang and Yang contend that the marriage and birth interval has dropped from 34 months in the 1950s to less than 18 months in the 1980s (1996:299). However, they argue that this reduction is not only due to the one-child policy, but also due to changes in marriage, education, nonfamilial employment, and young couples’ sexual behavior (Wang and Yang 1996:300). Thus, the reduction in the interval between birth and marriage has decreased, and premarital conceptions are believed to contribute substantially to fertility in China, whether directly or indirectly. Although Merli and Raftery (2000) exclude from their study all women who have births less than 9 months from a marriage date (4.5 percent of their sample), they assert that women are extremely reluctant to report births before marriage or less than 9 months after marriage. This exclusion of premarital conceptions in Merli and Raftery’s analysis reinforces traditional assumptions of premarital sex in China. However, in their very exclusion they acknowledge the existence of premarital conception. This chapter now turns to hypotheses and issues of conceptualization.

Hypotheses Traditional fertility paradigms suggest strong correlations between education, geographic locality, ethnicity, and women’s fertility. If China is indeed undergoing a silent sexual revolution, one might suggest that these social characteristics would similarly impact premarital conception. First, social demographic characteristics such as region, ethnicity, marriage age, and the one-child policy are used to determine the chances of a premarital conception. With regard to geographic region, Gao and associates (1993) found that 81 percent of women living in rural areas had premarital sex compared to 64 percent living in urban areas. In urban areas, there is presumably more social control regarding marriage and premarital sex than in rural areas. Therefore, if a woman lives in an urban locale, as opposed to a rural one, she will be less likely to have a premarital conception. The second social demographic characteristic is ethnicity. Given current governmental and social policies related to reproduction and stricter enforcement among Han women, it is expected that Han women will be less likely to have a premarital conception.

Fertility, family planning, and population policy in china

78

Third, age at marriage is another significant social demographic characteristic that should impact premarital conception. As a woman’s age increases, her fertility likewise increases. It has been suggested that older women have greater access to information and are thus more likely to recognize when they become pregnant. Also, if women have been encouraged to deter their childbearing until later in life, once an engagement has been arranged, a premarital coitus may occur. Thus it is hypothesized that age will be positively associated with the likelihood of having a premarital conception. Fourth, in 1979, concerned about population growth, the Communist party implemented the one-child policy, which has forcefully governed women’s fertility (see Chapter 1 by Liang and Lee in this volume). The policy was grounded in the idea that “one child is best, at most two, never a third” (Poston and Falbo 1992:427). This policy has been strictly enforced in urban areas, but less rigorously in rural areas. For instance, in rural areas if a couple’s first child is a daughter, they are frequently allowed to have a second child (Poston and Falbo 1992:427). Thus, it is hypothesized that if the pregnancy occurred in 1980 or later, the woman would be more likely to have had a premarital conception. Fifth, the wealth flows and feminist paradigms of fertility argue that as women’s education increases, fertility decreases (Greenhalgh 1990; Riley 1999). For example, Gao and associates (1993) have shown that as educational level increases, age of first intercourse also increases, thus limiting opportunities for premarital conception. It is thus hypothesized that education will be negatively associated with having a premarital conception. Sixth, research on the effects of sex education on fertility has demonstrated mixed results. For instance, Wang (1997) concludes that women in China who received better education were more likely to use contraception. However, Kaufman and colleagues (1995:153) note that sex information given by friends had a positive effect on men’s attitudes toward premarital sex, and that sex education itself, as taught in schools, had little influence on attitudes on both men and women. Although Kaufman and her associates assert that sex education through the schools has no influence on attitudes toward premarital sex, it is contended here that women who acquire sex education will learn about birth control and signs of pregnancy. Thus it is hypothesized that sex education will be positively associated with the likelihood of having a premarital conception. Finally, if a sexual revolution has indeed been occurring in China in the last 20 years, changes in attitudes should also be expected. If there has been attitudinal change, it has likely been from conservative views of premarital sex to more liberal ones. Thus, if one agrees with the statement that premarital sex may occur between two people about to marry, it is hypothesized that such a person will have increased odds of having a premarital conception. Further, knowing couples who participate in premarital sex is also expected to increase one’s odds of having a premarital conception.

Data The data used to test the hypotheses given earlier are taken from the 1997 Sample Survey of Population and Reproductive Health in China conducted by the China Population

Emerging patterns of premarital conception

79

Information and Research Center and the State Family Planning Commission of China (see Chapter 2 in this volume by Wu and Walther for more discussion). This is a national survey of women aged 15–49, and it includes questions about pregnancies, abortions, miscarriages, contraceptive usage, and marital status. Over 60 percent of the pregnancies resulted in a birth of a son or daughter. Most of the pregnancies occurred between 13 and 24 months after marriage, suggesting that most conceptions occur within marriage. The sample is necessarily restricted to all ever-married women who have had at least one birth. Researchers have suggested numerous problems with survey data from China. Both Western and Chinese scholars agree that Chinese births tend to be under-reported on census, surveys, and registration systems. This is further complicated by respondents’ failures to report births and infant deaths (Banister and Preston 1981; Lavely 1982; Banister 1987, 1989; Merli and Raftery 2000). One possible reason for the underreporting is the one-child policy and its enforcement. Merli and Raftery write that “in recent years, however, underreporting has become more widespread because of the major criterion for evaluating cadres’ job effectiveness at every level” (2000:111). In other words, couples fear punishment for exceeding birth quotas or having a child without official permission, and officials underreport to maintain quotas. Moreover, conceptions before marriage may be misreported as conceptions occurring just after marriage. Therefore, the first pregnancy may not be reported if official permission has not been given, and thus, the second birth may be reported as the first. Finally, women may have difficulty recalling dates of pregnancies or may fail to report miscarriages. The data used in this chapter may be affected by any or all of these problems, especially in rural areas.

Operationalization and description The dependent variable is the occurrence of a premarital conception. It is operationalized by first determining the number of months between the date the woman was married and the date of her first birth. A premarital conception is defined as a birth that occurred eight months or less since the date of marriage, or as a stillbirth that occurred six months or more since the date of marriage, or as an induced abortion or miscarriage that occurred three months or less since the date of marriage. For each woman this is a dummy variable, coded 1 if the woman had a premarital conception, 0 if not. About 2 percent of the women report having had a premarital conception, as just operationalized (Table 6.1). The first independent variable is rural residence, a dummy variable coded 1 if the woman resides in a rural area. About 77 percent of the women in the sample are rural residents. The next variable is age at marriage. On average women in the sample married at age 21.5 years. The third variable is Han nationality, a dummy variable coded 1 if the woman is a Han. Over 91 percent of the women in the sample are Han. Another variable endeavors to capture the effects of the one-child policy. If a woman had her first pregnancy in 1980 or later, she is coded 1 on the policy variable; if her first pregnancy occurred before 1980, she is coded 0 on this variable. Almost three-quarters of the women had their first pregnancy after 1980 (Table 6.1).

Fertility, family planning, and population policy in china

80

Table 6.1 Dependent and independent variables: means and standard deviations, China, 1997 Variables

Mean

Standard deviation

Dependent Premarital conceptions

0.02

0.15

Rural

0.77

0.42

Marriage age

21.5

2.80

Han

0.91

0.29

One-child policy

0.74

0.44

Education

6.97

4.53

Sex education

0.01

0.30

Premarital

0.13

0.33

Knows no one

0.38

0.49

Knows some

0.44

0.50

Knows lots

0.18

0.38

Independent

Two education variables are included. The woman’s formal education is measured in years. On average women report just under seven years of education completed. Data are also available in the survey on whether each woman reports having ever received sex education. This is operationalized as a dummy variable, coded 1 if the woman reports having received sex education. Over 99 percent of the women have not received any sex education. The last 4 independent variables refer to premarital sex with regard to the woman’s attitudes and friendship networks. The survey asked the respondents if they approve of 2 people who are going to marry having premarital sex. Women who responded “yes” are coded 1 on the premarital sex dummy variable. Of the women, 13 percent answered yes to this question. The survey also asked each woman, “As far as you know, is there anybody around you who had sex before marriage? If yes, how many? None, Some, Lots.” The responses to this question were coded into 3 dummy variables about friends having premarital sex: knows no one, knows some, and knows lots. Around 38 percent of the women reported knowing no one having premarital sex, 44 percent reported knowing some, and 18 percent reported knowing lots of persons having premarital sex.

Method Logistic regression is used for model estimation because the dependent variable of premarital conception is a dichotomous variable, scored one if the woman had a premarital conception.

Emerging patterns of premarital conception

81

The regression equation predicting the likelihood of a woman having a premarital conception is: L=Intercept+Rural(X1)+Marriage Age(X2)+ Han(X3)+One-Child Policy(X4)+Education(X5)+Sex Education(X6)+ Premarital(X7)+Knows Some(X8)+Lots(X9)+e. The “Knows No One” variable is used as the reference. The logit coefficient for each independent variable from the logistic regression may be interpreted as the increase or decrease of the log-odds of a premarital conception. The odds ratio is defined as: Odds ratio=eL, where L=Log odds. Odds ratios range from zero to infinity (Hamilton 1992:220) and are interpreted as the increase or decrease in the odds of an event occurring (Agresti and Finlay 1997:581; Hosmer and Lemeshow 1989).

Results Table 6.2 presents the results of the logistic regressions. Three models are presented. The first model includes only sociodemographic characteristics, the second model adds educational variables, and the third adds individual and peer attitudes towards premarital coitus. In Table 6.2, in Model 1, 4 sociodemographic characteristics are examined. Women who live in rural areas are more than 31 percent less likely to report having had a premarital conception (i.e. 100*(1−e−0.36)), a result not consistent with the hypothesis. An increase in age of marriage is positively associated with having a premarital conception. Also, women who had their first birth after 1980 are 25 percent more likely to have had a premarital conception. The Han nationality coefficient is not statistically significant. The results in Model 1 indicate that sociodemographic characteristics are indeed associated with the likelihood of having a premarital conception. In Model 2, in addition to sociodemographic characteristics, two education variables are added. In this model, the rural and age at marriage variables remain statistically significant, while the variable, one-child policy, becomes insignificant, but remains positive. The Han variable is still not significant. Of particular interest is the fact that women who report having received sex education are more than twice as likely to have had a premarital conception compared to those who have not received sex education. There is also a slight positive, and statistically significant effect of formal education on the likelihood of having a premarital conception. In Model 3, attitudes toward premarital sex and knowing peers who engage in premarital coitus are added. Of the sociodemographic characteristics, only the age at marriage variable remains statistically significant. Of the education variables, only the sex education variable remains statistically significant. Women who have received sex education are still twice as likely as women not receiving such education to have had a premarital conception. With respect to the attitude variables, women who agreed that it was alright for two people about to marry to have premarital sex are 59 percent more likely to have had a premarital conception. And women who have many friends

Fertility, family planning, and population policy in china

82

Table 6.2 Logits and odd ratios of Chinese women having a premarital conception, 1997 Model 1 logits odd ratios

Model 2 logits odd ratios

Model 3 logits odd ratios

Sociodemographic characteristics Rural

Han

Marriage age

One-child policy

−0.360**

−0.290**

−0.210

0.690

0.750

0.810

−0.140

−0.140

−0.160

0.870

0.870

0.850

0.008**

0.007**

0.008**

1.008

1.007

1.008

0.220*

0.071

0.050

1.250

1.073

1.050

0.021*

0.021

1.022

1.022

0.762**

0.720**

2.133

2.063

Education Education

Sex education

Attitudes Premarital

0.461** 1.592

Some

0.110 1.112

Popular

0.520** 1.690

Knows no one Constant Person χ

2

Pseudo R

2

Log likehood N Degrees of freedom

Reference −5.3650**

−4.83**

−5.19**

1040.39**

1612.08**

5528.99**

0.0170

0.03

0.04

−1623.79

−2324.02

−2299.54

12012

12012

12012

4

6

9

Emerging patterns of premarital conception

83

Notes * significant at 0.05. ** significant at 0.01.

who participated in premarital sex are 69 percent more likely to have had a premarital conception compared to women who know no one engaging in premarital sex.

Conclusion In conclusion, premarital conceptions are occurring in China, but they remain a small proportion of conceptions (well under 3 percent). Given this small sample, this may suggest some underreporting. Merli and Raftery (2000) suggest that women in China who become pregnant before marriage may wait to report conception until after marriage has occurred for fear of social punishment and to ensure that birth quotas remain low. Furthermore, in some cases, the reporting of abortions and miscarriages depend on the woman remembering the date of pregnancy, which may not always occur. In the results reported in the previous pages, the age at marriage variable remained significant in all three models. With increases in women’s age at marriage, there will be increases in the probabilities of having premarital sex resulting in increases in premarital conception. Thus, it would appear to be reasonable to expect that in China if the average age of marriage increases for women due to marriage laws or educational attainment, the country may well see an increase in premarital conception. Furthermore, attitudes about premarital sex seem to be changing, which also may alter behavior. As discussed earlier, a sexual revolution has been occurring in China with increases in the prevalence of more liberal attitudes toward premarital sex and the frequency of young people engaging in premarital sex. In the analysis reported in this chapter, women who had more liberal attitudes about premarital sex, and those who knew a lot of couples engaging in premarital sex were much more likely to have had a premarital conception. Researchers have debated the potential effects of sex education for women and men. For instance, Gao and associates (1993) suggest that sex education should occur through the media, schools, work units, and families so as to decrease premarital sex and conception. Fraser (1976) calls for more sex education for teenagers so that sexual education is not so discrete, muted, and selective. Wang and Yang (1996) contend that family planning is sex education. In the research reported here, women who received sex education were much more likely than those not receiving sex education to have had a premarital conception. This supports research by Cheng (1995) that sex education increases premarital coitus. These research results have some fairly significant policy implications related to marriage laws, family planning, and sex education in China. However, two limitations of the study need to be addressed. First, there could well be serious problems with the measurement of the dependent variable because women may not remember the exact date of conception. Second, a baby born eight months from the date of marriage may not be a premarital conception, but a premature birth.

Fertility, family planning, and population policy in china

84

Finally, the results reported here are an important addition to the current literature. Few studies about China and sexual activity have described premarital conception that may lead to marriage. Most have examined the effects of sex education on the reduction of premarital sex and attitudes leading to increases in premarital coitus. Moreover, the Chinese government has failed to implement programs to reduce premarital conception. By not examining premarital conceptions, China may continue through a sexual revolution for which it may be unprepared.

References Agresti A. and B.Finlay. 1997. Statistical Methods in the Social Sciences. Third Edition. Upper Saddle River, NJ: Prentice Hall. Banister, J. 1987. China’s Changing Population. Stanford, CA: Stanford University Press. Banister, J. 1989. A New Survey of Infant Mortality in China: A Research Note. Washington, DC: US Bureau of the Census. Banister, J. and S.H.Preston. 1981. “Mortality in China.” Population and Development Review, 7(1):98–110. Bianchi, S.M. and D.Spain. 1996. “Women, Work, and Family in America.” Population Bulletin, December 51(3):1–48. Bo, Z. and W.Geng. 1992. “Sexuality in Urban China.” The Australian Journal of Chinese Affairs, 28 July, 1–20. Cernada, G.P., M.C.Chang, H.S.Lin, and T.H.Sun. 1986. “Adolescent Sexuality and Family Planning Awareness, Knowledge, Attitude and Behavior: Taiwan.” Amherst, MA: International Area Studies Programs, Asian Studies Committee Occasional Papers Series No. 12. Cheng, Y. 1995. “Study on Early Marriage and Early Childbirth in China.” pp. 8–88 in W.Li, R.M.Li, H.C.Chen, and D.W.Hahn (eds), Proceedings of the Beijing International Symposium on Fertility Regulation, Fertility Regulation: Present and Future. Bethesda, MD: National Institute of Health, National Institute of Child Health and Human Development. Fan, S. and A.Pau. 1998. “More Open Attitude Towards Sex Among Hong Kong Youth.” People and Development Challenges, 5(9):24–25. Fraser, S.E. 1976. “Sex in China: A Matter of Numbers.” Australian Journal of Interpersonal Relations, 4(11):6–14. Gao, E., Z.Wu, and X.Gu. 1993. “Survey On Sexual Experiences Among Unmarried Women in Shanghai and Solutions.” Chinese Journal of Population Science, 5(2): 95–105. Graefe, D. and D.Lichter. 1999. “Life Course Transitions of American Children: Parental Cohabitation, Marriage and Single Motherhood.” Demography, 36(2):205–217. Greenhalgh, S. 1990. “Toward a Political Economy of Fertility: Anthropological Contributions.” Population and Development Review, 16(1):85–106. Hamilton, L. 1992. Regression with Graphics: A Second Course in Applied Statistics. Belmont, CA: Duxbury Press. Hosmer, D.W. and S.Lemeshow. 1989. Applied Logistic Regression. New York: John Wiley & Sons, Inc. Jankowiak, W.R. 1993. Sex, Death and Hierarchy in a Chinese City: An Anthropological Account. New York: Columbia University Press. Kaufman, G., D.L.Poston Jr, T.A.Hirschl, and J.M.Stycos. 1995. “Teenage Sexual Attitudes in China.” Social Biology, 43(3–4):141–154. Lavely, W.R. 1982. “China’s Rural Population Statistics at the Local Level.” Population Index, 48(4):665–611.

Emerging patterns of premarital conception

85

Lee, S.B. 1987. “Analysing Birth Interval in Korea” pp. 271–312 in Fertility Changes in Korea. Edited by Korea Institute for Population and Health. Seoul, Korea. Lee, T.M. 1988. “A Study of the Stability of First Marriage for Married Women of Childbearing Age in Taiwan Area—An Exploration From the Age of First Marriage, Dimensions of Premarital Pregnancy, and the Way of Marriage Decision-Making.” Journal of Population Studies, 11:33–54. Merli, M.G. and A.E.Raftery. 2000. “Are Births Underreported in Rural China? Manipulation of Statistical Records in Response to China’s Population Policies.” Demography, 37(1):109–126. Poston, D.L., Jr and T.Falbo. 1992. “Effects of the One-Child Policy on the Children of China.” pp. 427–443 in D.L.Poston, Jr and D.Yaukey (eds), The Population of Modern China. New York: Plenum Press. Reiss, I. 1967. The Social Context of Premarital Sexual Permissiveness. New York: Holt, Rinehart and Winston. Riley, N.E. 1999. “Challenging Demography: Contributions from Feminist Theory.” Sociological Forum, 1999. Rindfuss, R.R. and S.P.Morgan. 1983. “Marriage, Sex, and the First Birth Interval: The Quiet Revolution in Asia.” Population and Development Review, 9(2):259–278. Ruan, F.F. 1991. Sex in China: Studies in Sexology in Chinese Culture. New York: Plenum Press. She, W. 1998. “Survey of Premarital Pregnancy in Seven Provinces.” China Population Today, 15(3):14–15. Wang, F. and Q.Yang. 1996. “Age at Marriage and the First Birth Interval: The Emerging Change in Sexual Behavior Among Young Couples in China.” Population and Development Review, 22(2):299–320. Wang, J. 1997. “Pre-Marital Pregnancy in Shanghai.” China Population Today, 14(5): 18–19. Wolf, M. 1972. Women and the Family in Rural Taiwan. Stanford, CA: Stanford University Press.

7 Changing patterns of desired fertility Li Zhang, Xiaotian Feng, and Qingsong Zhang

Introduction China has the world’s largest population, and the most systematic family planning program. In the early 1970s, the Chinese authorities initiated the “later-longer-fewer” family planning program to control the birth rate, and in 1979 they implemented a “onechild” family planning program. Young couples are encouraged to have only one child, regardless of sex (see Chapter 1 in this volume by Liang and Lee for more discussion). If one were to trace China’s total fertility rate (TFR) over the past few decades, a definite declining pattern is seen (see specifically Figure 12.1 in this volume by Poston and Glover). In 1950, China’s TFR was 5.6; it dropped to 3.6 in 1975, and to 2.8 in 1979 (Coale and Chen 1987). Ten years later, the TFR decreased to 1.8 children per woman. In the 1990–2000 period, the annual population growth rate was 1.1 percent, which is 0.4 percentage points lower than the annual growth rate for the 1980–1990 period. China’s actual TFR is described as having “undergone a fertility transition since the late 1990s— from below replacement level fertility to a substantially lower fertility” (Wong 2001:69– 70). In demography, actual fertility and desired fertility are assumed to be related to each other. Pritchett described their relationship as follows: Empirically, desired fertility plus a constant is an excellent prediction of actual fertility; desired fertility can be used as an explanatory variable by using both retrospective and prospective data together with the statistical estimation techniques of instrumental variables; in explaining differences in fertility such measures as contraceptive use, “unmet need,” and family planning effort have an empirical small impact and explain very little of fertility variation once differences in desired fertility are accounted for. (1994:621) This chapter examines desired fertility patterns in China, as opposed to actual fertility. The main focus is on the period after the “one-child” family planning policy was initiated. Knowing this could assist in the prediction of the fertility patterns in the future, and would provide useful information to the writers of population policy in China. The World Fertility Survey (WFS) and the Demographic and Health Survey (DHS) programs have conducted numerous household surveys in a large number of countries. Using data from these surveys, researchers have derived three indicators of fertility preferences. The first draws on women’s responses to a question about their ideal number of children and is used to compute the “average ideal number of children” (AINC). A second measure of fertility preferences is the “desired total fertility rate” (DTFR), which recalculates the TFR in each country from age-specific birth rates after subtracting the

Changing patterns of desired fertility

87

number of actual births that exceed each woman’s reported desired family size (Lightbourne 1987; Westoff 1991). A third approach calculates the “wanted total fertility rate” (WTFR) by using answers to questions about women’s future desire for children to classify births (or current pregnancies) as wanted or unwanted (Bongaarts 1990). The first of these measures is used in the analysis to follow.

Approach In this chapter, the main focus is on the AINC and the desired gender of children. Researchers have undertaken many surveys in the last two decades in urban and rural China and have derived data on this measure. The assumption here is that such data give rough, if conservative, indicators of the number and gender of offspring Chinese people would like to have. A meta-analysis will be conducted in this chapter that includes articles and reports on desired fertility from 1979 through 1998. The emphasis throughout will be at the level of description and simple analysis because very little of the original data are now available. The implications of the changing patterns of desired fertility will also be analyzed. Before conducting the meta-analysis of the data, however, a review of some of the literature on Chinese fertility and the desired fertility among the world is needed.

Prior analyses Chinese fertility studies may be categorized into three main areas. The first type of study focuses on the changing patterns of China’s actual fertility. Coale (1984) reported the declining patterns of TFR from 1952 to 1982 in both rural and urban China, Lavely (1984) examined the rural Chinese fertility transition during 1950s to 1980 based on interviews with ever-married women in Shifang Xian of Sichuan province, Zeng (1986) indicated China’s declining fertility through examining the shrinking family size from 1949–1982, while Poston (1992) reviewed developments in fertility and family planning in China from the founding of the People’s Republic of China (PRC) in 1949 to 1982. Some demographers further focused their attention on special time periods. For instance, Banister (1984) examined the contraceptive use, marital pattern, and especially the fertility trends in China during 1978–1983; and found that the fertility rates in China were still in a slowly declining pattern and underreporting of births was serious. Gu and Yang (1991) discussed the fertility trends in rural China during the 1980s; and noted that except for Xinjiang and Tibet all subregions had a TFR 4.0 in the 1980s and the fertility of third and higher parity had declined. Zhang extended the analysis to the 1990s and showed a continuous declining pattern of China’s fertility. Zeng (1996) also indicated that “even after adjustments for serious under-reporting of births, fertility fell substantially between 1990 and 1992…but fertility in China was not far below replacement levels” (Zeng 1996:33). In addition, Feeney (1994) compared the fertility decline in China with that of other East Asian countries and concluded, “the whole of East Asia may have completed a demographic transition in early 1990s” (Feeney 1994:1518). These analyses illustrate the declining pattern of actual fertility in China from 1949 through the 1990s.

Fertility, family planning, and population policy in china

88

A second mainstream of Chinese fertility studies focuses on explaining the declining pattern of actual fertility in China. Some studies showed the important role of China’s family planning program in reducing fertility. Wolf’s (1986) study demonstrates the preeminent role of government intervention in the fertility revolution. Li (1990) also showed that family planning programs played a decisive role in the reduction of fertility. In addition, in their analysis of parity progression and birth intervals in China, Freeney and Wang (1993) pointed out the influence of policy in hastening fertility decline. Other studies show the importance of both family planning programs and socioeconomic change on the fertility transition (Tien 1984; Poston and Gu 1987; Peng 1989; Poston 1990, 2000). Poston’s (Poston and Zhong 1990; Poston 2000) studies illustrate that besides the impact of family planning programs, higher levels of social and economic development are also associated with lower fertility rates. Basically, this body of research reveals that family planning programs, along with socioeconomic factors, are important in explaining the declining patterns of China fertility. A third focus used when examining China’s fertility is the consideration of son preference and its effect on fertility. Arnold and Liu (1986) noted that “incontrovertible evidence of the persistence of son preference is in almost every part of China” (Arnold and Liu 1986:243); but their study showed that son preference was not a major obstacle to the overall success of the family planning program in China (Arnold and Liu 1986). Yet, this body of literature indicated a significant impact of son preference on sex ratio in China. These studies used a variety of levels to conduct the analyses. Lavely and his colleagues (2001) limited their study to the community level by examining the sex preference for children in a community in the Hainan province. Graham and associates’ (1998) focus is at the province level and reports the impact of son preference by comparing the higher sex ratios of the second, third, and higher order births to the first birth. Others look at son preference and sex ratio at the national level (Arnold and Liu 1986; Zeng et al. 1993; Gu 1994; Poston 2002). These studies illustrate that the hazard of having another birth is increased if the first birth was a female. In addition, sex ratios of the second or third or higher order births are higher compared to the first birth. The above studies have mainly focused on actual fertility and its related issues. In terms of desired fertility, Whyte and Gu (1987) studied the Chinese people’s response to China’s fertility transition by examining the desired number of children and desired family size in the 1980s. They discovered that on average “two children” is the most commonly stated preference in both rural and urban areas, but preferred family size seems to be somewhat higher in rural than in urban areas. In addition to this analysis, some other studies have examined desired fertility. Much of this research has been published between 1979 and 1999 and beyond in sociological and demographic journals, and governmental reports, many of them in China. Using these studies, this chapter conducts a meta-analysis of the changing patterns of desired fertility in rural and urban China, as well as the changing patterns of desired gender. The literatures have similar limitations. Whyte and Gu have noted the following: The procedures used in these literatures varied somewhat and are often less than ideal. For example, strict probability sampling procedures were not used in most studies, the wording of the questions varied, and the populations were not always comparable. Perhaps more important, given

Changing patterns of desired fertility

89

the official emphasis on the one-child policy, respondents in these surveys might be expected to understate their family size preferences. Finally, in several cases insufficient information is presented in the sources available to us to ascertain what survey methods were used. (1987:475) In spite of such problems with the data, this chapter, it is believed, should still be of some import in understanding the changing patterns of desired fertility and desired gender of children.

The changing patterns of the desired number of children The changing patterns of desired fertility in rural and urban areas are discussed separately because actual fertility differences are well documented. These differences are due to many causes, and desired fertility should be added as a causal factor of total fertility. Moreover, considering the different labor force participation patterns in rural and urban China, differences in preferences of the number of children desired have emerged In rural China, due to low levels of mechanization and modernization, more laborers, particularly male laborers, are required. In urban China, such requirements are rarely found. Also, the health insurance system and retirement pensions are more readily available for urban residents than for rural residents, resulting in the latter desiring to have more children. These children are a benefit to the parents because of the support they provide their parents during their later years. Based on these reasons, patterns of desired fertility will be separated for rural and urban areas. In China the rural areas are categorized into five subregions: Beijing and Shanghai, Northwestern China, Western China, Middle (Central) China, and Southeastern China. Table 7.1 presents the results from a variety of ideal number of children surveys in several locales. In general, the question used in these surveys was the following: “How many children do you feel it is best to have?” The respondents included husbands, wives, married women, single people, and farmers, as well as people who had children and persons 15–30 years old. “Farmers” were the most popular respondents in these surveys. Several conclusions may be drawn from these studies. First, a majority of respondents in these studies said they preferred 2 children. The exceptions were the people in Huilongguan County, women in Gansu Province, single people in Zhejiang Province, and husbands and wives in Guangdong province. Higher percentages of people in Huilongguan County and higher percentages of single persons in Zhejiang Province preferred to have only 1 child; in contrast, higher percentages of Gansu women and Guangdong husbands and wives favored 3 and more than 3 children. In other words, with few exceptions, 2 children are the most commonly stated preference among respondents in the 5 subregions of rural China. Second, the average desired number of children is slightly lower in the Beijing and Shanghai rural areas compared to other rural subregions, and slightly higher in Southeastern rural areas. For instance, in the Beijing and Shanghai rural areas, the lowest average desired number of children is 1.5 and the highest desired number of children is 1.9 (Gao and Gu 1984; Shaji 1995). On the other hand, in the Southeastern rural areas,

Fertility, family planning, and population policy in china

90

the lowest is 1.4 and the highest is 3.1 (Chen and Lou 1985; Shaji 1995). And in Southeastern rural areas, the average desired numbers of children for most respondents are around 2 or 3. Therefore in terms of the average desired number of children, Beijing and Shanghai rural respondents intend to have less than 2 children; respondents in Northwestern China, Western China, and Middle China prefer to have around 2 children; and the Southeastern respondents favor having 2 or 3 children. Third, the research suggests a declining pattern over time in the number of children desired in most of the 5 rural subregions although there are some fluctuations in each region. Specifically, the average preferred number of children in Northeastern rural areas in 1986 was 2.5; this decreased to 1.6 in 1995 (Chinese Rural Investigation Group 1993). The percentages of respondents who report wanting just 1 child increased dramatically from 13.6 percent in 1987 to 46.0 percent in 1995 (Sun 1988). The average preferred number of children in Western rural areas in 1987 was 2.7 and decreased to 2.1 in 1993 (Liu 1990; Jin 1995). And a similar pattern is also found in the Middle (Central) rural areas and the Southeastern rural areas. Within region differences also occur. For example, an examination of the subregions of Beijing and Shanghai shows that the Shanghai respondents report a higher mean number of desired children than the Beijing respondents. Western rural areas differ from Beijing, Shanghai, and Northwestern rural China. Specifically, socioeconomic conditions in western areas of China have developed more slowly than in other areas. This resulted in cultural fertility traditions remaining more prevalent. Therefore, a more accurate comparison

Table 7.1 Patterns of the desired number of children in rural areas of China Area

Year Provinces/counties

Characteristics of the respondents

Desired number of children (%) 0

1

2

3+ Mean

Beijing and 1979 Rural Beijing Shanghai rural

Had children and 15–30 year old

2.4 26.7 66.5

4.5

1.78

areas

1981 Suburban of Shanghai

Fanners

—

—

—

—

1.97

1986 Rural Counties

Husbands Wives

— —

— —

— —

— —

1.82 1.84

1991 Rural Shanghai Areas

Husbands Wives

— 11.1 82.3 — 12.3 81.8

6.5 5.9

.97 .96

1991 Jiading County, 1992 Shanghai Beijing

Farmers Women who have children

— — — 1.6 29.8 67.4

— 1.2

1.96 1.68

1994 Huilongguan County

Husbands Wives

1.8 50.4 43.3 1.2 51.1 43.1

— —

1.50 1.50

Northeastern

1986 Jilin Province

Farmers

—

—

2.49

rural areas

1987 Jilin Province

Women

— 13.6 62.3 24.1

2.21

—

—

Changing patterns of desired fertility

Western rural areas

91

1988 Jilin Province

Married women

—

5.5 70.5 24.4

2.20

1990 Fengchen County, Liaoning Province

Farmers

—

5.3 69.0 25.7

2.20

1995 Jilin Province

Farmers

— 46.0 48.0

6.0

1.60

1979 Rural Area of Sichuan Predominantly Province single

— 20.6 75.2

4.2

1.79

1987 Gansu Province

Women

—

2.0 27.0 71.0

2.69

1991 Shanxi Province

Husbands Wives

— —

4.6 65.7 29.5 3.8 67.0 29.1

2.35 2.35

—

2.30

1991 Xiaoyi County, Shanxi Women Province

Middle rural area

Southeastern rural areas

—

—

1991

Luochuan County, Shanxi Province

Women

— —

—

2.95

1992

Shanxi Province

Fanners

— 11. 59.9 27.1 0

2.25

1993

Gan and other poor Counties

Husbands Wives

— — — —

— —

2.23 2.05

1983

Tu County, Anhui Province

Women

— 2.7 80.6 16.7

2.14

1984

West Part of Hubei Province

Rich farmers

0.1 2.7 62.5 33.8

2.45

1988 1991 1998 1984

Ten Counties and Cities in Hunan Province Taoyuan County, Hunan Province Rural Area of Hubei Province Zhejiang Province

Farmers Farmers Farmers Single

0.2 3.8 50.6 45.4 — — — — 1.7 15. 73.0 9.8 0 4 40.5 — 59. 1

2.45 2.23 1.91 1.41

1986

Zhejiang Province

Farmers

— —

—

—

2.34

1986

Shandong Province

Farmers

— —

—

—

2.18

1986

Fujian Province

Farmers

— —

—

—

3.16

1989

Wendeng County of Shandong Province

Middle-class farmers

— 15. 78.6 3

6.1

1.91

1989

Shan County of Shandong Province

Poor farmers

— 4.8 65.4 29.8

2.25

1991

Shandong Province

Husbands Wives

— 5.6 69.3 24.7 — 6.2 71.9 21.9

2.26 2.21

1991

Guangdong Province

Husbands Wives

— 0.4 31.0 67.6 — 0.3 28.3 71.2

3.06 3.13

— —

Fertility, family planning, and population policy in china

1994

Taishun County of Zhejiang Province

Poor farmers Middle-class farmers Upper-class farmers

92

— 15. 62.1 22.4 — 5 70.7 12.2 — 17. 68.6 5.7 1 25. 7

2.10 2.00 1.90

of attitude changes in these provinces can be undertaken. The percentages of respondents who favor 1 child are lower compared to Beijing and Shanghai and the Northeastern rural areas, with the exception of the single respondents. In contrast, the percentages of respondents who want 3 and more children are relatively higher than those of the respondents in Beijing and Shanghai and in the Northeastern rural areas. In the rural areas in Middle (Central) China, the results show that the percentages of respondents who “do not want children” and respondents who “want one child” increased from 1983 to 1998. But the majority wanting 2 children stayed above 50 percent. One unique aspect of this area is the increasing percentages of respondents who prefer 3 and more children. In 1983, only 16.7 percent desired 3 or more children, but this percentage increased to 33.8 percent in 1984 and 45.4 percent in 1988 (Zhong 1984; Zhong and Liu 1985; Ye 1988). One explanation may be that the characteristics of the respondents differed in these studies. The respondents in the survey in 1983 were restricted only to women, while the 1984 and 1988 surveys included farmers and rich farmers. Therefore, the respondents in the later surveys may have been able to support more children. Thus, such figures showing an increasing pattern in terms of favoring three and more children in 1984 and 1988 may be a result of the differences in survey design. The studies of the Southeastern rural areas of China show that within these areas, the desired number of children in Zhejiang province is lower than in any of the other provinces. The average desired number of children was only 1.4 in 1984 (Chen and Luo 1985; Chinese Rural Family Investigation Group 1993; Shaji 1995; Zhong and Huang 1996). This low number may also be due to survey design. This study included only single women, which may influence their attitudes toward their desired number of children while the other studies included married males and females. In other words, marital status might have influenced the respondents’ attitudes toward their desired number of children. Nevertheless, examining the changing patterns of desired number of children among different provinces from 1986 to 1994, the research shows that with the exception of single people in 1984, the average desired number of children in Zhejiang province declined from 2.3 in 1986 to 1.9 in 1994 (Chinese Rural Family Investigation Group 1993; Zhou and Huang 1996). However, there is little variance in Shangdong province where the average desired number of children remained at 2.2 except for the middle-class farmers in the 1989 survey (Shaji 1995). Few studies were conducted in Fujian and Guangdong provinces. But the desired number of children in both provinces is relatively high which may be due to cultural influences that promote higher fertility. Based on the earlier analysis, all of the rural areas of China show a decrease in the desired number of children from the 1970s to the 1990s. The percentages of people who desire only 1 child have increased and the numbers who want 3 or more children have decreased. Having 2 children appears to be preferred by the majority of the rural residents in all of the rural areas. However, the patterns are not consistent across the different

Changing patterns of desired fertility

93

areas. For example, the mean number of desired children in the Beijing and Shanghai rural areas are lower than those of the other rural areas, and the desired numbers of children in Fujian and Guangdong provinces are higher than in any other areas. Urban areas in China are administrative entities and officially approved by the State Council or the provincial-level governments. Officially approved cities and towns, also known as “designated cities” (jianzhi shi) and “designated towns” (jianzhi zhen) respectively, are the two major components of the Chinese urban system. There exists a four-level urban system as follows: the provincial-level municipalities directly under the jurisdiction of the Central government (Zhi xia shi), the prefecture-level cities (diqu shi), the county-level cities (xian shi), and the towns (zhen). According to Whyte and Gu (1987), urban areas contained only around 20 percent of China’s total population in the 1980s. With less enforcement of China’s registration system (the hukou system) and the increased need for laborers in the urban areas, the percent of the urban population increased in the 1990s. Table 7.2 reports the changing pattern of desired number of children in urban areas from the late 1970s to the 1990s. The questions in the urban surveys are similar to those in the rural surveys. The respondents varied from predominantly single persons to married women, mothers, husbands and wives, husbands and wives who just got married, people who were married and had jobs, and urban citizens. According to Table 7.2, with the exception of the predominantly single people in 1979, a declining pattern of desired children can be seen during the past 20 years, especially in the large cities, like Beijing and Shanghai. For example, in Beijing, the mean number of desired children in the 1980s ranged from 1.7 to 1.9 (Gao and Gu 1984; Beijing Evening 1988; Feng 1990; Lin 1992), and from 1.1 to 1.7 in the 1990s (Wang 1994; Hao 1995; Zhao 1997; Feng 1998; Shao 1999). Also, in other urban areas during the six years from 1992 to 1998, the mean number of desired children decreased from 1.7 for husbands, and 1.7 for wives to an average of 1.5 for all citizens (Feng 1998). Moreover, the most preferred number of children also changed over time. Two children appears to be the most commonly stated preference in the 1980s, but in the 1990s, the preference of most people decreased to only 1 child. Respondents do not seem to prefer the traditional large family that included 3 or more children. For instance, in the 1980s, more than two thirds of respondents preferred 2 children (Gao and Gu 1984; Lin 1992), while in the 1990s, more than two thirds wanted to have only 1 child (Hao 1995; Zhao 1997). Granted, these surveys were conducted in different urban areas, but the general trend is found in most urban areas, for instance, the urban citizens who live in cities in Hubei province. The percentage desiring 2 children decreased from 76.6 to 55.1 percent in just 10 years (Feng 1990, 1999), and those wanting 1 child increased from 19.8 to 39.8 percent (Feng 1990, 1999). Regarding rural and urban comparisons, the preferred number of children is lower for urban citizens than for rural citizens. Since the 1980s, the average number of desired children is below 2; however, the overall average number of desired children for most of the rural areas is still above 2. For example, in rural Jilin province in 1988, the mean number of desired children was 2.2, while in urban Jilin, the average number was 1.8 (Lin 1992). This, along with the previous discussions, indicates that 2 children are almost always the predominant choice of rural Chinese citizens, but having 1 child seems to be the desired family size for urban citizens.

Fertility, family planning, and population policy in china

94

Table 7.2 Pattern of the desired number of children in urban areas of China Area

Year

Characteristics of the respondents

Desired number of children (%) 0

Beijing

1

2

3+

Mean

1979 Predominantly single

5.8 57.7 35.8

0.7

1.32

1987 Married women

— 19.9 66.3 13.8

1.94

1994 Women

13.2 48.9 34.4

2.3

1.25

Beijing, Haidian District

1992 Women who have kids

1.6 29.8 67.4

1.2

1.68

Shanghai

1981 Women

— 31.0 68.0

—

1.70

Jilin Province

1988 Married women

— 30.6 63.0

6.4

1.76

Five cities in Hubei Province

1988 Husbands and wives

— 19.8 76.6

3.6

1.84

Ten cities and provinces

1992 Husbands Wives

— —

— —

— —

1.71 1.65

Tianjin

1995 Husbands and wives who just got married

7.4 63.2 22.2

0.8

1.10

Ha-er-bin

1996 People who got married and had jobs

11.9 62.6 25.2

0.1

1.25

Fourteen cities

1998 Urban citizens

1.6 45.7 51.3

1.4

1.53

Cities in Hubei Province

1999 Urban citizens

1.4 39.8 55.1

2.7

1.58

— —

Accordingly, three basic conclusions may be drawn. First, in general, the number of desired children is declining in both rural and urban areas. The traditional attitude such as “the more children the better” appears to be more and more outdated. Second, although there is a general declining pattern, there is also some variation among the areas. For example, the desired number of children for Beijing rural citizens declined from 1.8 in 1979 to 1.5 in 1994 (Zhang 1982; Feng and Ma 1996), while in Northeastern areas, the decrease was from 2.5 in the mid 1980s to 1.6 in the mid 1990s (Chinese Rural Family Investigation Group 1993). The variation in the desired number of children for Shanghai, Shandong rural citizens, and the Western rural areas is relatively small. Thus, desired fertility varies among the different areas in China. Third, the changing patterns of desired number of children are different in rural and urban areas. Generally speaking, desired fertility is lower in urban areas than in rural areas. During the 1980s, the most preferred desired number of children for rural people was between 2 and 3 children. However, it was slightly below 2 for the urban citizens. But, in the 1990s, the desired number of children for rural citizens decreased to 2 children, and most urban citizens preferred only

Changing patterns of desired fertility

95

1 child. These figures show the different declining patterns of desired fertility for rural and urban citizens.

Changing patterns of the desired gender and gender composition of children The analysis in the previous section explored the ideal number of children without considering the gender of the offspring. It is often the case that the desired number of children is based on preferred gender or the gender composition of previously born children. The desire for an additional child is often contingent upon the gender of the existing child or children. Thus, this part of the chapter will focus on the changing patterns of the desired gender of children and the desire for additional children based on the gender of the existing child or children. Rural and urban areas will also be differentiated here. Unfortunately, only a few studies in China have asked about the desired gender of the children, but those studies that have been conducted tell an interesting story. Table 7.3 presents the results from studies conducted in rural areas. The respondents varied from single persons to pregnant women, married women, married men, fanners, rich farmers, husbands, and wives. The results show the percentages of the respondents who prefer 1 girl, 1 boy, 1 child either sex, 2 girls, 2 boys, or 2 children of either gender. Several conclusions may be drawn. First, among those interviewed, the overwhelming preference is for 2 children, but specifically, 1 son and 1 daughter. These results are consistent with the average number of desired children shown in the previous discussion and may be one explanation for the higher desired fertility in rural areas. Most people who have had 2 boys or 2 girls (as well as those with only 1 child) will be less than satisfied with their family and desire to continue with their childbearing until their desired gender outcome is reached,

Table 7.3 Patterns of the desired gender of children in rural areas of China Year

Area

Characteristics of the respondents

Desired gender of children (%) 1 1 1 2 2 1 son 2 Others son daughter either sons daughters and 1 either sex daughter sex

1979 Suburban Predominantly boundary single males of Beijing

8.7

1.5

17.9

—

—

54.6

—

17.3

Predominantly single females

7.3

1.5

16.9

—

—

50.6

—

23.7

Single

4.3

1.8

14.6

—

—

65.9

—

13.4

Rural area of Sichuan Province

Fertility, family planning, and population policy in china

1983 Chu 15–49 women County of Anhui Province

2.5

0.2

1984 Yangzhou Married males of Jiangsu Married females Province

6.0 4.1

1.0 4.6

36.0 35.0

10.2

2.5

—

—

1984 Zhejiang Province

Single

1984 West of Hubei Province

Rich farmers

1986 Sha Farmers County of Sichuan Province 1988 Jilin Province

68.0

—

16.6

2.0 1.0

1.5 2.6

47.0 43.3

6.5 7.2

— 2.2

52.4

2.3

0.5

27.4

1.4

3.3

—

—

—

—

70.3

—

29.7

—

—

0.6

0.3

67.2

31.9

—

86.3% of them intended to have male children

1993 Gan-ning- Farmers qing etc. poor counties 1994 Urban Husbands boundary Wives of Beijing 1999 Rural Hubei Province

0.6

Married women

Farmers

— 12.1

96

More than 85% of them intended to have male children

25.4 17.1

6.9 9.6

18.1 24.4

1.0 0.8

0.6 0.4

35.7 35.2

6.0 6.7

7.3 5.8

7.6

1.2

17.0

3.8

0.6

64.2

2.4

3.2

resulting in a higher desired number of children. Conversely, of those who expected to have only 1 child, a high percentage had no gender preference. This pattern is particularly found in Jiangsu and Zhejiang provinces (Zhang 1982; Zhong 1984, Zhong and Liu 1985; Chen 1985; Li 1987; Lin 1992; Jin 1995; Feng 1996, 1999). Consistent with the son preference literature, rural residents prefer sons to daughters in most areas regardless of the number of children they desire. In Jilin Province in 1988, 86.3 percent of the respondents preferred sons; and more than 85 percent of the respondents in Gan-ningqing’s poor counties in 1993 wanted at least 1 son (Sun 1988; Zhou and Huang 1996). Thus, we can conclude that in rural areas, people have strong son preferences in terms of their desired fertility. Stratifying based on the characteristics of the respondent further shows that son preference is stronger for males than for females. For example, in 1994, 25.4 percent of male respondents wanted to have a son, while only 6.9 percent wanted to have a daughter. In contrast, 17.1 percent of the female respondents wanted a son and 9.6

Changing patterns of desired fertility

97

percent wanted a daughter (Feng 1996). This shows that the gender of the parent accounts for some of the differences in the desired gender of offspring. Table 7.4 presents the respondents’ gender preference for their next child taking into account the number and gender of their existing children. Certain gender expectation patterns emerge. First, the respondents who had 2 daughters have the strongest desire for their next child to be a son. For example, data from a survey in 1987 in Jilin Province shows that all of the respondents (100 percent) who already had 2 daughters preferred their next child to be a son (Sun 1988). Following this pattern are respondents who had 1 daughter or 3 or more children. In fact Li (1987) found that in 1991, 96 percent of respondents of Luochuan County, a county of Shanxi Province, who had only 1 daughter desired that their next child be a male. Son preference does appear to decrease as the number of existing children decreases. Respondents who had no children were less likely than those who already had female children to prefer their next birth produce a son. According to the survey conducted in Gansu Province in 1987, 61 percent of people who had no children preferred their first child to be a son while only 1 percent preferred a daughter (Liu 1990). But for this group, respondents also showed the tendency of not caring about the gender of their first. For instance, in Jilin Province in 1987, 45.2 percent of the respondents said they had no preference about their firstborn’s gender (Sun 1988). This indicates that although the respondents who had no child still had some son preference, they began to show a tendency of caring less about their first child’s gender. In the previous discussion, most rural respondents said that 2 children was their preferred number, and many desired the gender composition to include 1 male and 1 female. Respondents with this gender combination become somewhat less likely to desire sons over daughters for any additional children that they may have. In fact, in Gansu Province, only 78 percent of respondents with a male/female combination of children desired their third child to be a son (Liu 1990); in Jilin Province only 50 percent had this desire (Sun 1988), and in Luochuan County of Shanxi Province, only 42 percent desired their third to be a son (Li 1987).

Table 7.4 Patterns of the desired gender of the next child and the number of existing children in rural China Year

Area

Gender Number of existing children (%) expectation None 1 1 1 son 2 2 More to the next son daughter and 1 sons daughters than child daughter 3

1987 Gansu Province

Male Female It doesn’t matter

1987 Jilin Province

Male Female It doesn’t

61 1 38

16 60 24

94 — 6

78 3 19

3 95

99 — 21

80 12 8

47.1 1.7 7.7 89.1 45.2 9.1

85.2 0.8 14.0

50.0 25.0 25.0

— 100 —

100 — —

100 — —

Fertility, family planning, and population policy in china

98

matter 1991 Luochuan County of Shanxi Province

Male Female It doesn’t matter

— — —

14 61 25

96 0 4

1996 North of Anhui Province

Male Female

— 86.7 — 13.3

77.7 22.3

42 14 34

5 68 27

96 0 4

— — —

84.6 81.4 15.4 18.6

90.5 9.5

85.2 14.8

Those least likely to prefer their next child to be sons and favored daughters appear to be those respondents who had 1 or 2 sons already. Daughter preference then actually emerges. In the 1987 surveys in Jilin and Gansu Provinces, virtually all of the respondents who had 2 sons wanted their next child to be a female (Sun 1988; Liu 1990). The percentages are slightly lower for 2-son parents in Luochuan County of Shanxi Province and Anhui Province, namely, 96 percent and 90.5 percent respectively (Graham et al. 1998). And of those respondents who already had 1 son, 60 percent in Gansu Province, 89.1 percent in Jilin Province, and 61 percent in Luochuan County desired that their second child be a female (Li 1987; Sun 1988; Liu 1990; Graham et al. 1998). This would appear to imply that although there is strong son preference in rural China, having at least 1 daughter in addition to 1 son is desired in order to have an ideal family. There are some differences in urban areas compared to rural areas with regard to the desired gender of children. The respondents in these surveys varied from single persons to married women, women who have children, married people who are in the labor force, and urban citizens. Table 7.5 reports the desired gender of children in the surveyed urban areas. As discussed earlier, 1 child is the preferred number of children for urban respondents. Gender preference does not seem to matter for the majority. In other words, preferring 1 child, either male or female, is the overwhelming pattern for most urban citizens. And the 1 child of either gender preference pattern is particularly popular among the urban single respondents of both genders. In the 1984 survey in Zhejiang Province, 60.5 percent of the single respondents said they wanted 1 child and 52.4 percent of all the respondents (86.6 percent of the single respondents) stated they did not care about the gender of their single child (Chen and Lou 1985). In Beijing, 43.5 percent of single males and 51.8 percent of single females desired only 1 child and did not have a gender preference (Wang 1994). But, in Jilin Province, 69.9 percent of married women still preferred male children (Lin 1992). Therefore, it appears that son preference has become somewhat weakened in urban China, and some daughter preference has even emerged. For instance, in the 1994 survey in Beijing, the female respondents stated they intended to have female rather than male children, and the desired sex ratio in Har-er-bin city was even below 99.8 which means that males are only desired 99.8 percent of the time (Wang 1994; Zhao 1997). This indicates that son preference has recently been challenged. While most urban citizens expect to have only 1 child, of those who desire 2, the most common gender preference is 1 male and 1 female. In surveys of Hubei Province cities and 14 other cities, the predominant desire was for 1 male and 1 female child, but when desiring only 1 child, the gender did not seem to make a difference (Feng 1998, 1999).

Changing patterns of desired fertility

99

Regarding similarities and differences among respondents from rural and urban areas, it seems that when only one child is expected, there are higher percentages of people in both rural and urban areas who did not seem to care about their single child’s gender.

Table 7.5 Patterns of the desired gender of children in urban areas of China Year

Area

Charac teristics respondents

Desired gender of children (%) 1 1 1 2 2 dau 1 son 2 Others son dau either sons ghters and 1 either ghter sex daughter sex

1979 Urban Boundary of Beijing

Predominantly 8.2 single males 8.6 Predominantly single females

1984 Zhejiang Province

Single

1988 Jilin Province

Married women

1992 Beijing Haidian District

Women who had children

1994 Beijing

Urban women

8.1

3.1 3.2

43.5 51.8

— —

— —

26.5 19.1

— —

18.7 17.3

0

52.4

2.2

0

30.9

0.2

6.2

69.9% of them preferred male children There is no apparent gender preference, most of them preferred one male and one female child To have male children was not the main motivation. There are more women intended to have female children rather than male children

1996 Ha-er-bin Married people who were in labor force

The desired sex ratio was 99.86, which is much lower than the real sex ratio 137.28

1998 14 cities

Urban citizens

9.1

10.5

26.1

0.7

2.0

47.9

—

3.7

1999 Hubei cities

Urban citizens

9.2

2.4

35.6

0.7

0.3

46.1

—

5.7

Fertility, family planning, and population policy in china

100

This pattern indicates that there is a certain relationship between the desired number and the desired gender of children; as esired family size decreases, son preference also decreases. But, for those respondents who desired 2 children, 1 son and 1 daughter seem to be the predominant preference in both rural and urban areas of China. Although there still is son preference in both rural and urban areas, having 1 son and 1 daughter seems to be the desire of most rural respondents, while having only 1 child with no gender preference seems to have become the preference of most urban citizen. With regard to the rural areas, in the 1980s preferring 1 son and 1 daughter was the predominant model. In the 1990s, more people turned to desiring 1 child and not caring about the gender of the single child. In urban areas, people showed a son preference pattern in the 1980s (Lin 1992). In the 1990s, some urban respondents began to actually demonstrate a preference for daughters (Zhao 1997). Thus, like the desired number of children, the desired gender of children also changed over time and showed different patterns in rural and urban areas. Based on the given results of the meta-analysis, several conclusions may be presented in terms of the desired gender of children. First, people typically have a preference about the gender of their future children, and it is mainly exhibited as son preference. Second, over the past two decades, the pattern of desired gender has been changing rather than remaining traditional and stagnant.

Discussion and conclusions The data presented and reviewed in this chapter are from diverse areas with differing methodologies. But several conclusions emerge. First, there has been a declining pattern in the desired number of children from the 1970s to the 1990s in both rural and urban areas of China. A family that consisted of 2 children was preferred for rural and urban citizens in the 1980s, but beginning in the 1990s, more urban people began to state a preference for 1 child. Generally speaking, urban citizens have lower desired number of children compared to their rural counterparts. Beijing and Shanghai residents prefer to have small numbers of children, and Fujian and Guangdong residents desire larger numbers of children compared to other areas. Second, son preference exists, but seems to be declining. According to Arnold and Liu (1986), son preference in China is the result of deeply rooted Confucian traditions, and sons are desired for the purpose of “family propagation, old-age security, the provision of labor, and the performance of ancestral rites” (Arnold and Liu 1986:222). Daughters do not pass on the family line. This is a most important reason for the strong son preference. In addition, many Chinese still hold to the old Chinese belief that “many sons [bring] much happiness” (Greenhalgh and Li 1993:10). Also Gu has noted that Chinese women from both urban and rural areas believe they will be at risk of discrimination and insult if they do not produce a son (Gu 1994). But, in some urban areas, daughters are becoming more desired, and most residents who expect only 1 child have little or no gender preference. And those who expect to have 2 children seem to desire 1 son and 1 daughter rather than only sons regardless of whether they are from urban or rural areas. There are several explanations for the declining desired fertility pattern. Admittedly, socioeconomic change, a reason frequently cited in explaining the drop in fertility, has

Changing patterns of desired fertility

101

played an important part. But this seems inadequate in explaining desired fertility changes in China. This is particularly so in areas with high socioeconomic levels with high desired fertility. This is the situation in Fujian and Guangdong provinces. Fujian and Guangdong provinces are two of the most economically developed areas of China. However, the desired number of children in these two provinces is also higher than in other areas. This suggests that the effect of the economy on people’s traditional desired fertility attitudes may not be felt right away. In other words, desired fertility may lag behind the rapid progress of the local economy. Other factors are also influencing the changes in the desired number of children. Strong and effective government intervention may have played an important role in reducing the desired number of children. With the implementation of the one-child policy, the nuclear family is gradually becoming the ideal family type, especially in urban China. Caldwell’s (1982) wealth flows theory states that in premodern societies, the wealth flow is from children to parents or, more broadly, from the younger to the older generation. Wealth is defined as money, goods, and resources. With the transition from the extended to the nuclear family, the pendulum swings, and the direction of the flow is from parents to children. In this situation, being childless would be the most rational economic behavior. Thus, people would begin to desire fewer children than before. From this point of view, the family planning policy in China has not only been a major factor for the reduction in actual fertility, but also for the reduction in desired fertility. The goal of the one-child policy is primarily to convince families that their interests can be best advanced by having small, rather than large, numbers of children, and particularly by having only one. From the data reviewed here it appears that the Chinese authorities have achieved some success in this direction. On the other hand, the underlying preference for the gender of children is still male. There are several reasons for this preference. The first reason can be referred to as the Chinese Confucian tradition. The Chinese Confucian tradition (especially in rural China) states that sons are the family’s descendants. As mentioned before, daughters are not allowed to carry the family lineage. Second is a family support consideration, especially in rural areas where there is a lack of social welfare and health insurance services. Sons are expected to be the main supporters of the families and the elders. And finally, the Chinese birth control policy can only influence the number of children per family. It cannot restrict people’s attitudes about the desired gender of their children. To some extent, it strengthens people’s preference of sons since most of the families can only have one child and due to tradition, males are preferred. Thus, the changing pattern of desired fertility occurs first, and is then followed by a change in the desired gender of children. Unlike the case in Western countries where fertility changes occurred in conjunction with the industrial revolution, fertility changes in China came about differently. China is still a socially and economically developing country. It did not experience a “natural change” in fertility through modernization, but a forced fertility decline due in large part to the birth control policy of the Chinese government. The fertility policies have expanded such that they have also influenced desired fertility, but they have not to date had much of an influence on the desired gender of children. The desired number of children shows a declining tendency, but son preference remains dominant in desired gender of children. Such a f fiding indicates that the birth

Fertility, family planning, and population policy in china

102

control policy in China has achieved some success, but some provisions must be included to lessen the desire for male children over female children.

References Arnold F. and Z.Liu. 1986. “Sex Preference, Fertility and Family Planning in China.” Population and Development Review, 12:221–246. Banister, J. 1984. “Population Policy and Trends in China, 1978–83.” The China Quarterly, 100:717–741. Beijing Evening. 1988. April 24:1. Bongaarts, J. 1990. “The Measurement of Wanted Fertility.” Population and Development Review, 16:487–506. Caldwell, J.C. 1982. Theory of Fertility Decline. New York: Academic Press. Chen, J. and B.Luo. 1985. “The Survey of the Fertility Motivation and the Fertility Need of the Unmarried Youth.” Population and Economy, 3:38–43 and 5:27–32. Chinese Rural Family Investigation Group. 1993. Rural Families in Contemporary China. China: Social Science Press. Coale, A.J. 1984. Rapid Population Change in China, 1952–1982. Washington, DC: National Academy Press. Coale, A.J. and S.Chen. 1987. Basic Data on Fertility in the Provinces of China. 1940–1982. Honolulu East-West Center. Feeney, G. 1994. “Fertility Decline in East Asia.” Science, New Series 226, (5190): 1518–1523. Feeney, G. and F.Wang. 1993. “Parity Progression and Birth Intervals in China: The Influence of Policy in Hastening Fertility Decline.” Population and Development Review, 19:61–101. Feng, X. 1990. “The Desired Fertility of the Parents of the Single-Child,” Population Studies (Renkou Yanjiu), 5:5–10. Feng, X. 1998. “Report of Elementary and Middle School Students’ Education Survey in Fourteen Cities of China.” Hubei Province: Hubei Foundation of Social Sciences. Feng, X. 1999. “Report of Household Survey in Hubei Rural and Urban Areas.” Hubei Province: Hubei Foundation of Social Sciences. Feng, L. and Y.Ma. 1996. “Marriage, Family, Fertility Investigation Reports of the Huilongguan Village in Beijing.” Population and Economy (Renkou yu Jingji), 3:55–62. Gao, E. and X.Gu. 1984. “The Fertility Rate Changes in Shanghai in the past 30 years and Its Influential Factors.” Population Studies (Renkou Yanjiu), 1:26–33. Graham, M.J., U.Larsen, and X.Xu. 1998. “Son Preference in Anhui Province, China.” International Family Planning Perspectives, 24:72–77. Greenhalgh, S. and J.Li. 1993. “Engendering Reproductive Practice in Peasant China: The Political Roots of the Rising Sex Ratios at Birth.” New York: The Population Council. Gu, B. 1994. “Synthesis of Findings of 8 Country Reports: Implications for Policy and Programs.” China: CPIRC. Gu, B. and S.Yang. 1991. “Fertility Trends in Rural China in the 1980s: Cohort Effect Versus Period Effect.” Asia Pacific Population Journal, 6(4):3–34. Hao, M. 1995. Chinese Women Newspaper, March 3. Jin, H.H. 1995. “The Reproductive Rights of Rural Women and Their Fertility Rates.” Chinese Population Sciences (ZhongGuo Renkou Kexue), 1:33–44. Lavely, W.R. 1984. “The Rural Chinese Fertility Transition: A Report from Shifang Xian, Sichuan.” Population Studies, 38:365–384. Lavely, W., J.Li, and J.H.Li. 2001. “Sex Preference for Children in a Meifu Li Community in Hainan, China.” Population Studies, 55:319–329.

Changing patterns of desired fertility

103

Li, B. 1990. “Levels, Trends and Determinants of Fertility in China: 1973–1987.” Asia-Pacific Population Journal, 5(2):3–19. Li, J. 1987. “The Investigation and Prediction of People’s Desired Fertility in Rural Shanghai.” Population Studies (Renkou Yanjiu), 6:44–45. Lin, F. 1992. “The Dramatically Changing Pattern of Women’s Desired Fertility.” Chinese Population Science, 3:36–42. Lightbourne, R. 1987. “Reproductive Preferences and Behavior.” pp. 838–861 in J.Cleland and C.Scott (eds), The World Fertility Survey. Oxford: Oxford University Press. Liu, G. 1990. “The Economic Status and the Women’s Desired Fertility.” Western Population (Xibei Renkou), 1:6–8. Peng, X.Z. 1989. “Major Determinants of China’s Fertility Transition.” The China Quarterly, 111:1–37. Poston, D.L., Jr. 1992. “Fertility Trends in China.” pp. 277–286 in D.L.Poston, Jr and D.Yaukey (eds), The Population of Modern China. New York: Plenum Press. Poston, D.L., Jr. 2000. “Social and Economic Development and the Fertility Transitions in Mainland China and Taiwan.” Population and Development Review, 26:(Supplement) 40–60. Poston, D.L., Jr. 2002. “Son Preference and Fertility in China.” Journal of Biosocial Science, 34:333–347. Poston, D.L., Jr and B.Gu. 1987. “Social Economic Development, Family Planning, and Fertility in China.” Demography, 24:531–551. Poston, D.L., Jr and J.Zhong. 1990. “Socioeconomic Structure and Fertility in China: A Countylevel Investigation.” Journal of Biosocial Science, 22:507–515. Pritchett, L.H. 1994. “The impact of population policies: Reply.” Population and Development Review, 20:621–630. Shaji, C. (ed.) 1995. Women’s Status in Contemporary China. Beijing: Peking University. Shao, X. 1999. “Comparative Study of the Desired Fertility In Rural and Urban Chinese Families.” Chinese Population Science (Zhongguo Renkou Kexue), 1:20–26. Sun, Y. 1988. “The Desired Fertility of Contemporary Women.” Population Bulletin (Renkou Xuekan), 6:1–6. Tien, H.Y. 1984. “Induced Fertility Transition: Impact of Population Planning and Socio-Economic Change in the People’s Republic of China.” Population Studies, 38:385–400. Wang, S. 1994. “The Changes of Beijing Women’s Fertility Attitudes.” Population and Economy (Renkou Yu Jingji), 1:42–46. Westoff, C.F. 1991. Reproductive Preferences: A Comparative View, Demographic and Health Surveys Comparative Studies No. 3. Columbia, MD: Institute for Resource Development/Macro Systems. Whyte, M.K. and S.Z.Gu. 1987. “Popular Response to China’s Fertility Transition.” Population and Development Review, 13:471–493. Wolf, A.P. 1986. “The Pre-eminent Role of Government Intervention in China’s Family Revolution.” Population and Development Review, 12:101–116. Wong, J. 2001. “China’s Sharply Declining Fertility: Implications for its Population Policy.” Issues and Studies, 37:68–86. Ye, Y.Z. 1988. “A Discussion of Rural Citizens’ Attitudes Towards Fertility.” Population Research (Renkou Yanjiu), 4:28–31. Zeng, Y. 1986. “Changes in Family Structure in China: A Simulation Study.” Population and Development Review, 12:675–703. Zeng, Y. 1996. “Is Fertility in China in 1991–92 Far Below Replacement Level?” Population Studies, 50:27–34. Zeng, Y, T.Ping, B.Ch.Gu, X.Yi, B.H.Li, and Y.P.Li. 1993. “Causes and Implications of the Recent Increase in the Reported Sex Ratio at Birth in China.” Population and Development Review, 19:283–302. Zhang, Z. (ed). 1982. The Desired Fertility of Chinese Youth. Sichuan: Sichuan Press.

Fertility, family planning, and population policy in china

104

Zhao, J. 1997. “The Desired Fertility of the Urban Citizens.” Population Studies (Renkou Yanjiu), 5:42–45. Zhong, S. 1984. “The Survey of the Rural Women’s Desired Fertility in Tu County of Anhui Province.” Population Bulletin (Renkou Xuekan), 3:32–36. Zhong, G. and Y.Liu. 1985. “The Desired Fertility of the Rich Farmers in the West Part of Hubei Province.” Population and Economy (Renkou yu Jingji), 4:18–21, 42. Zhou, C.H. and L.Huang. 1996. “The Comparative Analysis of Rural Women Who Are in Different Economic Status.” Population and Economy (Renkou yu Jingji), 3:49–52.

Part III Biological and social determinants of fertility

8 Age at menarche and the timing of the first birth Sherry L.McKibben All cultures have markers or rites of passage that indicate when members are ready to move from one phase to the next phase of privileges and responsibilities. These markers include chronological age, governmental policies, and religious doctrine or traditions. They are usually reinforced through socialization and norms. Privileges such as dating, marriage, and parenthood are granted to an individual as he or she passes from and through the socially constructed phases of childhood and adolescence, and into adulthood. These privileges are based on the perceived maturity of the individual and are often considered individual decisions. When to allow one’s child to begin dating, when to get married, when to start a family, and how many children are desirable are decisions left to the individual. Or are they? All societies have inputs into these decisions. Society deems a person “ready” for the next level of responsibility. The question then is what is the basis of these judgments, and how are they reinforced through socialization? The importance a society places on these markers may well affect future behaviors. For instance, many cultures believe that when a girl reaches menarche, she is transformed into a woman. This biological function, menarche, has traditionally been the basis of marking the passage from one phase in life to the next. Menarche signals the time when a female first becomes fecund and has the biological potential for motherhood. Marriage and/or motherhood allow her to become a full member of her social world. Ceremonies marking this phase of “womanhood” vary from culture to culture (Stattin and Magnusson 1990), but virtually all cultures have at one time or another used this marker as the timing of entry into womanhood. Upper class white families in western cultures have “Coming Out” parties. These are typically after the girl’s sixteenth birthday. Hispanic families have the Quinceañera (Serrato 2003), which is held after the girl’s fifteenth birthday. Historically, these “celebrations” coincided with a girl reaching menarche and signaled to all single men that the girl had become a “woman.” Most sociologists and demographers focus on social explanations of human behavior, while minimizing, if not discounting, biological factors. There are exceptions (see the work of Richard Udry 1979, 1988), but most demographic and sociological theories concentrate primarily on social explanations. With the exception of the proximate determinants paradigm, fertility theories such as demographic transition, wealth flows, political economy, household economy, and others, pay little if any attention to biological predictors. And the numerous other fertility theories, namely, ecological, feminist, and diffusion, also fail to include the one major event that must precede fertility, menarche.

Fertility, family planning, and population policy in china

108

This is an important oversight in fertility theories because menarche and how girls’ social worlds react to this biological function should have an influence on the timing of her future fertility behavior. With modernization, the average age at menarche has decreased. This is due largely to better nutrition and healthier lifestyles (Frisch 1988; Wahrenforf 1993). In the United States, as in other northern European countries, the average age at menarche has decreased by about 2 years in the past 100 years (Pollard 1994). This pattern is expected to be repeated in China as China continues to modernize. As the age at menarche decreases, the timing of future fertility behavior should also decrease. Decreasing age at menarche may also have larger social and economic consequences that lie outside the individual woman or even beyond her immediate social world. Therefore, this chapter investigates whether a biological factor such as age at menarche has an independent effect on the hazard of the first birth. The central goal is to model the likelihood of giving birth to a first child during the period of time beginning with the onset of menarche. The analysis is conducted for Han women and for high fertility minority women.

Literature Udry and Cliquet (1982) found that a girl’s age at first intercourse is correlated with her age at menarche. Using data from 6 different sources and 4 different countries, they found that cross-cultural differences in age at menarche and the amount of social controls did not negate the effect of menarche on age at first sexual intercourse, first marriage, or first birth. The mean ages at menarche varied from 12.6 years for US Whites to 14.3 years for Malaysian Chinese. Furthermore the strict religious practices of Muslims in Pakistan did not alter the effect of menarche on the ages at marriage and first birth. Their study found a mean age of about 8 years between menarche and first intercourse for women in Belgium and the United States. The authors concluded as follows: The differences in timing of reproductive events are due to some more or less universal social processes by which the time of onset of menstruation and the rate of development of physical sexual maturity are picked up as social information on the basis of which social processes differentially propel women with different biological timing into mating and reproductive activity. (Udry and Cliquet 1982:60) This research would seem to indicate that the transition to the adult responsibilities of marriage and motherhood are rooted in a biological process, menarche, which becomes a social marker and is deemed important in almost all cultures. While correlations between menarche and first intercourse, first marriage, and first birth were found, regression analysis was only conducted for the Malaysian Malays and Malaysian Chinese. In both instances there was reported a significant relationship between menarche and first birth. Since menarche seems to be related to first sexual intercourse, regardless of social

Age at menarche and the timing of the first birth

109

factors, the biological motivation contributing to the desire for sexual intercourse must be examined. The links between age at menarche and age at marriage and age at first birth have been shown in cross-cultural studies using United States, Belgium, and several Asian data-sets (see earlier discussion of Udry and Cliquet 1982). A direct relationship has been shown to exist between age at menarche and age at marriage. As the age at menarche increases, the age at marriage also increases and the age when a female gives birth to her first child also increases. Using longitudinal data collected since 1935, Sandler and associates (1984) extended this research to the United States. They noted significant relationships between age at menarche and both age at marriage and age at first birth. As age of menarche increases, age at marriage and age at first birth also increase. They also found a relationship between age at menarche and fertility, but it disappeared when controlling for other factors such as education and residency. Riley and colleagues (2001) examined the duration between age at menarche and age at marriage and first birth and found that after controlling for social variables, age at menarche had no effect on either marriage or first birth. These findings were derived from two sources of data that may be questionable. The Tremin Trust data were collected from three cohorts of only White women who were attending the University of Minnesota. The first cohort attended from 1935 to 1939, the second from 1961 to 1965, and the third from 1965 through 1980. Their birth years are from 1900 to 1950. These women are not representative of the population. The second data set only included women from the cohort born in 1900–1910. First, marriage patterns have evolved and changed and other research has utilized more current cohorts that are more representative of the population. And, since marriage patterns have changed, using marriage as the beginning of the risk period for a first birth may skew the results. Second, the duration to first birth is measured from first marriage. However, the risk period for a birth does not begin at marriage, but at menarche. Therefore, the risk of having a first birth may be biased. Also, controlling for the woman’s age could well negate most other age related variables because menarche is also age related. Therefore, their results may not be without problems. In endeavoring to explain teenage fertility changes, Manlove and colleagues (2000) used life-course analysis with US data from the 1995 Cycle U of the National Survey of Family Growth to examine the differences between three cohorts of women and their hazard of experiencing a first birth. The sample included 4,883 women. Age at menarche was shown to be significant for the first two cohorts in predicting the hazard of experiencing a first birth for sexually active teenagers. Unfortunately, there was little discussion about the effects of age at menarche. This study used menarche only as a control for the timing of first intercourse. The literature on age at menarche and the timing of the first birth tends to be limited and somewhat controversial. Nevertheless, menarche remains significant in most studies and should be entertained as an effect. We turn next to a brief discussion of female human reproduction in order to fully comprehend the complex mechanisms responsible for the two events that indicate the beginning and end of a woman’s fecund period, her menarche and menopause.

Fertility, family planning, and population policy in china

110

Reproduction Menarche and menopause are two distinct and seemingly unrelated events in a woman’s reproductive life. Menarche signals the time when a female becomes capable of reproduction. But changes have already been occurring in her body for some time prior to this “marker,” and a period of non-ovulatory menses occurs after the onset of menarche. Primary sex characteristics begin to develop in the girl when she is between 8 and 16 years of age. This is when she begins to realize an increase in muscle strength, body fat, the development of pubic hair, and the development of breasts (Golub 1983:31). These changes begin as a result of hormonal changes brought on by the initial activation of the gonadotropin-releasing hormone (GnRH) pulse regulator. The cause of the activation is not fully known which presents a problem when explaining the onset of menarche. Activation of the GnHR can be seen in the increase in luteinizing hormone (LH) secretion during sleep. It then follows a pulsating pattern. Activation seems to begin in the Central Nervous System (CNS) and is independent of the ovaries (Wood 1994:402). After the onset of menarche, a woman does not automatically ovulate each cycle. A period of subfecundity occurs (Wood 1994:401). Some studies indicate that women with an earlier menarche increase in ovarian function earlier than their late menarche counterparts (Ellison 2001). Ellison (2001) has reported that in three different cultures, increased levels of ovarian steroid production (necessary for ovulation) in young maturers are higher than late maturers and appears to be consistent over a woman’s lifetime. This means that those who reach menarche early produce more steroids than those who reach menarche later. Evidence from other studies indicates that the time from first ovulation may be shorter for women with a late menarche compared to those with an early menarche (Foster et al 1986; Wood 1994). This would mean that the ability to become pregnant following menarche would be sooner for women reaching menarche late rather than early. It also suggests that women with later menarche tend to catch up with women with early menarche in terms of fertility performance (Foster al. 1986). Another reason for this period of subfecundity may be evolutionary. A woman who is still maturing while in her early teens needs her energies for her own development. A fetus requires many of the reserves available for development. As a female matures, her biological capacity to reproduce increases as she matures into her twenties. It is then a cost/benefit trade-off between reproduction and her own survival. Postponing reproduction increases the woman’s odds of future reproduction and the survival of the fetus (Ellison 2001:226). But if the body is already mature because it has a later menarche, ovulation should begin sooner and the body should be more able to sustain pregnancy. Fetal loss is more likely to occur after 30 or 35 and in the teenage years (Wood 1994:250). Each pregnancy lengthens the birth interval and causes a reduction in fertility. Each one of these adds gestation days until the next fertile period. Also, the development of each dominant follicle takes approximately three cycles to become capable of fertilization; this is a loss of one complete cycle (Wood 1994:73, 244). So, a younger woman is more likely to experience fetal loss even if she is unaware she has conceived. This would add time to her next possible conception; thus the younger a woman is at menarche, the more likely she is to experience more fetal loss. And “early menarche may

Age at menarche and the timing of the first birth

111

predispose [women] toward a higher risk of fetal loss, at least during the first two or three pregnancies” (Wood 1994:252). There are many processes occurring at once that may well affect the timing of the first live birth. First, the time from menarche to ovulation appears to be shorter for women with later ages at menarche, leading to shorter subfecund periods and waiting times to conceive. Second, the mean menstrual cycle length may be shorter for women who reach menarche later, which would lead to more cycles over the lifetime and more opportunities to conceive. Third, women in their teen years are more likely to suffer fetal loss than women in their twenties leading to an increase in the number of days between conception and birth, adding days to the gestation of the fetus. So if a woman reaches menarche early, the time it takes to produce a live birth will increase as her gestational days increase. Fourth, in controlled fertility societies like China, women experience an increase in number of menstrual cycles, which could lead to an increased health risk. Early menarche women experience more cycles than later menarche women; therefore, the later the age when reaching menarche, the more likely the woman will be better fit physically to carry the fetus to term. Based on these reasons, we hypothesize that the older a female’s age of menarche, the shorter the duration between menarche and her first birth and the more likely she will experience a first birth.

Data and methods The data analyzed in this chapter are from the Sample Survey of Population and Reproductive Health (SSPRH) conducted in China in late 1997 (State Family Planning Commission of China 1998). (See the discussion of this data-set Chapter 2 by Wu and Walther.) The analysis in this chapter is restricted to the 11,818 currently married or evermarried women in the sample. The data are divided into two groups, one for ever-married Han women, consisting of 10,879 women, and one for women who belong to minority nationalities with average fertility higher than the Han, consisting of 936 women. We did not wish to keep the two groups together because most minorities have higher fertility than the Han owing in part to the different enforcement of the one-child policy (Poston 1993). Manchu and Korean women are excluded from this second analysis because their fertility rates are similar to the Han, and they are subjected to similar policy enforcement as the Han. Cox Proportional Hazard analysis is used to estimate the effect of menarche on a woman’s transition to her first birth (see Chapter 2 by Wu and Walther for more discussion of hazard analysis). Table 8.1 provides descriptive statistics for the Chinese Han women. Table 8.2 provides descriptive statistics for the Chinese minority women. The dependent variable consists of two components, the number of months from menarche to the first birth or to the survey date, and whether a woman has given birth to her first child. The duration from menarche to first birth or survey for Chinese Han

Fertility, family planning, and population policy in china

112

Table 8.1 Descriptive statistics of Chinese Han women for the hazard of a first birth, 1997 Variable Menarche to first birth

Mean

Standard deviation

Minimum

Maximum

94.92

42.32

6

434

0.95

0.22

0

1

Menarche

184.89

21.76

120

240

Education

6.73

4.46

0

18

Rural

0.76

0.43

0

1

Policy

0.76

0.46

0

1

Fecund

189.20

82.97

2

468

Age at first marriage

259.67

32.56

132

490

Born before 1961

0.42

0.49

0

1

Born between 1961 and 1970

0.43

0.50

0

1

Born after 1970

0.15

0.36

0

1

Ever given birth

Source: State Family Planning Commission of China, 1998. Note N=10,879.

Table 8.2 Descriptive statistics of Chinese minority women for the hazard of a first birth, 1997 Variable Menarche to first birth

Mean

Standard deviation

Minimum

Maximum

85.48

45.08

6

385

0.94

0.25

0

1

Menarche

185.55

21.56

132

240

Education

4.73

4.29

0

18

Rural

0.89

0.32

0

1

Policy

0.74

0.44

0

1

Fecund

183.60

86.94

6

433

Age at first marriage

247.61

36.60

139

389

Born before 1961

0.31

0.46

0

1

Born between 1961 and 1970

0.47

0.50

0

1

Born after 1970

0.22

0.41

0

1

Ever given birth

Source: State Family Planning Commission of China, 1998. Note

Age at menarche and the timing of the first birth

113

N=936.

women has a mean of 94.92 months (7.91 years) with a minimum of 6 months and a maximum of 434 months (36.17 years) with 95 percent giving birth. The Chinese minority women’s duration has a mean of 85.48 months (7.12 years) with a minimum of 6 months and a maximum of 385 months (32.08 years) with 94 percent giving birth. In Figures 8.1 and 8.2 are graphic illustrations of the numbers of women who do not survive having a birth from one month to the next. Figure 8.1 is the Kaplan-Meier survival estimates for the Chinese Han women. This estimates the number of Chinese Han women who succumb to the hazard of the first birth for

Figure 8.1 Kaplan–Meier survival estimates for the hazard of a first birth: Chinese Han women, 1997.

Fertility, family planning, and population policy in china

114

Figure 8.2 Kaplan–Meier survival estimates for the hazard of a first birth: Chinese minority women, 1997. each month. For instance, at 100 months (8.33 years), approximately 55 percent of the Han women have given birth to their first children. Figure 8.2 is the Kaplan-Meier Survival estimates for the Chinese minority women. At 100 months, approximately 60 percent have succumbed to the hazard of a first birth. The major independent variable, the woman’s age at menarche, for the Han women, has a mean of 184.89 months (15.41 years) and ranges from a low of 120 months (10 years) to a high of 240 months (20 years). Age at menarche for minority women has a mean of 185.55 months (15.46 years) and ranges from a low of 132 months (11 years) to a high of 240 months. Education is measured as the number of years of completed education a woman has as of the date of the survey. It has a range of 0–18. The Han women have a mean of 6.73 years of education; the mean for the Chinese minority women is 4.73 years. The rural variable controls for whether a woman lives in a rural or urban area (rural=1). Of the Han, 76 percent are rural residents, while 89 percent of the minority women are rural residents. Since the sample mainly comprises women who have not yet completed childbearing, it is important also to control for each woman’s exposure to the risk of childbearing. Thus included is the variable, fecund, which is calculated in months for each woman; it is the difference between her age at menarche and either, her age at sterilization, her age at menopause, or her age when the survey was conducted, that is, 1997, whichever is less, minus 8 months for each live birth. Among the sample of Han women, this covariate has a mean of 189.20 months (15.77 years) with a minimum of 2 months and a maximum of 468 months (39 years). Among the sample of minority women, fecund has a mean of 183.60 months (15.30 years) with a minimum of 6 months and a maximum of 433 months (36.08 years).

Age at menarche and the timing of the first birth

115

Another consideration which must be controlled is the effect on a woman’s childbearing of China’s one-child population policy. The one-child policy was first implemented in late 1979, and has been a major factor in determining how many children Chinese women are able to have (Poston and Yu 1986; Wolf 1986). It is assumed, other things being equal, that women whose fertility began after the policy was first initiated will be more conscious of the timing of their births. Thus a control variable, policy, is included, which is a dummy variable indicating whether the woman’s first birth occurred after 1980; it is scored one, if yes. The policy covariate is an imperfect measure of the effect of the policy on a woman’s fertility, but it is the best that can be done with the available SSPRH data. Of the Han women, 76 percent in the sample had their first birth after 1980 (Table 8.1) compared to 74 percent of the minority women (Table 8.2). The second additional covariate used is to assist “fecund” in controlling for the woman’s exposure to childbearing. Age at first marriage is measured in months and can be included in the models because the samples are restricted to ever-married women. This covariate has a mean of 259.67 months (21.64 years) with a minimum of 132 months (11 years) and a maximum of 490 months (40.83 years) for the Han women, and a mean of 247.61 (20.63 years) with a minimum of 139 months (11.58 years) and a maximum of 389 months (32.42 years) for the minority women. After controlling for the relevant social factors, consideration of the eras in which the women were socialized is needed. Culture is not stagnant and affects the individual’s behaviors. Because this sample includes women of varying ages, a rough control for societal changes is included. Each data-set was stratified into three cohorts of women, those born before 1961, those born from 1961 to 1970, and those born after 1970. Dramatic changes occurred in China during these eras. Women born in China before 1961 experienced the change to socialism after 1949. During the 1960s women growing up in China were experiencing the Cultural Revolution. After 1970, women in China have experienced the drastic reductions in fertility. Of the sample of Han women, 42 percent were born before 1961, 43 percent were born from 1961 to 1970, and 15 percent were born after 1970. Of the sample of minority women, 31 percent were born before 1961, 47 percent from 1961 to 1970, and 22 percent after 1970.

Results Tables 8.3 and 8.4 report the results of the hazard regressions. In the tables, the numbers in parentheses are the standard errors and the numbers in brackets are the hazard ratios. The hazard ratios are derived by exponentiating the hazard coefficients. This leads to a more intuitive interpretation. The hazard ratios are used in the discussions. The results will be discussed for the Chinese Han women, and then for the Chinese minority women. Chinese Han Table 8.3 presents results for the 10,879 Chinese Han women. In Model 1, only the biological variable of age at menarche is included. Menarche is significant and positive. It indicates that for every month increase in age at menarche, a Han woman’s hazard of

Fertility, family planning, and population policy in china

116

experiencing a first birth increases by 2.7 percent. The test of Pseudo R2 examines the degree of model fit, which shows the fit for Model 1 is far from perfect. In Model 2, two variables, years of education and rural residency, are added. All three variables are significant. It indicates that every month increase in age at menarche will increase the hazard of a first birth by 2.3 percent, holding education and rural residency constant. In Model 3, a control for the one-child policy is added. The policy variable is not significant while the other three variables remain significant In Model 4, the number of months a woman has been fecund is added. While this does not change the predictive value of age at menarche (the coefficient is the same as in Model 2 and Model 3), it does significantly increase the effect of education and decreases the effect of rural. Because childbearing is almost exclusively limited to married women, their age at first marriage is controlled in Model 5. After controlling for women’s age at first marriage, the effect of menarche increases; for every additional month in age at menarche, a woman’s hazard of experiencing a first birth increases by 3.8 percent, controlling for the effects of the other variables.

Table 8.3 Cox Proportional Hazard analysis of the hazard of a first birth: ever-married Han females, China, 1997 Independent variable Menarche

Education

Rural

Policy

Fecund

Marriage age

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

0.026***

0.023***

0.023***

0.023***

0.037***

0.036***

(0.001)

(0.001)

(0.001)

(0.001)

(0.001)

(0.001)

[1.027]

[1.023]

[1.023]

[1.023]

[1.038]

[1.036]

−0.032*** −0.032***

−0.042***

−0.006*

−0.001

—

—

—

—

—

(0.003)

(0.003)

(0.003)

(0.027)

(0.003)

[0.968]

[0.968]

[0.959]

[0.994]

[0.999]

0.500***

0.501***

0.257***

−0.067*

−0.056*

(0.027)

(0.027)

(0.028)

(0.027)

(0.028)

[1.649]

[1.650]

[1.294]

[0.935]

[0.945]

—

−0.007

−0.135***

0.239***

0.279***

(0.023)

(0.023)

(0.025)

(0.026)

[0.993]

[0.874]

[1.270]

[1.322]

—

−0.005***

—

—

—

−0.003*** −0.005***

(0.000)

(0.000)

(0.000)

[0.995]

[0.997]

[0.995]

—

−0.034*** −0.034***

Percent change

3.6

−0.1

−5.5

32.2

−0.5

−3.4

Age at menarche and the timing of the first birth

117

(0.000)

(0.000)

[0.967]

[0.966]

Born before 1961

—

—

—

—

— Reference

Born between

—

—

—

—

— −0.240***

1961 and 1970

−21.3

(0.026) [0.787]

Born after 1970

—

—

—

—

— −0.469***

−37.4

(0.040) [0.626] Pseudo R

2

Final log

0.171

0.0228

0.0228

0.0228

0.0764

0.0773

−84846.999 −84349.001 −84348.96 −83838.473 −79726.185 −79652.22

Note Numbers in parentheses are standard errors; numbers in brackets are hazard ratios *p<0.05; **p<0.01; ***p<0.001.

Table 8.4 Cox Proportional Hazard analysis of the hazard of a first birth: ever-married minority females, China, 1997 Indep endent variable Menarche

Education

Rural

Policy

Model 1

Model 2

Model 3

Model 4

Model 5

Model 6

0.016***

0.015***

0.015***

0.015***

0.025***

0.024***

(0.002)

(0.002)

(0.002)

(0.002)

(0.002)

(0.002)

[1.017]

[1.015]

[1.015]

[1.015]

[1.025]

[1.024]

—

−0.022*

−0.025**

−0.031***

−0.013

−0.012

(0.009)

(0.009)

(0.009)

(0.009)

(0.009)

[0.978]

[0.976]

[0.969]

[0.987]

[0.988]

0.223

0.209

0.112

−0.031

−0.044

(0.122)

(0.122)

(0.122)

(0.121)

(0.121)

[1.250]

[1.232]

[1.118]

[0.970]

[0.957]

—

0.217**

0.020

0.779***

0.830***

(0.083)

(0.086)

(0.102)

(0.105)

—

—

Percent change

2.4

−1.2

−4.3

129.3

Fertility, family planning, and population policy in china

Fecund

Marriage age

—

—

—

—

118

[1.243]

[1.020]

[2.177]

[2.293]

—

−0.004***

−0.003***

−0.004***

(0.001)

(0.001)

(0.001)

[0.996]

[0.997]

[0.996]

—

−0.026***

−0.026***

(0.001)

(0.001)

[0.975]

[0.974]

—

Born before 1961

—

—

—

—

—

Reference

Born between

—

—

—

—

—

−0.279*

1961 and 1970

−0.4

−2.6

−24.3

(0.109) [0.757]

Born after

—

—

—

—

—

1970

−0.445***

−35.9

(0.155) [0.641]

Pseudo R2 Final log

0.0112

0.0128

0.0135

0.0191

0.0756

0.0764

−5127.5062 −5118.8547 −5115.3396 −5086.2409 −4793.5055 −4789.2955

Note Number in parentheses are standard errors; numbers in brackets are Hazard Ratios *p<0.05*; **p<0.01; ***p<0.001.

Because not all women have been exposed to the same cultural influences and are of vastly different ages, three cohort variables are included in Model 6. Women born before 1961 are the reference group. Of importance in this chapter is the finding that menarche is still significant and positive after controlling for the effects of the social variables. Therefore, for Chinese Han women, an additional month in a woman’s age at menarche will increase her hazard of experiencing a first birth by almost 4 percent after controlling for the other variables. Chinese minority women Table 8.4 reports the results of six additional Hazard models for the 939 Chinese minority women. In Model 1, age at menarche is significant; for every month older a minority woman is at menarche, her hazard of a first birth will increase by 1.7 percent. This is the

Age at menarche and the timing of the first birth

119

only variable included in this model. Model 2 adds years of education and rural residency, and the menarche variable remains significant. In Model 3, the policy control is added, and this does not change the effect of menarche on the hazard of a first birth. In Model 4, the fecundity variable is added. Again, this does not change the value or significance of menarche. Model 5 adds the woman’s age at first marriage. The effect of menarche increases in that for every month increase in menarche, the hazard increases to 2.5 percent. Model 6 is the full model and controls for cohort differences. The variable of interest, menarche, maintains its significance; for every month increase in a Chinese minority woman’s age at menarche, her hazard of experiencing a first birth increases by 2.4 percent, holding the other independent variables constant. The last section of this chapter summarizes and discusses some of the implications of these findings.

Discussion Age at menarche is consistently significant and positive for both minority and Han Chinese women in predicting the hazard of a first birth. As a woman’s age at menarche increases, her hazard of experiencing a first birth also increases. The fact that age at menarche is not the most important variable in the models is not that central for this chapter. The important point is that it is significant after controlling for numerous social variables. For the Han women, as they get older when reaching menarche, their hazard of a first birth is increased by 3.6 percent and for the minority women, by 2.4 percent. This means that for every month older when reaching menarche, a Han woman has a 3.6 percent greater chance of giving birth to her first child and a minority woman has a 2.4 percent greater chance of having her first child. Biologically, women who reach menarche at a later age have a shorter period of subfecundity and are thus more likely to experience a first birth sooner after reaching menarche than those women with an early age at menarche. The findings reported here support this hypothesis. A woman’s chance of having a first birth and having it sooner increases. A woman who reaches menarche early is likely to waste many of her most viable follicles before she is ready to conceive, and her ovulation cycles will be spaced further apart. Therefore, if menarche is postponed, the woman’s chances of conception increase and she will experience her first birth soon after reaching menarche. Most demographic and sociological research does not include biological variables despite the fact that the two of the key dependent variables of demography, fertility and mortality, have obvious ties and linkages with biology (Poston 2000). This chapter has shown that a biological variable, age at menarche, has an important and statistically significant impact on fertility behavior, even after controlling for relevant social factors. The literature on age at menarche has clearly shown that increases in modernization in a society lead to decreases in women’s average age at menarche. In the United States, as with other northern European countries, the average age at menarche has decreased by about 2 years in the past 100 years (Pollard 1994). There is evidence that the decline has slowed and perhaps even stopped (Wood 1994:423). One may assume that as China

Fertility, family planning, and population policy in china

120

becomes more modernized, the mean age at menarche will also decrease. Some have argued that the secular decline is due largely to such features of modernization as better nutrition and healthier lifestyles (Frisch 1988; Wahrenforf 1993). Others place more importance on decreases in the “prevalences of infectious disease and decreased consanguinity” (Wood 1994:416). Ceremonies marking the transition to “adulthood,” such as the Quinceañeras and “Coming Out” parties, are part of the social structure. Since social structures influence fertility behavior, it stands to reason that the “social” behavior of fertility, namely a first birth, is rooted in a biological function. But, to date, the fertility-reduction effects of modernization have not been represented as including any biological causes. The decreasing age at menarche will lead to a lower chance of experiencing a first birth. In China, children are an important part of the social structure and childlessness is not desirable. Strict marriage laws that limit childbearing until the early twenties may well reduce fertility. More women will not be able to experience giving birth to their first child because they have reached menarche early and marry late. This leads to a widening of the duration between menarche and their first conception and a decreasing chance for a first birth.

References Ellison, P.T. 2001. On Fertile Ground. Cambridge, MA: Harvard University Press. Foster, A., A.K.M.Chowdhury, J.Menken, and S.L.Huffman. 1986. “Age at Menarche and Its Influence On Fertility.” Fertility Determinants Research Notes. New York: Population Council. Frisch, R.E. 1988. “Fatness and Fertility.” Scientific American, 253:70–77. Golub, S. 1983. Menarche: The Transition from Girl to Woman. Lexington, MA: Lexington Books. Manlove, J., E.Terry, L.Gitelson, A.R.Papillo, and S.Russell. 2000. “Explaining Demographic Trends in Teenage Fertility, 1980–1995.” Family Planning Perspectives, 32(4):166–175. Pollard, I. 1994. A Guide to Reproduction: Social Issues and Human Concerns. Cambridge: Cambridge University Press. Poston, D.L., Jr. 1993. “The Minority Nationalities of China” in Proceedings of the International Population Conference, Montreal 1993. Liege, Belgium: International Union for the Scientific Study of Population. Poston, D.L., Jr. 2000. “Conceptual and Methodological Advances and Challenges in Demography.” pp. 273–298 in Stella R.Quah and Arnaud Sales (eds), The International Handbook of Sociology. Thousand Oaks, CA: Sage Publications. Poston, D.L. Jr and M.Y.Yu. 1986. “The One Child Family: International Patterns and Their Implications for the People’s Republic of China.” Journal of Biosocial Science, 18:305–310. Riley, A.P., M.Weinstein, J.C.Ridley, J.Mormino, and T.Gorrindo. 2001. “Menarcheal Age and Subsequent Patterns of Family Formation.” Social Biology, 48(1):21–43. Sandler, D.P., A.J.Wilcox, and L.F.Horney. 1984. “Age at Menarche and Subsequent Reproductive Events.” American Journal of Epidemiology, 119(5):765–774. Serrato, A.M. 06/27/2003. “The Quinceañeras.” http://clnet.ucr.edu/research/folklore/%20quenceaneras State Family Planning Commission of China. 1998. Sample Survey of Population and Reproductive Health. Beijing: State Family Planning Commission of China. Stattin, H. and D.Magnusson. 1990. Pubertal Maturation in Female Development. Hillsdale, NJ: Lawrence Erlbaum Associates, Inc.

Age at menarche and the timing of the first birth

121

Udry, R.J. 1979. “Age at Menarche, at First Intercourse and at First Pregnancy.” Journal of Biosocial Science, 11:433–441. Udry, R.J. 1988. “Biological Predispositions and Social Control in Adolescent Sexual Behavior.” American Sociological Review, 53(5):709–722. Udry, R.J. and J.O.G.Billy. 1987. “Initiation of Coitus in Early Adolescence.” American Sociological Review, 52(6):841–855. Udry, R.J. and R.L.Cliquet. 1982. “A Cross-Cultural Examination of the Relationship Between Ages at Menarche, Marriage, and First Birth.” Demography, 19(1):53–63. Udry, R.J., L.M.Talbert, and N.M.Morris. 1986 “Biosocial Foundations for Adolescent Female Sexuality.” Demography, 23:217–230. Wahrenforf, J. 1993. “Dietary Fat and Sports Activity As Determinants for Age at Menarche” American Journal of Epidemiology, 138(4):217–224. Wolf, A.P. 1986. “The Preeminent Role of Government Intervention in China’s Family Revolution” Population and Development Review, 12:101–116. Wood, J.W. 1994. Dynamics of Human Reproduction: Biology, Biometry, Demography. New York: Aldine De. Gruyter.

9 The effect of floating migration on fertility Xiuhong You and Dudley L.Poston, Jr

Introduction On the eve of the 1990 Chinese New Year, the term “Child-bearing Guerilla” first came into existence in China. In a comedy sketch on the national TV channel which was viewed by hundreds of millions of people all over China, two famous Chinese entertainers, Huang Hong and Song Dandan, portrayed a rural couple who in their 4 years of marriage had produced 3 children, all girls. But they wanted to continue childbearing until they could have a boy. To escape the controls of local family planning workers, they were forced to continually move around the country, just like guerilla fighters who move to hide themselves from the enemy. The image of floating migrants as a high fertility population was thus impressed on the minds of many Chinese. Floating migrants in China are a unique group of temporary (or irregular) migrants (Jordan and Duvell 2002). They result from the household registration system initiated in China in the 1950s and the reforms of the economic and political systems of the 1980s. Under the household registration system, every citizen in China is registered as either agricultural or nonagricultural, and this is determined by the type and the official residence of the person’s household at birth (Yang 1993). People with agricultural registration usually live in the rural areas and make their living off the land. In contrast, people with nonagricultural registration usually work for the government or state-run enterprises. Most of them live in urban areas and account for a small proportion of the population. Nonagricultural citizens receive stable salaries from the government, almost never lose their jobs, and enjoy many benefits and services provided by the government. Not surprisingly, these advantages motivate agriculturally registered persons to change their status and move to urban areas (Harbison 1981). However, until very recently, the registration system has been strictly enforced to control changes in residence across administrative boundaries, especially for rural-tourban moves. As a result, household registration has become an ascribed status, which has an important impact on migration (Yang 1993), fertility (Cooney and Li 1994), and other behaviors. Therefore, in China two groups of migrants have emerged: the permanent migrants, who migrate and officially change their household registrations; and the floating migrants, who move without officially changing their household registrations. Although the government does not prohibit floating migration, floating migrants are often denied job opportunities, education, and access to many social services (Gaetano and Jacka 2004). After the economic and political reforms that began in 1978, a market economy developed which opened up non-governmental job opportunities and made the food rationing system unnecessary.

The effect of floating migration on fertility

123

These changes provide opportunities for floating migrants because the newly developed urban free markets tended to weaken the role of the household registration system in controlling migration (A.Goldstein et al. 1991). Government provision is no longer the only source of food and services in urban areas. People without a household registration at the current place of residence are now able to find jobs, buy food in the market, and obtain services. They can even send their children to private schools. However, the relaxing of the household registration system makes enforcement of the one-child fertility policy more difficult because it creates a “loophole” in family planning administration (Yang 2000). Most demographic studies report that floating migrants move to pursue economic benefits. Studies conducted by A.Goldstein and colleagues (1997) and Yang (2000) in Hubei Province do not conclude that floating migrants have higher fertility than other groups after controlling for social, demographic, and economic factors. But the belief of many people in China is that floating migrants indeed move to escape family planning controls in order to have more children. The comedy sketch mentioned at the beginning of this chapter is a reflection of this belief. To date there has not been research conducted on this topic using data for the whole country.

Fertility and migration Studies of migrant fertility have been conducted since before the Second World War, many of which were conducted in the United States and other developed countries. Analyses of rural-to-urban migration and fertility have also been conducted in developing countries that have undergone rapid industrialization and urbanization (Zarate and Unger de Zarate 1975). Yet despite the large amount of literature on this topic, there is but little consensus about the association between migration and fertility (Zarate and Unger de Zarate 1975; Goldstein and Goldstein 1981). Reasons include methodological differences (Zarate and Unger de Zarate 1975) and the different contexts in which migration takes place (Long 1970). The theories dealing with fertility and migration fall into four major categories. Each will be reviewed. Selectivity hypothesis The selectivity hypothesis argues that the fertility differentials of migrants from nonmigrants are not caused by the migration itself but are due to the fact that migrants are usually a very select group (Bacal 1988). Research shows that migrants and nonmigrants are different on many characteristics such as age, education, marital status, and occupation, all of which may affect fertility either directly or indirectly. This suggests that once the relevant social demographic factors are controlled, fertility differences between migrants and nonmigrants should diminish. Socialization hypothesis The socialization hypothesis asserts that fertility preferences are the result of socialization during childhood (Goldstein and Goldstein 1981; Lee and Farber 1984). Thus, migrants

Fertility, family planning, and population policy in china

124

will tend to have fertility attitudes and behaviors similar to those of the people in the places where they grew up. The acculturation and assimilation processes at the place of destination will take place gradually, and migrants’ fertility levels and other norms should converge to those of the destination only after at least a generation has elapsed. Disruption hypothesis The disruption hypothesis argues that migration has a disruptive effect on fertility and marriage, due to the disruptive factors associated with the migration process, including spousal separation (Millman and Potter 1984) and passive fecundity owing to stress typically associated with moving to a new place (Hervitz 1985). Stephen and Bean (1992) point out that the disruptive effects in reducing fertility are particularly salient in high fertility populations and tend to be age related, in that migrants tend to be younger than nonmigrants. Because of the nature of the effect, the fertility of the migrants is expected to be lower than nonmigrants and for only a short period of time, right before or after the spatial movement. Adaptation hypothesis The adaptation hypothesis assumes that migrants will adapt to the new norms and environment of their destination after they arrive (S.Goldstein 1978; Lee and Farber 1984; Campbell 1989). The convergence is expected to take place in less than ten years (Hervitz 1985), a time interval which is different from the socialization hypothesis. Although the public believe that floating migrants will tend to have higher fertility than nonmigrants, none of the above hypotheses suggest that migrants moving from rural areas to lower fertility urban areas should have higher fertility than their rural counterparts. However, the socialization hypothesis does suggest that a reduction in migrant fertility may take as long as one generation after the migration. Such an inconsistency between the public belief and theoretical perspectives provides further motivation to examine the relationship between floating migration and fertility.

Floating migration and fertility in China Previous studies in Shanghai (A.Goldstein et al. 1991), Hubei (A.Goldstein et al. 1997; Yang 2000), and other places in China have distinguished migrants as “temporary migrants” and “permanent migrants.” Using longitudinal data in Hubei province, Alice Goldstein and colleagues (1997) noted that after controlling for social demographic variables, temporary migrants did not differ much from nonmigrants in fertility behavior and “if anything, migration tended to lower the propensity to have a child” (A.Goldstein et al. 1997:488). Three years later, Yang (2000) used the same data and argued for a detachment hypothesis. He emphasized the “weakening of social and normative controls over fertility due to the temporary migrants’ detachment from their usual social context at the place of origin and their lack of strong attachments to the place of destination” (Yang 2000:166). Under this hypothesis, he found different results from those of Goldstein and colleagues (1997). His results show that temporary migrants have higher probabilities of

The effect of floating migration on fertility

125

having a second birth than permanent migrants and nonmigrants even after controlling for other factors. The conflicting conclusions, which may result from different study designs, require further study. Moreover, the results for one province may not be generalizable to the rest of the country (A.Goldstein et al. 1997). Therefore, in this chapter, the 1 percent sample data from the 1990 census of China are used to examine the fertility behavior of floating migrants in China. Floating migrants are “those living outside their places of household registration” (Sun 2000). There are two reasons for using the term “floating migrants.” First, the term “temporary migrants” refers to the length of time after migration. However, floating migrants are a unique group exhibiting different characteristics and fertility behavior from other migrant groups, namely, nonmigrants and permanent migrants; the length of time does not always reflect this distinction. Second, although it is true that most floating migrants in China are temporary, that is, they live at their place of destination for only a short period of time, for example, less than 5 years (Goldstein 1973), there are floating migrants who have resided at the place of destination for longer than 5 years. Furthermore, as Sun (2000) has noted, there is a tendency for the floating population to stay away from their places of residence for longer than a temporary period. Floating migrants have unique characteristics that need to be addressed. Like most migrants, floating migrants tend to be younger and relatively concentrated in the younger age groups (Sun 2000). They usually have less education than permanent migrants, and most of them originate in rural areas, where fertility is higher than in urban areas. Permanent migrants, on the other hand, often have urban backgrounds since “they are mostly cadres and skilled workers whose movement involves a job transfer approved by the government” (A.Goldstein et al. 1997:484). To study the fertility of floating migrants, the selectivity hypothesis should first be examined by controlling for relevant social, demographic, and economic factors, such as age and education. If the fertility difference still persists, further examination using the other three hypotheses may be undertaken. The socialization hypothesis holds that fertility norms are formed at the place of origin of the migrants, and that assimilation will take a long time. Under this hypothesis, floating migrants who move from rural to urban areas are expected to have fertility levels similar to their rural nonmigrant counterparts. The adaptation hypothesis, on the other hand, argues that migrants adapt to the norms at the place of destination shortly after their move. That is, the fertility of floating migrants should be similar to urban nonmigrants. The testing of the disruption effect of migration on fertility requires information on migrants’ moving time relative to the time they give birth to a child, which is not available in the census data used here. Therefore, this hypothesis cannot be tested directly. On the other hand, if the media is right in saying that floating migrants in China move to escape the one-child policy control, their fertility should be higher than that of rural and urban migrants, as well as that of permanent migrants.

Fertility, family planning, and population policy in china

126

Data and method The data used in this chapter are from the 1 percent sample of China’s 1990 census. More than 2.3 million ever-married women aged 15–49 are included. The census data provide information on the reported number of children ever born (CEB) to a woman, the dependent variable in this analysis. In the top row of Table 9.1 we show descriptive data for CEB for all 2,305,642 ever-married women in the sample, and then for the women broken out by migration status. The mean number of children ever born for all the women is 2.148. When the women are subdivided into their respective migrant categories, nonmigrants actually have higher CEB values than long-term floaters, followed by shortterm floaters, with permanent migrants showing the lowest CEB. The best and most ideal source of information about the extent to which the fertility of migrants is influenced by, or influences, migration would be longitudinal fertility data. Such data, however, are not available. With data on migration status and children ever born from the 1990 census of China, we are only able to appraise the relationship at the point of time in which the census was conducted. Some of the fertility for the women is most likely to have occurred prior to the migration. However, this is not that problematic an issue. In the famous Chinese television show about “child-bearing guerrillas” that motivated this analysis, all

Table 9.1 Social demographic variables by migration status: ever-married women in China, aged 15–49, 1990 Variable

All Non Long-term women migrants floating

Short-term floating

Permanent migrants

CEB (mean)

2.148

2.182

1.838

1.394

1.286

Age (mean)

33.3

33.5

33.2

28.5

28.9

Employed (%)

89.3

89.9

62.1

66.2

86.7

Rural (%)

79.5

79.9

71.2

86.6

57.9

Han (%)

92.7

92.7

94.0

92.5

92.6

2,305,642 2,204,882

14,809

35,891

50,060

(0.64)

(1.56)

(2.17)

Number of observations (%)

(100.0)

(95.63)

the childbearing of the floaters was not shown to occur after the migration, only some of their childbearing. Similarly, the migration-fertility theories reviewed above do not all posit that all the fertility needs to occur after the migration. Although longitudinal fertility and migration data would be preferred, most of the analyses of migration and fertility have used the kind of cross-sectional data we use here (e.g. see Rindfuss and Sweet 1977: chapter 8; Stephen and Bean 1992). As noted, we use data on the number of children ever born. We are aware that there may be an underreporting problem with fertility questions such as the CEB question in

The effect of floating migration on fertility

127

the 1990 census of China. This would occur mainly because of the one-child policy implemented in China in the late 1970s. However, this issue of possible undercount may not be as problematic an issue in this analysis. This is so because the major group being analyzed is the floating migrant group. By definition, they will be away from their home place at the time of the census and thus less under the control of the local officials. They should be less likely to underreport their CEB given their geographical location away from their home place. If the nonmigrants underreport their CEB, this will reduce the CEB difference between the migrants and nonmigrants, resulting in a more difficult test of our hypothesis. The children ever born variable is a “count” variable, that is, it is a “nonnegative integer-valued random variable” (Cameron and Trivedi 1998:1). Sometimes, count variables are treated statistically as if they are continuous and unbounded, and ordinary least squares (OLS) models are used to estimate the effects of the independent variables on their occurrence. However, OLS models for count outcomes sometimes produce biased, inefficient, and inconsistent estimates if the OLS regression assumptions are not met, particularly if average fertility is low (Poston 2002), as is the case for ever-married women in China in 1990. To get a better idea about this issue, we will examine the distribution of the CEB dependent variable. In Figure 9.1 we show the frequency distribution of the CEB variable (the solid line shown in the figure). The mean CEB value for all the women in the sample is 2.148 (Table 9.1). But with respect to the distribution for each count of CEB (from 0 to 16), we see the following: about 8 percent of the Chinese women have

Figure 9.1 Frequency distribution of CEB variable, and univariate Poisson distribution based on mean of 2.148. no children, followed by over 30 percent with 1 child, 29 percent with 2 children, and about 18 percent with 3 children. The percentages with more than 3 children drop off quickly to successively very low values, so that the CEB frequency distribution has a long positive tail. The distribution of the CEB variable in Figure 9.1 closely resembles a Poisson distribution. We have graphed in Figure 9.1, with a dotted line, the univariate Poisson distribution for a mean of 2.148, according to the univariate Poisson formula,

Fertility, family planning, and population policy in china

128

where Pr is the Poisson probability for each value of the count variable; µ is the mean of the univariate Poisson distribution; in the case of the univariate Poisson distribution shown in Figure 9.1, µ=2.148; and y is the value of the count variable, that is, the number of CEB, ranging from 0 to 18. The dependent variable of CEB has a distribution that is very close to a Poisson distribution; hence, we will estimate our regression models using Poisson regression rather than OLS. We discuss these issues in greater detail in the following section. We turn now to a discussion of the independent variables, the most important of which, for our purposes, is the woman’s migrant status. The major independent variable is migration status. The 1990 census did not ask directly about the respondents’ migration status, but answers to two related questions may be used. One question is “what is your current household registration?” with 5 answers: (1) Live in this city/county with household registration in this city/county; (2) Have lived in this city/county more than 1 year, with household registration in another city/county; (3) Have lived in this city/county less than 1 year, and have left the place of household registration for more than 1 year; (4) Live in this county and city with household registration yet to be decided; (5) Originally live in this city/county, currently working or studying abroad, temporarily without household registration. Because of the way this question was asked, Liang and White (1996, 1997) have noted that it fails to include floating migrants who have stayed at the place of destination less than 1 year. But there is no consensus established in the literature as to the amount of time after movement that should be used to define migrants and floating migrants. As Sun (2000) has pointed out, different explanations have been established for different purposes. Therefore, the use of the 1990 census data is appropriate to test the belief that floating migrants move to other places to escape the supervision of the one-child policy. As Alice Goldstein and colleagues (1997:483) have noted, commuters, those who “had not changed their place of residence, and spent only limited time in the nonresident location,” would not have stayed long enough at the place of destination to escape “the supervision of the family planning units at origin.” The other relevant question in the census is “What was your residence on July, 1, 1985?” (5 years before the census) with 3 answers: (1) This city/county; (2) Other city/county in this province; (3) Other provinces. The answers to the 2 questions may be receded to generate 2 variables. For the question of household nature, persons who marked option 5 (those who are working and studying abroad) are removed because the focus here is with internal migration. Also omitted are those who choose option 4 (household registration is yet to be decided) because the status of their household registration has not been set. Fortunately, removing these 2 categories of respondents only results in a loss of 3,901 observations, accounting for less than 0.2 percent of all the women aged 15–49. The household nature of those who choose option 1 is coded as “live in this city/county with household registration,” and that for those who choose options 2 or 3 is coded as “live in this city/county without household registration.” Thus there are 2 dichotomous variables for current household

The effect of floating migration on fertility

129

registration. For the question of respondents’ residence 5 years ago, the respondents are coded into 2 dichotomous categories. Those who choose option 1 are coded as “residence was in this county/city 5 years ago,” while those who chose options 2 or 3 are coded as “residence was in another county/city 5 years ago.” These 4 new variables, generate the primary explanatory variable, women’s migration status (see Figure 9.2), with 4 categories: (1) Nonmigrants (those who currently “live in this city/county with household registration” and whose “residence was this county/city five years ago”); (2) Long-term floating migrants (those who “live in this city/county without household registration” and whose “residence was this county/city 5 years ago”); (3) Short-term floating migrants (those who “live in this city/county without household registration” and whose “residence was another county/city 5 years ago”; and (4) Permanent migrants (those who “live in this city/county with household registration” and whose “residence was another county/city 5 years ago.”

Figure 9.2 Generation of the migration status variable. Note HR refers to Household Registration (Hukou). To farther specify the origin and the destination of the migrants, it is assumed that the place of household registration is the migrant’s origin, while the place where she completed the census form is her place of destination. Under this assumption, only the destinations of permanent migrants may be determined because permanent migrants are defined as people who move and have their household registration transferred to the destination. However, both the origin and destination of floating migrants may be determined by cross-referencing the place where they completed the census and their household registration. A migrant is of rural origin if she is registered as agricultural, and of urban origin if nonagricultural. The census categorizes all places into “city,” “town,” and “village.” A migrant’s destination is defined as rural if the place she marked in the

Fertility, family planning, and population policy in china

130

census form is “village,” and as urban if it is “city” or “town.” This may be then crossreferenced with her household registration to determine her place of destination. Five control variables are used, namely, age, education, state of employment, household registration status, and ethnicity. As the majority of the floating migrants move from rural to urban areas, only long- and short-term floating migrants flowing in this direction are analyzed. Two sets of Poisson models are estimated. The first set includes rural nonmigrants, rural-to-urban short-term and long-term floating migrants, and permanent migrants who were living in rural areas at the time of the census. The second set includes urban nonmigrants and permanent migrants who were living in urban areas at the time of census. Each set contains 6 models, with model 1 including only the primary variable, women’s migration status, and the next 5 models each adding one additional control variable.

Description Table 9.1 (bottom row) shows that among the ever-married women aged 15–49 in the sample, most (95.6 percent) are nonmigrants. Half of the migrants are permanent (2.2 percent) and the other half are floaters, most (1.6 percent) of whom have lived at places of destination for less than 5 years, with a smaller percentage (0.6 percent) living at their places of destination for more than 5 years. Although this last migration category of longterm floating migrants is only 0.64 of 1 percent of all the ever-married women in the sample, it includes over 14,800 women. The 1990 China census also provides other information about the respondents that we use in the analyses as control variables. These are the respondent’s age, her education, her employment status, the rural/urban status of her household origin, and whether she is a member of the majority Han nationality. Woman’s age is a continuous variable, ranging from 15 to 49. Educational attainment is based on her answer to the educational attainment question, with 7 responses: illiterate; and 6 categories of the highest level of school completed, namely, primary school, junior high, senior high, vocational, associate college degree, and 4-year college degree. For each woman, we developed a set of

Table 9.2 Percentage distributions of education status, for nonmigrants and groups of migrants: ever-married Chinese women, aged 15–49, 1990 Education status

Nonmigrants

Long-term floating

Short-term floating

Permanent migrants

Illiterate

22.10

15.94

14.77

10.03

Primary School

40.02

32.07

35.54

30.73

Junior High

26.55

36.76

38.48

35.93

Senior High

8.49

12.65

9.38

12.24

Vocational

1.77

1.63

0.96

5.83

The effect of floating migration on fertility

131

Associates

0.80

0.74

0.63

2.66

College

0.27

0.22

0.25

2.57

Total

100

100

100

100

7 dummy variables (one for each of the 7 responses) to measure education. The woman’s current employment status is coded as a dummy variable, employed=1, and unemployed=0. Current residence is also a dummy variable, coded 1 if rural, 0 if urban. And ethnicity is a dummy variable, coded 1 if Han, 0 if non-Han. Tables 9.1 and 9.2 present descriptive information for these independent and control variables. Short-term floaters and permanent migrants are younger than nonmigrants and long-term floaters. Employment is higher for nonmigrants and permanent migrants compared to both classes of floaters. Permanent migrants have the lowest percentage of persons with rural origins. Most of the women in all the groups are majority Han (Table 9.1). And the nonmigrants, on average, have lower levels of education than long- and short-term, and permanent migrants. Permanent migrants have the highest levels of education (Table 9.2).

Results In this part of the chapter we present the results of a series of Poisson regressions gauging the effects of a woman’s migration status on her predicted number of children ever born, controlling for other social variables. In the Poisson regression model, the dependent variable is a nonnegative integer, namely, the number of children ever born. This variable is assumed to have a Poisson distribution with a conditional mean that may be predicted by the woman’s characteristics on various X variables, that is, her migrant status, age, education, employment status, current residence, and ethnicity, according to the following structural model: µi=exp(a+X1ib1+X2ib2+…+Xkibk) where µi is the predicted CEB for the ith woman; X1i through Xki are the independent variables for the ith woman; a is the Poisson intercept; and b1i through bki are the Poisson coefficients for each of the k independent variables. The Poisson regression model is a nonlinear model, predicting for each woman her number of children ever born, that is, the number of times that µ, the count variable, has occurred. The X variables are related to µ nonlinearly.

Table 9.3 Poisson regressions of number of CEB: ever-married Chinese women, aged 15–49, 1990 Model

1

2

3

4

5

6

Migration Nonmigrants

Ref.

Ref.

Ref.

Ref.

Ref.

Ref.

Fertility, family planning, and population policy in china

132

Long-term floating

−0.17

−0.14

−0.10

−0.12

−0.11

−0.11

Short-term floating

−0.45

−0.17

−0.16

−0.17

−0.19

−0.19

Permanent migrants

−0.53

−0.28

−0.19

−0.19

−0.12

−0.12

0.05

0.05

0.05

0.05

0.05

Ref.

Ref.

Ref.

Ref.

Primary School

−0.12

−0.12

−0.10

−0.09

Junior High

−0.34

−0.34

−0.22

−0.21

Senior High

−0.50

−0.50

−0.28

−0.27

Vocational

−0.70

−0.70

−0.36

−0.36

Associates

−0.89

−0.89

−0.54

−0.52

College

−0.93

−0.93

−0.58

−0.57

−0.04

−0.07

−0.07

Ref.

Ref.

Ref.

Rural

0.36

0.36

Non-rural

Ref.

Ref.

Age (continuous) Education Illiterate

Employment status Employed Not employed Place of origin

Ethnicity Han

−0.20

Non-Han

Ref.

Constant Pseudo R

2

0.78

−1.12

−0.75

−0.71

−1.14

−0.98

0.004

0.119

0.136

0.136

0.144

0.146

Note All Poisson coefficients are statistically significant at p<0.05.

Table 9.3 shows the result of a series of Poisson regressions for all women. Reported are the Poisson regression coefficients and, at the base of the table, the Pseudo R2 values. All the Poisson coefficients shown in Table 9.3 are significant at levels of 0.05 or less. The first Poisson regression includes only the 3 migrant status variables (nonmigrants are the reference category). The Poisson coefficients for these 3 categories are all negative indicating that all 3 categories of migrants have lower fertility than nonmigrants. The coefficients may be exponentiated to produce Poisson odds ratios, and they provide a better interpretation than the Poisson regression coefficients (Poston 2002). In Model 1, the Poisson odds ratio for long-term floaters is e−0.17=0.84. This indicates that the expected number of children ever born of long−term floaters is about 16 percent less (i.e.

The effect of floating migration on fertility

133

Poisson odds ratio minus 1 x’s 100) than that of nonmigrants (the reference category). Short-term floaters have around 36 percent less children ever born than nonmigrants, that is e−0.45=0.64. The second through the sixth Poisson regression models include increasingly more control variables. In Model 6, the final model containing the migrant status variables along with all the control variables, long-term and short-term floating migrants still report negative Poisson coefficients. Even after controlling for age, educational attainment, employment status, place of origin, and ethnicity, shortterm floaters have 17 percent less children than nonmigrants (e−0.19=0.83), and long-term floaters have 10 percent less (e−0.11=0.90). Floating migrants have less, not more, fertility than nonmigrants. The results in the table also indicate that the fertility of the permanent migrants is less than that of the nonmigrants; permanent migrants have 11 percent less children ever born than nonmigrants (e−0.12=0.89). So, short-term floaters not only have lower fertility than nonmigrants, they also have lower fertility than permanent migrants. There is no evidence in Table 9.3 that floating migrants are childbearing guerrillas. There is no evidence in the table to support the story line of the famous Chinese television show (discussed at the beginning of this chapter) that floating migrants move away from their areas of origin to escape the family planning officials so as to be able to have more children than they would have been able to have had they not migrated, and hence have more children than the nonmigrants they left in their areas of origin. In fact, the results in Table 9.3 are consistent with the selectivity and disruption hypotheses reviewed earlier. However, the Poisson regression results reported in Table 9.3 are, at best, imperfect tests because no distinctions are made with regard to the nonmigrants and the permanent migrants. For instance, the nonmigrants captured in Table 9.3 include nonmigrants that the floating migrants left behind in the origin areas, as well as the nonmigrants in the destination areas that received the floaters. The permanent migrants included in Table 9.3 are comprised of permanent migrants who migrated to both urban areas and rural areas. We need to better distinguish among the nonmigrants. And we must decide what to do with the permanent migrants. Also, the analysis reported in Table 9.3 includes all floating migrants irrespective of whether they migrated to urban or rural areas. Since we are mainly interested in rural-tourban floating migrants, we need to distinguish the directionality of their migration stream. Data are available in the 1990 China census on the geographical location (urban or rural) of the hukou (household registration) of the floating migrant. So it is possible to select only those floaters with hukous in rural areas who are now (in 1990) living in urban areas. These will be the stream of rural-to-urban migrants we will next analyze. In the early 1990s it is estimated that there were around 80–100 million floating migrants in China (Roberts 1997; Solinger 1999:17–18). Of the floating migrants enumerated in the 1 percent sample of China’s 1990 census (the data used here), more than half (53.6 percent) are rural-to-urban floating migrants; the remaining 46 percent of the floaters are urban-to-rural floaters (4.6%), rural-to-rural floaters (28.5%), and urbanto-urban floaters (13.3%). We restrict the next analyses only to rural-to-urban floating migrants.

Fertility, family planning, and population policy in china

134

Now we must deal with the nonmigrants and the permanent migrants. The nonmigrants may be differentiated according to whether their residence in 1990 was urban or rural. We separate them into two groups according to their residence in 1990, rural and urban. We estimate one Poisson regression model including the rural-to-urban floaters and the rural nonmigrants. This analysis will enable us to compare the fertility of the floaters with the fertility of the nonmigrants in the rural areas of origin of the floaters. The selectivity and disruption hypotheses reviewed earlier are concerned with comparing migrants with persons in the sending areas who do not migrate. The famous Chinese television show about the “childbearing guerrillas” also compared the fertility of the floaters with the fertility of those in the areas of origin. We also estimate a second model including the same rural to urban floaters, but in this model we include the urban nonmigrants. This analysis will enable us to compare the fertility of the floaters with the fertility of those in the destination areas that received them. The socialization and adaptation hypotheses are applicable in tests comparing the floaters with persons in the receiving areas. But what should be done with the permanent migrants? The permanent migrants are persons who migrated legally within China between 1985 and 1990 from origin to destination areas. But all we know about them is whether their residence in 1990 is urban or rural. We do not know about the urban-rural status of their residence in 1985. Since we are restricting the analyses to floaters who moved from rural to urban areas, we will separate the permanent migrants into those who in 1990 are living in urban areas and those who are living in rural areas. We will include the permanent urban migrants only in the second model, namely, the model comparing the rural-to-urban migrants with urban nonmigrants. It does not make conceptual sense to include permanent migrants who migrated to rural areas in either of the two tests. Table 9.4 reports the findings from several Poisson regressions estimating the number of CEB for rural-to-urban floaters and rural nonmigrants. These results permit us to compare the fertility of the floaters with the nonmigrants in the rural sending areas. All the Poisson coefficients shown in Table 9.4 are significant at levels of .05 or less. The first Poisson regression includes only the 2 migrant status variables (nonmigrants are the reference category). The Poisson coefficients for the 2 migrant categories are both negative indicating that both categories of floating migrants have lower fertility than rural nonmigrants. In Model 1, the Poisson odds ratio reported for long-term floaters is e−0.21=0.81. The predicted number of CEB of long-term rural to urban floaters is about 19 percent less than that of rural nonmigrants (the reference category). Short-term rural-tourban floaters have around 36 percent less CEB than rural nonmigrants, that is e−0.44=0.64. Thus the rural-to-urban floaters, both short-term and long-term, have significantly lower fertility than the rural non-migrants they left behind when moving to urban areas. The second through the fifth Poisson regression models include more and more independent variables as controls. The final model, Model 5, contains the migrant status variables along with all the control variables; long-term and short-term floating migrants still report negative Poisson coefficients. Even after controlling for age, educational attainment, employment status, and ethnicity, short-term floaters have 17 percent less children ever born than nonmigrants (e−0.19=0.83),

The effect of floating migration on fertility

135

Table 9.4 Poisson regressions of number of CEB: ever−married Chinese women, aged 15–49, 1990 (rural nonmigrants and rural−to−urban floating migrants) Model

1

2

3

4

5

Migration Non-migrants

Ref.

Ref.

Ref.

Ref.

Ref.

Long-term floating

−0.21

−0.20

−0.17

−0.18

−0.17

Short-term floating

−0.44

−0.21

−0.19

−0.19

−0.19

0.05

0.05

0.05

0.05

Ref.

Ref.

Ref.

Primary School

−0.10

−0.10

−0.09

Junior High

−0.21

−0.21

−0.20

Senior High

−0.24

−0.24

−0.23

Vocational

−0.64

−0.63

−0.62

Associates

−0.90

−0.90

−0.88

College

−1.14

−1.13

−1.12

−0.02

−0.02

Ref.

Ref.

Age (continuous) Education Illiterate

Employment Employed (=1) Not employed (=0) Ethnicity Han (=1)

−0.19

Non-Han (=0) Constant Pseudo R2

Ref. 0.84

−1.06

−0.84

−0.82

−0.66

0.0011

0.1276

0.1319

0.1319

0.1338

Note All Poisson coefficients are statistically significant at p<0.05.

and long-term floaters have 16 percent less (e−0.17=0.84). Rural-to-urban floating migrants have less, not more, fertility than the non-migrants they left behind in the rural areas. The results in Table 9.4 are as predicted by the selectivity and disruption hypotheses. These results are inconsistent with the theme of the famous Chinese television show that floating migrants are “child-bearing guerrillas.”

Fertility, family planning, and population policy in china

136

Finally, we estimate a series of Poisson regressions comparing the fertility of rural-tourban floaters with the urban nonmigrants; in these regressions we also include permanent migrants who are residing in urban areas in 1990. Two of the fertility— migration theories reviewed earlier, the socialization and adaptation hypotheses are concerned with comparing the migrants with the nonmigrants at the destination. They differ mainly according to the amount of time it takes the migrants to adjust to the lower fertility of the nonmigrants. Table 9.5 reports a series of Poisson regressions estimating the number of CEB for long-term and short-term rural-to-urban floating migrants, permanent migrants residing in urban areas, and nonmigrants residing in urban areas. In the first model only the three migrant status variables are included; in the second through fifth models, we add increasingly more of the control variables. Model 5 is the final model. This model contains the migrant status variables and all the control variables. The Poisson results in the fifth model indicate that long-term rural-to-urban floaters, short-term rural-to-urban floaters, and permanent urban migrants all have

Table 9.5 Poisson regressions of number of CEB: ever-married Chinese women, aged 15–49,1990 (urban nonmigrants vs. rural-to-urban floating migrants and urban permanent migrants) Model

1

2

3

4

5

Migration Non-migrants

Ref.

Ref.

Ref.

Ref.

Ref.

Long-term floating

0.22

0.30

0.16

0.11

0.11

Short-term floating

−0.01

0.32

0.17

0.12

0.12

Permanent migrants

−0.16

0.17

0.09

0.08

0.08

0.06

0.05

0.05

0.05

Ref.

Ref.

Ref.

Primary School

−0.15

−0.11

−0.11

Junior High

−0.35

−0.28

−0.28

Senior High

−0.44

−0.36

−0.35

Vocational

−0.48

−0.39

−0.39

Associates

−0.65

−0.56

−0.55

College

−0.73

−0.63

−0.63

−0.21

−0.21

Ref.

Ref.

Age (continuous) Education Illiterate

Employment Employed (=1) Not employed (=0)

The effect of floating migration on fertility

137

Ethnicity Han (=1)

−0.21

Non-Han (=0) Constant Pseudo R

2

Ref. 0.41

−1.78

−1.24

−1.08

−0.90

0.0016

0.1108

0.1222

0.1253

0.1267

higher fertility than urban nonmigrants in the receiving areas. The long-term and shortterm floaters have, respectively, 12 percent higher fertility (e0.11=1.12) and 13 percent higher fertility (e0.12=1.13) than urban nonmigrants. The permanent urban migrants also have higher fertility than the urban nonmigrants. The socialization hypothesis posits that it takes a generation or more for the fertility of the migrants to equal that of the nonmigrants; the adaptation hypothesis stipulates that a shorter interval of time must pass before the migrants adapt the norms of their new areas of residence. Hervitz (1985) has stated that about 10 years are needed before the fertility of the migrants may be expected to converge with that of the nonmigrants. The results in Table 9.5 indicate that the migrants (floaters and permanent) have higher fertility than the nonmigrants. But most of them have not resided for a sufficiently long enough time period to have adapted to the lower fertility norms. Hence the results in Table 9.5 are consistent with the expectations of the socialization hypothesis and to a lesser extent, the adaptation hypothesis.

Conclusion Using the 1 percent sample data from the 1990 China Census, we have explored in this chapter the relationship between fertility and migration in China. Our investigations do not provide evidence for the view that rural-to-urban floating migrants have higher fertility than the nonmigrants they left behind in the rural areas. There is no evidence in our data that they are the “child-bearing guerrillas” that have been portrayed in the popular media as persons who move in order to be able to bear more babies. Rather, after controlling for relevant demographic and socioeconomic factors, rural-to-urban floating migrants have lower fertility than the nonmigrants they left behind in the rural areas. Floating migrants can be distinguished by the length of time after their migration. We showed in this chapter that both long-term and short-term rural to urban migrants have lower fertility than that of the nonmigrants they left behind in the rural areas. These findings are very much consistent with the fertility-migration literature dealing with the selectivity and disruption hypotheses. We also compared the fertility of the rural-to-urban migrants with that of nonmigrants residing in the urban areas receiving the migrants. Although this comparison was not directly relevant for our addressing the issue of whether floating migrants are “childbearing guerrillas,” there is a body of migration theory comparing the fertility of migrants with that of nonmigrants in the receiving areas. This literature, namely, the socialization and adaptation hypotheses, expects that initially the fertility of the migrants will be higher than that of the nonmigrants, but after a generation or so (the socialization

Fertility, family planning, and population policy in china

138

hypothesis), or after 5 or 10 years (the adaptation hypothesis), the fertility of the migrants will converge with that of the nonmigrants. Our results showed that long-term and shortterm floating migrants, and permanent migrants, had higher fertility than the nonmigrants receiving them. These results are consistent with the socialization hypothesis, and to a limited extent, the adaptation hypothesis. Our analyses could be improved by extending the study to include an alternate measure of fertility. In addition to measuring fertility with data on CEB to a woman, a more recent fertility indicator might also be used. The 1990 census also provides information on whether a woman had a baby born in the past 18 months before the census. This measure has the advantage of only including very recent fertility. Our preliminary work using this alternate measure shows very similar results to those reported here. In China there are great differences in fertility and migration across the different regions and provinces of the country. Future work in this area would do well to use a multi-level framework in which the characteristics of the women, as well as the characteristics of the provinces in which they reside, would be used to predict the number of CEB. Incorporating aggregate effects would doubtless improve the overall adequacy of the models.

References Bacal, R.A. 1988. “Migration and Fertility in the Philippines: Hendershot’s Selectivity Model Revisited” Philippine Population Journal, 4:53–67. Cameron, A.C. and P.K.Trivedi. 1998. Regression Analysis of Count Data. Cambridge, UK: Cambridge University Press. Campbell, E.K. 1989. “A Note on the Fertility-migration Interrelationship: The Case of Men in Western Area, Sierra Leone.” Demography India, 18:103–114. Cooney, R.S. and J.Li. 1994. “Household Registration Type and Compliance with the ‘One Child’ Policy in China, 1979–1988.” Demography, 31:21–32. Gaetano, A.M. and T.Jacka, eds. 2004. On the Move: Women in Rural-to-Urban Migration in Contemporary China. New York: Columbia University Press. Goldstein, S. 1973. “Interrelations Between Migration and Fertility in Thailand.” Demography, 10:225–241. Goldstein, S. 1978. “Migration and Fertility in Thailand, 1960–1970.” Canadian Studies in Population, 5:167–180. Goldstein, S. and A.Goldstein, A. 1981. “The Impact of Migration on Fertility: An ‘Own Children’ Analysis for Thailand.” Population Studies, 35:265–284. Goldstein, A., M.White, and S.Goldstein. 1997. “Migration, Fertility, and State Policy in Hubei Province, China.” Demography, 34:481–491. Goldstein, A., S.Goldstein, and S.Guo. 1991. “Temporary Migrants in Shanghai Households, 1984.” Demography, 28:275–291. Harbison, S.F. 1981. “Family Structure and Family Strategy in Migration Decision Making.” in G.De Jong and R.Gardner (eds), Migration Decision Making, pp. 225–251, New York: Pergamon Press. Hervitz, H.M. 1985. “Selectivity, Adaptation, or Disruption? A Comparison of Alternative Hypotheses on the Effects of Migration on Fertility: The Case of Brazil.” International Migration Review, 19:293–317.

The effect of floating migration on fertility

139

Jordan, B. and F.Duvell. 2002. Irregular Migration: The Dilemmas of Transnational Mobility. Cheltenham, UK: Edward Elgar. Lee, B.S. and S.Farber. 1984. “Fertility Adaptation by Rural-urban Migrants in Developing Countries: A Case of Korea.” Population Studies, 38:141–155. Liang, Z. and M.White. 1996. “Internal Migration in China, 1950–1988.” Demography, 33:375– 384. Liang, Z. and M.White. 1997. “Market Transition, Government Policies, and Interprovincial Migration in China: 1983–1988.” Economic Development and Cultural Change, 45:321–336. Long, L. 1970. “Fertility of Migrants to and Within North America,” Milbank Memorial Fund Quarterly, 48:297–326. Millman, S.R. and R.Potter. 1984. “The Fertility Impact of Spousal Separation.” Studies in Family Planning, 15:121–126. Poston, D.L. Jr. 2002. “The Statistical Modeling of the Fertility of Chinese Women.” Journal of Modern Applied Statistical Methods, 1:387–396. Rindfuss, R.R., and J.A.Sweet. 1977. Postwar Fertility Trends and Differentials in the United States. New York: Academic Press. Roberts, K.D. 1997. “China’s ‘Tidal Wave’ of Migrant Labor: What Can We Learn From Mexican Undocumented Migration to the United States?” International Migration Review, 31:249–293. Solinger, D.J. 1999. Contesting Citizenship in Urban China: Peasant Migrants, the State, and the Logic of the Market. Berkeley, CA: University of California Press. Stephen, E.H. and F.D.Bean. 1992. “Assimilation, Disruption, and the Fertility of Mexican-origin women in the United States.” International Migration Review, 26: 67–88. Sun, C. 2000. “The Floating Population and Internal Migration in China.” pp. 179–191 in Xizhe Peng and Zhigang Guo (eds), The Changing Population of China. Maiden, MA: Blackwell Publishers. Yang, X. 1993. “Household Registration, Economic Reform, and Migration.” International Migration Review, 27:796–818. Yang, X. 2000. “The Fertility Impact of Temporary Migration in China: A Detachment Hypothesis.” European Journal of Population, 16:163–183. Zarate, A. and A.Unger de Zarate, A. 1975. “On the Reconciliation of Research Findings of Migrant-Non-migrant Fertility Differentials in Urban Areas.” International Migration Review, 9:115–156.

10 The impact of language dialect on fertility Xiaodong Wang and Xiuhong You

Introduction In China there are many spoken language dialects. These dialects may be roughly categorized into 7 or 8 groups, namely, Mandarin, Wu, Jin, Min, Yue, Xiang, Gan, and Hakka (Yuan 1960; Norman 1988; You 1992). What is most striking is that some of the Chinese dialects are so different from one another that a speaker of one dialect will not be understood by a speaker of another. In many ways therefore the Chinese dialects may be seen as different languages (Norman 1988:187). In fact, in some universities in the United States, Cantonese (the Yue dialect spoken mainly in Guangdong province) “is offered alongside the standard language in Asian language departments, just as German and Dutch are both taught in departments of Germanic languages” (Norman 1988:187). In her illuminating analysis of European fertility patterns, Susan Cotts Watkins found that linguistic boundaries tended to act as “temporary brakes to fertility decline” while a common language would be both a vehicle for the diffusion of innovation and a necessity for intimate interaction (Watkins 1991:173). This chapter uses the Watkins research as a springboard and explores whether Chinese dialects constitute a barrier to the diffusion of innovation resulting in a slower change in fertility.

Prior literature There has not been an extensive amount of research on language and fertility and even less on the relationship between dialects and fertility. Patterns of language usage and proficiency might be associated with fertility based on the fact that language can be seen as a key indicator of cultural attachment and identification (Fishman 1981). The disappearance of distinctive linguistic behavior may be viewed as a necessary but not sufficient condition for acculturation and eventual assimilation (Lieberson 1970). In China, people tend to speak their own dialect, which constitutes barriers to the diffusion of new ideas between dialect areas. This results in the maintenance of the local culture, including norms governing fertility. Internal migrants from the same hometown tend to gather together in destinations and retain their hometown values, beliefs, and norms, including speaking in their home dialect. This kind of social network is a double-edged sword. On the one hand, it helps the migrants settle sooner and receive various kinds of support from their fellow migrants. On the other hand, this usually hinders their process of assimilation and acculturation in the new destination because it often hinders their adaptation to the new environment (Lopez and Sabagh 1978). In Watkins’s work on French fertility, she found that when she added a language variable into her model, the income variable, which had a highly significant association with marital fertility, illegitimacy, and marriage in 1871, became insignificant in 1961. In

The impact of language dialect on fertility

141

contrast, whether or not the population of an area was largely French speaking was associated with all three of these demographic behaviors in the nineteenth century (Watkins 1991:166). Watkins found that language significantly impacted fertility in France. A common language is more than a vehicle for the diffusion of innovation. Diffusion means a decline in the linguistic boundaries existing within areas and among people. Thus, a common language is necessary for personal interactions such as courtship and gossip while a dialect, “much like a distinctive dress,” is a sign of difference (Watkins 1991:173). Distinctive languages are also markers of distinctive cultures. Since dialects reflect distinctive cultures, they are often slow to disappear. For instance, even though the French government legislated that education was to be conducted in French, there was little success in the diffusion of French until the Third Republic (Watkins 1991:153). Some scholars, in China and worldwide, insist that dialects or distinctive languages are legacies that need to be preserved. Furthermore, languages tend to persist over time because they are rooted in the local culture. Whether language diversification is good for cultural preservation is not the concern in this chapter. The objective here is to explore whether the unification of languages facilitates the decline of fertility. Swicegood and colleagues (1988) presented evidence about the relationship between English proficiency and fertility among the Mexican-origin population in the United States. In this population, it was found that women who speak only English had 2 children on average as compared with more than 3 children for women speaking no English at all. The cumulative fertility of women in the sample monotonically decreased with rising levels of English proficiency. In general this pattern was observed in each age group. In another study, Sorenson (1988) explored cultural and socioeconomic explanations of the effects of language characteristics on the fertility of Mexican-American and nonHispanic white couples. The relevance of pronatalist values among Spanish-speaking Mexican Americans was found. He also identified a subset of English-speaking MexicanAmerican couples with lower fertility and a subset of non-Hispanic whites whose higher fertility was tied to language characteristics. Diffusion can occur temporally or spatially (Tolnay 1995). If one area is near high fertility areas and another near areas with low fertility, then that area tends to have higher fertility than lower fertility. However, language is a very basic condition for diffusion. If dialects prohibit communication between areas, subcultures or traditions tend to be maintained over time, and fertility decline will come about slowly. Mass media actively acts in the process of diffusion. In his work in the United States measuring the penetration of media into counties, Tolnay (1995) used a scale which combined (1) the number of county residents who subscribed to 15 popular national magazines, and (2) the percent of occupied dwellings in each county having at least one radio (Tolnay 1995:304). He found that the extent of media penetration into counties is the most effective variable in reducing the overall impact of fertility potential on actual fertility. Kim and associates’ (1974) study of fertility change in South Korea included an index of exposure to mass media. It was concluded that the South Korean government emphasized family planning via mass media. However, mass media had no effect on

Fertility, family planning, and population policy in china

142

women who were 40–49 years old because of the lateness in their reproductive years. The use of mass media was most effective in decreasing fertility for the two younger age groups. Martine (1996) conducted research on Brazil’s fertility decline during 1965–1995. He found influences of mass media on social behavior, including reproductive behavior. In Brazil, television is now the most popular vehicle for entertainment and information throughout the country. It becomes increasingly important because the great majority of the population does not read a daily newspaper. Martine writes, “[T]elevision has contributed to the homogenization of culture and language; it is a prime source of information and is aggressively manipulated in the formation of public opinion” (Martine 1996:67). A similar phenomenon is in effect now in China. In 1988, television sets were very popular, even in rural China. For every 100 rural households, 94 had television sets, ranging from a minimum value of 12 television sets in Tibetan households to a maximum of 142 in Shanghai households (National Bureau of Statistics 1988). Urban households had an average ratio of 107, with a minimum of 94 in Anhui households to a maximum of 133 in Beijing households (no data are available for Tibetan households) (National Bureau of Statistics 1988). However, one needs to be cautious in using mass media as a proxy for diffusion. Tolnay (1995) noted that the significant effect of media might be due in part to the influence of other omitted variables. For example, exposure to mass media is clearly related to a person’s income and education status.

Hypotheses Based on the given literature, it is hypothesized that provinces in which Mandarin or dialects very similar to Mandarin are the principal dialects will have lower fertility than those where this is not the case. Without a language barrier deterring understanding, better conditions for the diffusion of fertility practices will exist. Family planning policies should thus spread more easily because they are transmitted in Mandarin. Mass media act as the important means to diffuse policies and ideas, including family planning policies. Governments usually make use of mass media to promote policies, and people rely increasingly on mass media to receive information. Therefore the degree of presence of mass media is hypothesized to be positively associated with fertility.

Data and operationalization All the data used here, except for the language variables, were taken from the 1998 China Statistics Yearbook (National Bureau of Statistics 1998). The analysis conducted in this chapter is a provincial level aggregate analysis, consisting of 31 area units, namely, the provinces and autonomous regions. Counties would have been preferred but language data were not available at this lower level of aggregation. The language variables were based on data in the Language Atlas of China (1987), compiled jointly by the Australian

The impact of language dialect on fertility

143

Academy of the Humanities and the Chinese Academy of Social Sciences (Simmons 1999). The dependent variable is the 1998 crude birth rate (CBR), that is, the number of births per 1,000 persons. CBR is not the preferred method for measuring fertility but is widely used when data availability is limited. The independent variables include control variables commonly used in fertility models such as illiteracy (percentage of illiterate or semi-illiterate population aged 15 or older), per capita Gross Domestic Product (GDP) (as an income indicator), and percentage of minority population. The main independent variables are language variables along with variables depicting the diffusion of the mass media. The first independent variable, Mandarin, is a dummy variable that divides all the dialects of China into “speaking Mandarin or a dialect that can be understood by Mandarin-speaking people” (coded 1) and “speaking dialects that cannot be understood by Mandarin-speaking people” (coded 0). The other independent language variable, affinity, attempts to distinguish the relative affinity to Mandarin of each province’s dialect. The provinces are divided into 5 ordered groups from 1 (most dissimilar to Mandarin) to 5 (speaking Mandarin or very similar to Mandarin). The grouping of Chinese dialects was not an easy task. Some provinces have many dialects so that it was difficult to categorize them into one group. However, these provinces are precisely those where it is most difficult for cultural diffusion to occur because the diversified dialects constitute barriers. They are coded as 0 in the first language variable (not speaking Mandarin) or 1–3 in the second language variable (very remote from Mandarin). Yuan Jiahua has produced a “standard reference and most informative overview of Chinese dialectology” while more recent surveys shed new light on the study of it (1960; Forrest 1973; He 1984; Norman 1988; You 1992; Chen 2000). There is rather general agreement that Chinese dialects may be divided into seven major groups, namely, Mandarin (the Northern China dialect), the Jin dialect, the Wu dialect, the Gan dialect, the Xiang dialect, the Min dialect, and the Yue dialect (Norman 1988; Chen 2000). Table 10.1 provides the population size and main locations in China of speakers of these dialects. To code each province into Mandarin/non-Mandarin, the dialect spoken by the majority of the population in the province was determined. In cases where there was no obvious majority dialect, the province was coded as non-Mandarin (0) because this situation constituted diffusion difficulties. The same strategy was applied to the second dialect variable, affinity.

Table 10.1 Chinese dialects and their major locations in China Dialects Mandarin

Population speaking the dialect (millions)

Main location in China

662.2 North of Yangzi River, and in southwestern provinces

Jin

45.7 Shanxi, northern Shaanxi, western Hebei

Wu

69.8 Southern Jiangsu, Zhejiang, southeastern

Fertility, family planning, and population policy in china

144

Anhui Gan

31.3 Jiangxi, eastern Hunan

Xiang

30.9 Hunan

Min

55.1 Fujian, eastern Guangdong, Hainan

Yue

40.2 Guangdong, eastern Guangxi

Hakka

35.0 Southern Jiangxi, west Fujian, eastern Guangdong

Table 10.2 Lu’s quantified indices for selected cities and the dialects groups City Beijing

Lu’s coefficents

Index code

Dialect the city represents

The province the city is in

2 5

N Mandarin

Beijing

Jinan

1.8803 5

N Mandarin

Shandong

Xi’an

1.8674 5

NW Mandarin

Shaanxi

Taiyuan

1.1079 1

Jin

Shanxi

Hankou

1.7167 4

SW Mandarin

Hubei

Chengdu

1.6932 4

SW Mandarin

Sichuan

Yangzhou

1.6586 4

Wu

Jiangsu

Wenzhou

1.0846 1

N Min

Zhejiang*

ChangSha

1.4048 3

Xiang

Hunan

Nanchang

1.3145 2

Gan

Jiangxi

Guangzhou

1.0124 1

Yue

Guangdong

Xiamen

1.2482 2

S Min

Fujian

Note * Although Wenzhou is a city of Zhejiang Province, it is not a city representing the majority of dialects spoken in Zhejiang, which is Wu dialect. Hakka dialect group has no sampled city in this table.

In order to obtain an ordered affinity index of each province’s dialect relative to Mandarin, research by Zhiji Lu (1992) was used. Because of the diversity of geographic dialects, Lu utilized computer software to compare geographic dialects with Beijing, based on the relationship of each area’s dialect characteristics on consonants/vowels and tones. Lu’s research resulted in two sets of correlation coefficients for consonants/vowels and tones (Lu 1992:61). Two coefficients for each area were summed to obtain a comprehensive coefficient of sounds and tones. The summed coefficients were ordered and divided into 5 groups, from low to high, with 5 for Beijing and 1 for the cities and areas of the most remote dialect. Each province was coded with the index value of the

The impact of language dialect on fertility

145

representative dialect in the province. Table 10.2 (second data column) lists the index values of each city’s dialect. Table 10.3 is a complete list of affinity coding for each province. These two language variables will be used alternatively in the models to compare the effects of language on fertility. Other independent variables used in this chapter are, (1) television sets per 100 rural households; (2) income (per capita GDP); (3) education (illiteracy rate of population age 15+, according to the National Bureau of Statistics, the illiterate and semi-illiterate population refers to the population aged 15 and over, who have no ability or limited ability to read; and (4) proportion of minority population per province. Television sets per 100 households is the primary diffusion variable in this study. It is used to examine the mass media effects on fertility. Only television sets in rural households are used because (1) there is not much variation in fertility

Table 10.3 The coding of Chinese dialects by province Region

Coded as Mandarin/ non-Mandarin

Ordered index

Anhui

0

4

Beijing

1

5

Chongqing

1

5

Fujian

0

1

Gansu

1

5

Guangdong

0

1

Guangxi

0

1

Guizhou

0

1

Hainan

0

1

Hebei

1

5

Heilongjiang

1

5

Henan

1

5

Hubei

1

5

Hunan

0

3

Inner Mongolia

1

5

Jiangsu

0

4

Jiangxi

0

2

Jilin

1

5

Liaoning

1

5

Ningxia

1

5

Qinghai

0

1

Fertility, family planning, and population policy in china

146

Shaanxi

1

5

Shandong

1

5

Shanghai

0

4

Shanxi

0

1

Sichuan

1

5

Tianjin

1

5

Tibet

0

1

Xinjiang

0

1

Yunnan

0

1

Zhejiang

0

4

Sources: Yuan (1960), Norman (1988).

among the urban areas where the one-child policy is implemented very strictly; and (2) there is less variation among the number of television sets in urban areas. After conducting model diagnostics, two provinces were dropped from the models as outliers (Shanghai and Shanxi). Shanghai has been a pioneer of fertility decline, and it reached replacement level in the early 1970s, even before the one-child policy gained full-scale implementation in other parts of China (Guo 1990). As a long-time port city exposed to western culture, Shanghai formed low fertility values probably as early as 1955 (Zhang 1981). Although the dialect in Shanghai is the Wu dialect, which is phonetically very dissimilar to Mandarin, Shanghai provides an exception for the diffusion perspective. The other excluded province is Shanxi Province, where Jin dialect is spoken. The Jin dialect is very similar to Mandarin in pronunciation but very dissimilar in intonation. Therefore the sum of correlation coefficients of pronunciation and intonation was low and was coded as more dissimilar to Mandarin than many other dialects. However, Jin dialect is more understandable to Mandarin-speaking people than those dialects with high index values. The problem is with the calculation of the index that gives weight to pronunciation and intonation. This problem only exists in the Shanxi Province. Moreover, no literature can be found as to how to weight pronunciation and intonation. Therefore, the Shanxi Province was dropped from the model.

Methods and description Ordinary Least Squares (OLS) regression is used to determine the effect of language on fertility at the province level. Descriptive data for the variables are presented in Table 10.4. The mean CBR is 14.5 live births per 1,000 births with a minimum of 5.2 and a maximum of 23.7. Second, the mean of television sets is 94.2 television sets per 100 households.

The impact of language dialect on fertility

147

Table 10.5 shows the zero-order correlation of all the variables. The correlation between the dependent and independent variables supports the general hypotheses regarding direction, that is, speaking Mandarin is negatively associated with

Table 10.4 Descriptive statistics of independent and dependent variables Variable

Mean

Minimum

Maximum

CBR (per 1000)

14.5

5.2

23.7

Mandarin (1–0)

0.52

0

1

Affinity (1–5)

3.41

1

5

94.16

11.9

142

64.00

29.1

107.9

17.7

6.5

60.0

Per capita GDP (in 1,000)

7,139

2,301

25,192

% of minority

12.20

0

94.4

Television Sets (per 100 HH) Radios (per 100 HH) Illiteracy 15+ (%)

Table 10.5 Zero-order correlations of dependent and independent variables CBR CBR

Mandarin Affinity Television sets Illiteracy Per capita GDP

1.000

Mandarin

−0.505

1.000

Affinity

−0.670

0.876

1.000

Television sets

−0.838

0.415

0.589

1.000

0.691

−0.312

−0.373

−0.714

1.000

−0.693

0.105

0.190

0.664

−0.440

1.000

0.726

−0.356

−0.525

−0.697

0.681

−0.3791

Illiteracy Per capita GDP Minority

Table 10.6 Standardized regression coefficients of effects of independent variables on CBR Model Mandarin Affinity Television Sets GDP

1

2

3

4

5

6

−0.51**

7

8

−0.33** −0.67***

0.42*** −0.86***

−0.75*** −0.71***

−0.16

−0.31*

−0.34* −0.32**

Fertility, family planning, and population policy in china

148

Illiteracy

0.43**

0.31*

0.30**

Minority

0.34**

0.32**

0.24*

Adj. R

2

0.23**

0.43***

0.73***

0.48***

0.73*** 0.67*** 0.77*** 0.83***

Notes 1 *P<0.05, **P<0.01, ***P<0.001. 2 All coefficients are standardized.

fertility. Standardized language likely has a positive effect on the diffusion of contraceptive use and the practice of limiting fertility. The more dissimilar a province’s dialect from Mandarin, the higher the fertility. There are a few moderate correlations between several of the independent variables, namely, illiteracy and television sets, illiteracy and per capital GDP, per capita GDP and television sets, and minority population and television sets. Television sets has a strong correlation with per capita GDP and seems to be a good indicator of the province’s income level. In the regression analyses, both television sets and income are not used in the same models. The results of OLS regressions are reported in Table 10.6, alternatively including 1 of the 2 independent language variables, Mandarin and affinity. The models shown in the table present only the standardized regression coefficients for the independent variables, along with values for the adjusted coefficient of determination (R2). Model 1 reports that the language variable, Mandarin, by itself explains 23 percent of the variation. The unstandardized regression coefficient for Mandarin (not shown in Table 10.6) indicates that the CBR of those provinces speaking Mandarin is about 4 babies per 1,000 less than provinces speaking dialects dissimilar to Mandarin. In Model 2, affinity (i.e. the ordered categories of dialects) by itself explains 43 percent of the variation. This means that provinces in an upper category of dialect group (closer to Mandarin) will on average have CBRs about 1.5 less children per 1,000 live births. Model 3 shows the effect of television sets. This variable by itself explains 73 percent of the variation. Considering that television sets are a strong indicator of income, GDP is entered as a single variable in Model 4; it is not as powerful a predictor as television sets. Models 5 and 6 are the two most important models in the table. Model 5 is the regression of fertility on Mandarin controlling for illiteracy, income, and minority percentage. Mandarin remains significant, and the three control variables are also significant. Mandarin is a good predictor for CBR even after controlling for the effects of education, income, and minority percentage. The CBR of a province in which people speak Mandarin is about 2.6 children lower than a province in which people do not speak Mandarin. Comparing the standardized coefficients of the 4 independent variables, Mandarin is the second largest coefficient. In fact, these 4 coefficients are all similar. Model 6 includes affinity controlling for the same 3 variables; this is a slightly better model than the preceding. According to the unstandardized regression coefficient for Mandarin (not shown in the table), for provinces speaking a dialect one category closer to Mandarin, their CBR on average is 0.93 children less. A province that is coded as 1 (most dissimilar to Mandarin) has a CBR 3.7/1,000 higher (0.93*4) than a province that is coded as 5 (Mandarin or very close to Mandarin). Also, in Model 6 affinity is the most effective predictor. All 3 control variables remain significant, which indicates that the

The impact of language dialect on fertility

149

affinity variable explains a good amount of the variance that is not explained by the 3 control variables.

Conclusion In the analyses presented in this chapter, two of the key independent language variables are only coded at the provincial level. However, within provinces, dialects often change depending upon the area. Thus province-level coding is likely to introduce error into the language data. One way to improve this is to use county level data, mainly because a county has little variation in dialect groups. This was not possible to undertake for this analysis because data were not available. Second, more variables need to be included in the study. Women’s employment status, which is one of the key variables in other models, should be introduced. Furthermore, language variables may have interactive effect with education and income. Though the models estimated here indeed control for education and income, these variables may not be the most appropriate indicators. The affinity variable is also not without question. The two coefficients of consonants/vowels and tones were summed to form the affinity index. This means that consonants/vowels and tones share the same weight in the variable. For example, with the consonants/vowels and tones sharing the same weight, Shanxi province becomes an outlier. Some may argue that tones may not have the same weight as the consonants/vowels when understanding another language. Had data been available on the proportions of people in a province speaking a specific dialect, it would have been possible to use a weighted index. Again, requisite data were not available. The major finding is that speaking different dialects does indeed affect fertility. Those provinces speaking Mandarin have about 2.6 less births per 1,000 persons, controlling for the effects of income, education, and minority percentage, compared to those not speaking Mandarin. Compared with provinces speaking the most dissimilar dialects from Mandarin (code 1), the provinces speaking Mandarin have 3.7 fewer births per 1,000 population. These results are similar to those of Watkins (1991) and Swicegood and his colleagues (1988). The language variable examined here is of special interest because dialects have not been utilized to explain fertility differences. It does appear to be an effective indicator for explaining CBRs. The results reported here indicate that fertility differences across the provinces are significantly related to dialect differences. China’s diversified dialects may be affecting fertility levels by constituting a barrier to the diffusion of modern fertility norms. Therefore, if Mandarin is promoted more within China, fertility differences among the provinces could well be reduced. The promotion of Mandarin could well be introduced, with less attention given to the promotion of the one-child policies. Another interesting finding is that prevalence of television sets, as the major mass media in China, has a significant negative relationship with fertility. As television sets begin entering the majority of households, diffusion should increase even more so, and the CBR will also decline.

Fertility, family planning, and population policy in china

150

In conclusion, as China becomes more and more dominated by one language, Mandarin, and with increases in access and viewing of mass media, especially television, the results reported here would suggest that provincial level fertility should decline.

References Australian Academy of the Humanities and Chinese Academy of Social Sciences. 1987. Language Atlas of China. Hong Kong: Longman. Chen, M.Y. 2000. Tone Sandhi: Patterns across Chinese Dialects. Hong Kong: Cambridge University Press. Fishman, J.A. 1981. “Language Maintenance and Ethnicity.” Canadian Review of Studies of Nationalism, 8:229–247. Forrest, R.A.D. 1973. The Chinese Language, London: Faber and Faber Ltd Guo, S. 1990. “Shanghai: Pioneer of Fertility Decline in People’s Republic of China—Trends and Determinants of Fertility Transition, 1950–1984.” Ann Arbor, MI: University Microfilms International. He, G. 1984. A Brief History of Chinese Dialects Research. Shanxi: Shanxi People’s Press. Kim, M., R.V.Rider, P.A.Harper, and J.Yang. 1974. “Age at Marriage, Family Planning Practices, and Other Variables as Correlates of Fertility in Korea.” Demography, 11: 641–656. Lieberson, S. 1970. Language and ethnic Relations in Canada. New York: Wiley. Lopez, D.E. and G.Sabagh. 1978. “Untangling Structural and Normative Aspects of the Minority Status-fertility Hypothesis.” American Journal of Sociology, 83:1491–1497. Lu, Z. 1992. A Preliminary Probe for Quantifying Chinese Dialects. Beijing: Language Press. Martine, G. 1996. “Brazil’s Fertility Decline, 1965–95: A Fresh Look at Key Factors.” Population and Development Review, 22:47–75. National Bureau of Statistics. 1998. 1998 China Statistics Yearbook. Beijing: National Bureau of Statistics. Norman, J. 1988. Chinese. Cambridge, MA: Cambridge University Press. Okun, B.S. 1994. “Evaluating Methods for Detecting Fertility Control: Coale and Trussell’s Model and Cohort Parity Analysis “Population Studies, 48:193–222. Simmons, R.V. 1999. Chinese Dialect Classification, Amsterdam and Philadelphia, PA: John Benjamins Publishing Company. Sorenson, A.M. 1988. “The Fertility and Language Characteristics of Mexican-American and NonHispanic Husbands and Wives.” The Sociological Quarterly, 29:111–130. Swicegood, G., F.D.Bean, E.H.Stephen, and W.Opitz. 1988. “Language Usage and Fertility in the Mexican-Origin Population of the United States.” Demography, 25: 17–33. Tolnay, S.E. 1995. “The Spatial Diffusion of Fertility: A Cross-Sectional Analysis of Counties in the American South, 1940.” American Sociological Review, 60:299–308. Watkins, S.C. 1991. From Provinces into Nations: Demographic Integration in Western Europe, 1870–1960. Princeton, NJ: Princeton University Press. You, R. 1992. An Introduction to Chinese Dialectology. Shanghai: Education Press. Yuan, J. 1960. An Introduction to Chinese Dialectology. Beijing: Wenzi Gaige Press. Zhang, C., M.Liu, and Y.Hu. 1981. “Shanghai: Population Developments Since 1949.” pp. 129– 135 in Z.Liu and J.Song (eds), China’s Population: Problems and Prospects. China Studies Series. Beijing: New World Press.

Part IV Implications and the future

11 The managed fertility transition in rural China and implications for the future of China’s population Che-Fu Lee and Qiusheng Liang Assessments of the impact of China’s population policy often cite the statistic that “more than 300 million” births were averted over the latter half of the twentieth century, and this is attributed to the government’s administration of birth planning (Wei and Wang 1996; Ca et al. 1999). Some assessments include evaluations of socioeconomic benefits as extrapolated from the reduced population growth due to government intervention. The central government, for example, put out a white paper on “China’s Population and Development in the 21st Century,” in which it stated, Since the implementation of planned child births, the cumulative number of prevented births nationwide has been over 300 million. This saves the nation and the society an enormous amount of costs of child rearing, eases the population pressure on the natural resources and environment, and helps accelerate the economic development and raise the people’s level of living. (People’s Daily, December 20, 2000) The National Committee on Planned Fertility in 1999 organized a number of experts from universities to undertake a project titled, “China’s Planned Fertility: Its Investment and Return.” In its final report, the group said, “based on trend analyses,” there have been 338 million prevented births. However, none of these reports have ever referred to their sources of information, nor have any discussed how their estimated number of prevented births was produced. This chapter evaluates the effects of birth control in China via a comparative analysis (Liang and Lee 2003). China’s population of 1.27 billion, as reported by the most recent 2000 census, will serve as a baseline reference for comparison. Various projections extrapolating China’s population and its growth rate in earlier eras, with explicit assumptions of the policy effects (or lack of), will be compared with China’s population in 2000. The comparative differences will be shown to be attributed to birth planning programs introduced at various points in time. Before doing so, however, it is necessary to revisit the transition process of China’s fertility, especially the transition in rural areas, and to judge on the basis of the most recently available data the accuracy of the belowreplacement rate of low fertility reported for both urban and rural China. Since about 70

Fertility, family planning, and population policy in china

154

percent of China’s population is found in rural regions, the credibility of the overall national fertility statistics depends largely on the reliable reports of rural births. A review will thus be undertaken here first of the managed fertility transition in rural China. Then an assessment will be undertaken of the effects of voluntary and involuntary aspects of China’s birth control policy on the size of the population in 2000. The implications will then be discussed for the future.

Urban and rural paths of fertility transition According to the latest report of the 2000 census, China has reached a below-replacement level of low fertility, nationwide in both urban and rural areas. The total fertility rate (TFR) in 2000 was 1.0 in urban and 1.4 in rural places. According to data from the 1990 census, which reported a TFR of 1.55 for urban and 2.54 for rural places, the rural and urban fertility rates in China have further converged from above-replacement levels to below-replacement levels, that is, a TFR nationwide from 2.35 to about 1.15. By the year 2000, urban China had literally conformed to the “one-child policy” expectation. The cities had a TFR of 0.87 and the townships, 1.08, both of which jurisdictions were included as urban. As reviewed in the introductory chapter in this volume, the precipitous decline in fertility began in China in the mid-1960s. Before that time, in fact, urban fertility, measured by the crude birth rate (CBR, or the number of births per 1,000 people) was higher than the rural CBR. In 1957, the year before the calamity period of 1958–1961, the urban CBR was 44.5 per thousand and the rural, 32.8 (Table 11.1). Given the lower death rate in urban than in rural places, the natural growth of the urban population was higher than the rural population until the early 1960s. This urban-rural differential birth rate may have been due, in part, to the underreporting of rural births. For example, still births and infant deaths may tend to be unreported or unregistered more in the countryside than in the city. And, the higher rate of births in urban areas is not as readily detected with the TFR, which reflects the age pattern of fertility more than the ratio of births to the population (compare TFRs of urban and rural as shown in Table 1.1). At any rate, both urban and rural fertility rates in China before mid-1960 are near what demographers term “natural fertility,” or fertility in which there is little or no conscious regulation. As such, the average number of births per woman (TFR) fluctuates around 6.0. The mid-1960s appear to be the watershed period for the urban, but not the rural, fertility decline. In 1965, the urban CBR came down to 26.6 from 44.5 per thousand in 1957; and the TFR from 5.9 to 3.7. The rural counterpart, in reverse, went from 32.8 to 39.5 per thousand in CBR; and 6.5 to 6.6 in TFR. Clearly indicated by these statistics is that the post Great Leap Forward program of birth planning did not have much of an effect in rural places, but exerted a noticeable effect on the fertility behavior of the urban population. The political turmoil of the Great Cultural Revolution (1966–1976) delayed the early 1960s planned birth campaign in rural areas for a few more years, but did not disrupt very much its implementation in most cities. Urban fertility continued

The managed fertility transition in rural China

155

Table 11.1 Crude rates of birth, death and natural increase for urban and rural China, selected years, 1957–1999 Selected years

Urban (per thousand) Birth

Death

Rural (per thousand)

Increase

Birth

Death

Increase

1957

44.48

8.47

36.01

32.81

11.07

21.74

1962

35.46

8.28

27.18

37.27

10.32

26.95

1965

26.59

5.69

20.90

39.53

10.06

29.47

1975

14.71

5.39

9.32

24.17

7.59

16.58

1978

13.56

5.12

8.44

18.91

6.42

12.49

1980

14.17

5.48

8.69

18.82

6.47

12.35

1989

16.73

5.78

10.95

23.27

6.81

16.46

1990

16.14

5.71

10.43

22.80

7.01

15.79

1991

15.49

5.50

9.99

21.17

7.13

14.04

1992

15.47

5.77

9.70

19.09

6.91

12.18

1993

15.37

5.99

9.38

19.06

6.89

12.17

1994

15.13

5.53

9.60

18.84

6.80

12.04

1995

14.76

5.53

9.23

18.08

6.99

11.09

1996

14.47

5.65

8.82

18.02

6.94

11.08

1997

14.52

5.58

8.94

17.43

6.90

10.63

1998

13.67

5.31

8.36

17.05

7.01

10.04

1999

13.18

5.51

7.67

16.13

6.88

9.25

to decline and reached a low of less than 15 per thousand by 1975. Since then, the urban CBR has been maintained around this level or lower with only minor fluctuations. With a low death rate of around 5 to 6 per thousand, the natural growth rate of the urban population has been reduced to a level of 10 per thousand or less since 1975. Turning to rural fertility, the transition process has followed a very different path. The rural TFR continued to hover around 6.0 into the early years of the 1970s. The rural CBR did not fall below 30 per thousand until the mid-1970s, nor below 20 per thousand till the end of the 1970s. With a demographic echo of the 1960s baby boom, the rural CBR exceeded 20 per thousand again in the late 1980s and early 1990s. This was followed by a decline, but the rural CBR has yet to achieve the low level of 15 per thousand as that of urban fertility in 2000 (see Table 11.1). With a slightly higher death rate than in urban China, the natural growth rate of the rural population first reached a low level of 10 per thousand in the late 1990s. In short, the path of fertility transition traveled by the urban

Fertility, family planning, and population policy in china

156

population in the mid-1970s was delayed by 20 years for the rural population, till the mid-1990s.

Managed rural births and birth records China, with a vast countryside of 800 million persons, which represents about 70 percent of the total population, depends on the success of the policy-induced decline in rural fertility in order to arrest the fast rate of population growth. The fertility transition in rural China just depicted is an impressive and almost incredible accomplishment of policy administration in bringing overall fertility to about the replacement level in the most recent two decades. To what extent has the birth planning system faithfully enforced planning regulations to conform to policy expectations? To what extent, on the other hand, can the apparent success in the rural areas be attributed to the official and mass manipulation of birth records? These are legitimate and appropriate questions deserving a close scrutiny. Merli and Raftery (2000) argue that under stringent population policies, especially the one-child policy beginning in 1979, there are motivations for individual couples and local officials to underreport births. Couples are rewarded if they have only one child and penalized if they exceed that limit. Local officials are responsible for meeting the annual quota of births set by upper-level authorities. Out-of-plan births in their jurisdiction during a given period would likely be hidden. To assess the extent of underreported births in rural China, Merli and Raftery analyzed the distribution of first birth intervals from the time of marriage, using reproductive history data collected in 1992 from 4 rural counties in the northern provinces of Hebei and Shandong. They found evidence in 3 of the 4 sample counties of underreporting of the first birth of girl babies, and their being replaced with second births. This was seen in a sudden surge of reported “first births” at about 20 months since marriage, and an abnormally high sex ratio (many more male than female births than normally expected) of the first births at these extra-long intervals. There are other studies which also detected underreporting or under-recording of female first births. Needless to say, the extent of distorted birth statistics varies among the diverse rural places in different regions and uneven levels of socioeconomic development. The most important determining factor, however, is the effectiveness of local authorities in administering the birth planning program according to the central population policy and the varying intensity of the birth control campaigns from period to period. Pertinent to this, rural communities tend to vary in the processes of the fertility transition in association with their relative access to nodes of transportation more than with their level of economic development (Xie 1997). Isolated localities are more insulated from policy central supervision and are more likely to deviate from the uniformly designed program (Feeney and Yuan 1994). Does this mean that China’s current low fertility has been exaggerated because of underreported births?

The managed fertility transition in rural China

157

Policy adjustment and uneven regional implementation Beginning in 1971, the government at all levels reinforced the policy of “wan, xi, shao” (late marriage, spacing between births, and few in number). Women of reproductive age were closely monitored by the “neighborhood committee” in the city, and the committee on planned birth in the village, regarding the age of marriage, the timing, and number of births. Since the limitation of family size was relatively flexible, allowing for 2 and in some cases 3 children, few violent protests or confrontations with the enforcing officials were observed, even in rural areas. The overall impact of the 1970’s policy enforcement was seen in the precipitous decline of the fertility rate during the decade. Along with economic reforms, the Chinese government experimented with a compulsory “one child per couple” policy in 1979. The administrative measures to implement the new compulsory policy include welfare incentives offered to the only child whose parents pledged to have no more births, on the one hand, and a monetary penalty levied against couples who gave birth to any children beyond the first. Incidents of forced abortions dictated by over-zealous officials at the local clinics were reported, which touched off international outcries of human rights violations. Violence caused by the conflicting interests of officials and local people, especially in the rural areas, became almost inevitable. The central government, under these circumstances, was forced to modify the compulsory one-child policy in 1984, and decentralized the policy authority to the provincial and local level, allowing for diversity in population planning according to the local conditions. Except in major cities, couples were permitted to have a second pregnancy if the firstborn was a girl. In essence, since 1984, the policy has reverted to the more flexible control strategy of the 1970s. Moreover, in 1981, the government promulgated a new “marriage law” which lowered the minimal legal age for marriage to 22 for men and 20 for women, compared to previously enforced ages of 27 and 25. The decentralization of the birth planning policy and the lowering of age at marriage contributed to an earlier age pattern of fertility and a slight spurt in the birth rate in the mid-1980s, especially in rural areas. No generalized patterns of response to these policy adjustments and law changes, however, are found for the rural regions. There have been ample case studies of villages and rural counties in various provinces finding rural fertility at low levels since the early 1980s (Lavely 1984; Feeney et al. 1989; Lavely and Freedman 1990; Choe and Tsuya 1991; Feeney and Yuan 1994; Greenhalgh et al. 1994). Among the various explanatory variables, the most powerful determinant of a lowered fertility appears to be the leadership of a given jurisdiction. The key to produce successful results is to strictly carry out a scheduled childbirth plan for women of reproductive ages within the jurisdiction, rather than stringently limiting a couple to one child, In fact, the scheduled childbirth plan under a “one-boy-two-child” policy, as a revision of the one-child policy, has had a remarkable imprint on the age pattern of fertility. This can be shown by using single-year age-specific fertility rates to compare the age schedules of childbirth as reported by the 2000 census with those of the 1990 census, by cities, townships, and villages. Figure 11.1 shows that fertility rates for China as a whole became lower at all ages in 2000 than in 1989. This is true for cities, townships, and rural villages. Whereas the fertility schedule for townships closely coincided with that of cities in 1989, the city fertility pattern declined further at all ages

Fertility, family planning, and population policy in china

158

and was set apart from the township fertility pattern. Noticeable in the age pattern of village fertility in 2000 are the relatively high rates of births in the age range from the late twenties to the early thirties. This scheduled birth of a second or maybe third child after a waiting period of four or so years can be seen more distinctly in the data for Hebei province in 2000 (see Figure 11.2). A small peak of births occurred between the ages of the late 20s and early 30s in villages. The same pattern is discernible in townships, though not as dramatic.

Figure 11.1 Age-specific fertility rates by urban and rural places: all China, (a) 1989 and (b) 2000.

The managed fertility transition in rural China

159

The evidence is convincing that China’s administration at all levels of government has managed to depress fertility rates at a very low level, perhaps around the just-replacement level, even in the vast countryside. Of course, there are exceptions such as in the western provinces of Yunnan, Tibet, Shaanxi, Gansu, Qinghai, Ningxia, and Xinjiang, which are more sparsely populated and have concentrations of ethnic minorities. Some of these regions are not required to conform to the national policy and continue to maintain a slightly higher rate of fertility and hence higher rate of population growth than the national average. As pointed out

Figure 11.2 Age-specific fertility rates by urban and rural places: Hebei province (a) 1989 and (b) 2000.

Fertility, family planning, and population policy in china

160

by a study of fertility preferences (Merli and Smith 2002), however, in urban and rural counties where family planning policy has been rigorously implemented, the acceptance of the officially sanctioned family size, “one-boy-two-child” policy seems prevalent. To what degree this is simply a public acquiescence to the policy demand is difficult to ascertain. This chapter turns next to an assessment of the effect of birth control by way of a number of comparative analyses of population growth trajectories. By comparing alternative scenarios for the size of China’s population in 2000, the effects of various policy components can be evaluated, and the extent of underreporting of births over the past two decades may be properly estimated.

Comparative analysis of policy effects China’s population in 2000 was 1.27 billion, according to the 2000 census. What would have been the number of China’s population in 2000 if it had followed a different but plausible pattern of growth without a particular policy intervention? Presented in this section are a number of alternative scenarios. No change due to socioeconomic development or policy intervention In spite of a huge swing in the rate of population growth in the decades of 1950 and 1960, the normal rate of natural increase is about 25 per thousand per year during the period of the mid-1950s to 1970 (see Table 11.1). First, the level of population growth of 25 per thousand per year, taken as a growth rate for a population experiencing a decline in mortality while maintaining a “natural fertility” rate, is fairly modest and common. During the 1960s subsequent to the calamity years of 1958–1961, China had seen a precipitous decline in its mortality rate, while during the same period the fertility rate was maintained well above 30 per thousand per year. Thus, the actual annual growth rate during the decade of the 1960s was clearly above 25 per thousand. There were no reasons to expect changes in demographic behavior, since then, an extrapolation (with an exponential equation) from the 1954 population of 603 million projects a population of 1.90 billion in year 2000, a tripling of its population in 46 years. Of course, the assumption of no change over more than four decades may be deemed as unrealistic. In fact, the tripling of a population during the later half of the twentieth century was not at all unusual. Of 225 countries and territories worldwide for which there are records, 100 (44 percent) of them increased their population three fold or more over the decades 1950–2000. In the developing regions of Asia, population tripling occurred in 29 of 50 countries or territories (58 percent); in Africa, 41 of 56 (73 percent); and in Latin America and the Caribbean, 21 of 51 (41 percent). In fact, some countries where family planning programs had been implemented in the 1960s, still tripled their population by the year 2000 because of their young population base to begin with and a dramatic decline in death rates thereafter. Thus the number, 1.90 billion, or nearly 2 billion, as a potential size of China’s population in 2000, may be used as a baseline reference for later comparison.

The managed fertility transition in rural China

161

Impact of administrative monitoring of the compulsory birth control: India as a reference of voluntary planning By 1970, socioeconomic conditions in China had largely stabilized, though the political turmoil of the Cultural Revolution had not completely disappeared. Another baseline population for the purpose of comparison may be obtained by simply extrapolating the base population of 830 million in 1970 with its actual growth rate of 26 per thousand in 1970. This projects a population of 1.8 billion in the year 2000. This, interestingly, yields a figure of only about 100 million less than the 1.9 billion as projected from the mid1950s (see the previous section). However, this again assumes that no socioeconomic factors or policy influenced the changes in demographic behavior subsequent to the year of 1970. Alternatively, one may look at the Indian experience for an empirical reference of changing demographic behavior given a certain level of economic development and a population policy without a rigorous administrative enforcement. Except for this latter divergence in population policy since 1970, China and India were strikingly alike in their levels of development in the mid-twentieth century. Both countries established a new independent regime around the 1950s with the largest populations in the world. Both have a long cultural tradition of favoring early marriage, a large-family ideal, and son preference. There are sufficient indicators, moreover, showing that the two countries were similar in their standards of living in the early 1950s (see Table 11.2). The population of China was 575 million in 1952, and that of India, 367 million. In terms of GDP per capita, India in 1952 had an edge over China (US$ 154 vs. 101). The birth rates for both were high, 37 per thousand for China and 40 for India, but China had a considerably lower death rate (18) than India (28). Infant mortality rates were high for both (176 for China and 190 for India). Per capita calorie-intake levels were low for both (1,917 and 1,540), and the adult illiteracy rates were similarly high (25 and 20 percent). China’s population up to 1970 was fluctuating around the more stable level of India’s rate of population growth (see Figure 11.3). The clear divergence in the rates of growth began in 1970 with China lowering its rate precipitously thereafter. It is plausible, therefore, to assume India’s pattern of population growth, subsequent to 1970, for China’s, were China to adopt no compulsory population control policy. Population in India grew from 555 million in 1970 to 1.02 billion in 2000, or 1.84 times larger than 30 years ago.

Table 11.2 Development indices of China and India around the year 1950 Index (unit)

Year

China

India

Ratio (China: India)

GDP per capita (dollar)

1952

101

154

0.66

Population (million)

1952

575

367

1.57

Birth rate (per thousand)

1950

37

40

0.93

Death rate (per thousand)

1950

18

28

0.64

Life expectancy (year)

1950

40

32

1.25

Fertility, family planning, and population policy in china

162

Infant death rate (per thousand)

1950

176

190

0.92

Adult illiteracy rate (%)

1950

25

20

1.25

Calories (per capita)

1952

1917

1540

1.24

Source: Subramanian Swamy, “Can India Catch Up with China?” India’s National Magazine, 17(15), July 22-August 4, 2000.

Figure 11.3 Population growth rates of China and India, 1950–2000. Sources: (1) Yao, Xin-Wu and Hua Yi. General Demographic Statistics of China. Beijing: Chinese Population Publishers, 1994; (2) US Bureau of the Census. “International Data Base: 2000.” Washington, DC: US Government Printing Office, 2000; (3) Population Division, Department of International Economic and Social Affairs, United Nations. World Population Projections: 1998. New York: United Nations, 1999. Were China to follow the path of India, with no compulsory control of births, since 1970, China would have a population estimated at 1.53 billion in 2000. Compared with 1.80 billion as projected with the growth rate of China in 1970, 270 million less births could have been attributed to changing fertility behavior without a rigorous program of administrative monitoring of women’s reproductive behavior in China. But compared with the actual population of 1.27 billion as counted in the 2000 census, 1.53 billion is 260 million in excess. In other words, 260 million less births in China during 1970–2000 would have been due to the impact of the administrative monitoring of compulsory birth control. This figure of 260 million is not far from the “300 million” births averted that is

The managed fertility transition in rural China

163

often cited. Whether those official citations refer specifically to the number of prevented births, which are attributed to the rigorous compulsory birth control program, as is assumed here, is unclear. The impact of “one-child” policy? The short-lived “one child per couple” campaign beginning in 1979 generated a great deal of tension in China and received wide international attention. How much of a demographic dent did it make on the size of the population in the year 2000? In 1980, China’s population was 987 million and the rate of growth was 12 per thousand. Extrapolating from 1980 to the year 2000, the population of China in 2000 would have been 1.25 billion. This is a figure about 20 million less than the actual population of 1.27 billion according to the 2000 census. In fact, the growth rate in most of the 1980s rebounded to 15–16 from 12 per thousand in 1980, due to the policy adjustment of “one-boy-two-children” from the one-child policy, and the lowering of the legal marriage age, as discussed before. What is clearly evidenced here is that the muchfanfared “one-child” policy espoused by the Chinese government during the period of 1980–1984 has, in fact, done little in accomplishing immediately the intended population control. The population of 1.27 billion counted in the 2000 census is 20 million in excess of the population of 1.25 billion projected from the population and growth rates of 1980. The rebounded birth rate in the 1980s may account for most of this excess number of people. Alternatively, if the 1990 population of 1.14 billion is projected using its rate of natural increase, 14.4 per thousand, China’s population in 2000 would be 1.32 billion. This is 50 million more than the 1.27 billion reported in the 2000 census. This suggests another spectacular drop in fertility during the last decade of the twentieth century. Apparently, the administration of planned birth programs was reactivated more vigorously after a rise in period fertility during the 1980s. Some have raised reasonable doubts about the further reduction in fertility in the 1990s to the extent that the total fertility rate fell below the replacement level and the rate of natural increase reached just about 1 percent. Attene and Sun (1999) point out contradictory statistics reported from registration and survey data by different government agencies for the inter-census years of the 1990s. A special 1995 survey of rural counties by Zeng (1996) estimated an underreporting rate as high as 37 percent for rural births. It is assumed here that whatever error in the 2000 census counts failed to uncover the underreported births cannot be such a glaring discrepancy. It seems safe to conclude that China’s population of 1.27 billion is 20 million more than would be expected projecting a natural rate of increase from 1980, and 50 million less than would be expected by the same procedure from 1990. However, a closer analysis of the population at age 20 and below will be necessary when the detailed 2000 census data are made available to acquire a better understanding of the managed fertility and birth statistics in the 1980s and 1990s.

Fertility, family planning, and population policy in china

164

Conclusion In summary, government leadership in China noticed the problem of the rapid growth of the population as early as in the 1950s. At then-current rates of population growth, China’s population would have approached 2 billion at the turn of the twenty-first century. The effective intervention by an official control policy did not begin until 1970. Without that policy intervention, the size of China’s population in 2000 could have been as large as 1.8 billion. The actual population of China as at the end of year 2000 is 1.27 billion. This is more than a half billion less than it could have been. Assuming, furthermore, that even without the government’s rigorous administrative monitoring, China’s fertility could have declined over the last three decades, like India, because of socioeconomic change and available voluntary family planning services, China’s population would thus have still grown to a size of over 1.5 billion at the end of the twentieth century. In this scenario China would have a population 260 million more than was actually the case. Restated more succinctly, China’s population today is at least 500 million less than it would have been because of changes during the last three decades, including government efforts in implementing a compulsory population control policy. Of these more than 500 million births averted, the administrative and close monitoring of citizens’ reproductive behavior since 1970 in both the urban and rural regions accounted for more than half of the reduced population growth. The spurt of the radical “one-child” policy in the early 1980s was soon adjusted in 1984 to allow for a second birth if the first was a girl. The immediate effect was seen in a rise in period fertility rates in the subsequent years of the 1980s. Allowing for the possibilities of the underreporting of unplanned births, especially in rural places, fertility declined further in rural China in the 1990s to nearly converge with urban China. The rural fertility transition was met with rapid socioeconomic transition in China, especially in the 1990s. The residential registration system installed in the 1950s to control the mobility of the rural population was largely relaxed. City employers were able to acquire residential permits, however temporarily, for employees recruited from rural places. Young men and women migrant workers, if unmarried, tended to postpone their marriages to save money. Migrant married men tended to leave families behind at least for the earlier years of migration. Aided by the effective management of planned births of local authorities, all these factors helped keep down the period birth rate of the 1990s. Whether the low fertility rate in most of rural China is here to stay, and whether future levels will remain at the just-replacement levels are open questions. As envisioned in the government’s white paper, “China’s Population and Development in the 21st Century,” China’s population should reach its peak of 1.6 billion by mid-century and then begin to decline. For this to come true, the government’s role in managing the citizens’ reproductive behavior will have to continue for a long time to come. Moreover, the rapid speed of policy-induced fertility decline has already introduced drastic changes in the demographic structure. The aging of China’s population has begun to set in. The desire to have male posterity, at least one in a very limited-size family, has contributed to a very high ratio of boys relative to girls (see Chapter 12 in this volume by Poston and Glover). The uneven pace of development between urban and rural areas, and between regions, requires a major redistribution of resources, including human resources.

The managed fertility transition in rural China

165

The market economic transformation has replaced state-owned enterprises, which used to provide the bulk of workers’ health care, pension, and other welfare. The commune, or collective management, is no longer present to manage the public goods for the vast populace in the countryside. China’s successful management of population reproduction has undoubtedly helped to alleviate population pressures on capital and the environment. Innumerable other management tasks still remain.

References Attane, I. and M.Sun. 1999. “Birth Rates and Fertility in China. How Credible are Recent Data?” Population: An English Selection, 11:251–260. Ca, R., W.Hu, and J.Zei. 1999. Demography of A Century. Beijing: Arts and Literature Publishers. Choe, M.K. and N.O.Tsuya. 1991. “Why Do Chinese Women Practice Contraception? The Case of Rural Jilin Province.” Studies in Family Planning, 22(1):39–51. Feeney, G. and J.Yuan. 1994. “Below Replacement Fertility in China? A Close Look at Recent Evidence.” Population Studies, 48:381–394. Feeney, G., F.Wang, M.Zhou, and B.Xiao. 1989. “Recent Fertility Dynamics in China: Results from the 1987 One Percent Population Survey.” Population and Development Review, 15(2):297–322. Greenhalgh, S., C.Zhu, and N.Li. 1994. “Restraining Population Growth in Three Chinese Villages, 1988–93.” Population and Development Review, 20(2):365–395. Lavely, W.R. 1984. “The Rural Chinese Fertility Transition: A Report from Shifang Xian Sichuan.” Population Studies, 38:365–384. Lavely, W. and R.Freedman. 1990. “The Origins of the Chinese Fertility Decline.” Demography, 27(3):357–367. Liang, Q. and C.F.Lee. 2003. “Comparative Analysis of the Effects of Birth Control in China.” Population Research, 1:5–10. Merli, M.G. and A.E.Raftery. 2000. “Are Births Underreported in Rural China? Manipulation of Statistical Records in Response to China’s Population Policies.” Demography, 37(1):109–126. Merli, M.G. and H.L.Smith. 2002. “Has the Chinese Family Planning Policy Been Successful in Changing Fertility Preferences?” Demography, 39(3):557–572. People’s Daily (overseas edition). 2000. “China’s Population and Development in 21st Century.” A White Paper by the Press Office of the State Council, December 20. Population Division, Department of International Economic and Social Affairs, United Nations. 1999. World Population Projections: 1998. New York: United Nations. She, C. 1988. A History of Planned Birth Activities in China. Urumqi, China: Xinjiang People’s Publishers. Subramanian Swamy, “Can India Catch Up with China.” India’s National Magazine, 17(15), July 22-August 4, 2000. U.S. Bureau of the Census. 2000. “International Data Base: 2000.” Washington, DC: U.S. Government Printing Office. Wei, J. and S.Wang. 1996. Evaluation and Strategy of China’s Population Control. Beijing, China: Higher Education Publication.

Fertility, family planning, and population policy in china

166

Xie, W. 1997. Family Changes in Rural and Urban China, 1950s to 1980s: A Multilevel Model Analysis. Unpublished Ph.D. Dissertation. Washington, DC: The Catholic University of America. Yao, Xin-Wu, and Hua Yi. 1994. General Demographic Statistics of China. Beijing: Chinese Population Publishers. Zeng, Y. 1996. “Is Fertility in China in 1991–1992 Far Below Replacement Level?” Population Studies, 50(1):27–34.

12 China’s demographic destiny Marriage market implications for the twenty-first century Dudley L.Poston, Jr and Karen S.Glover

Introduction Auguste Comte, the French mathematician, philosopher, and the founder of sociology, is believed to be the first person to have written about the effects that demography and the demographic processes have on society and society’s future (see the selection from Martineau’s The Positive Philosophy of Auguste Comte in Thompson (1976:156–157)). Were Comte writing today and in English, he would likely be stating that “demography is destiny.” This phrase means that population processes and their effects on the changing composition of their members have important impacts on the future society. Demography is destiny precisely because once the demographic shifts and changes have occurred, the future is set. Based on the demographic events that have occurred, astute observers are able to figure out what will happen in the future, or at least what might happen in the future, before the events of the future actually occur. This chapter deals with the demographic destiny of the People’s Republic of China (PRC). Earlier chapters in this book have shown that since the 1960s China has experienced a dramatic and extremely rapid fertility transition from around 6 children per woman to around 1.7 children per woman in 2001 (Figure 12.1). This rapid fertility transition has resulted in a remarkable demographic occurrence in the numbers of boys and girls born each year. In every year beginning in 1980 to the year 2001, many more Chinese boys have been born than Chinese girls. This demographic event has occurred, and China’s demographic destiny has been determined. This event has important and relevant implications for China’s marriage market starting around the year 2000. In this chapter it is estimated that there will be more than 23 million boys born between 1978 and 2001 in China who will not be able to find Chinese brides. The cause of this demographic shift is China’s demographic dynamics since 1950 that culminated in a dramatic fertility transition. This chapter first considers the demographic occurrence of many more boys being born each year than girls and its implications for the destiny of China.

Population growth in China and the fertility transition China is now completing its demographic transition from high rates of fertility and mortality to low rates. The country is in the third phase of the transition, that

Fertility, family planning, and population policy in china

168

Figure 12.1 TFRs: China, 1950–2001. of low fertility and mortality rates. Since the early 1960s, China has experienced a pronounced and rapid decline in its fertility, to a level in 2001 of 1.8 children per woman (see Figure 12.1). This section traces the history of population dynamics in China since Mao Zedong and the Chinese Communist Party (CCP) assumed control in 1949 (see Liang and Lee’s Chapter 1 in this volume for more discussion). When Mao Zedong and the Chinese Communists took over China in 1949, relatively little attention was paid to the size and growth of the population. This was a period of doctrinaire Marxism (Aird 1972). But when the initial results of the 1953 census became available in 1954, there was anxiety expressed about the size and growth trends of the country, and by the summer of 1956 a birth control campaign was underway. With the introduction in 1958 of communes and the Great Leap Forward however, China started to reverse its new birth control policy: “A large population was once more regarded as advantageous, and the vicious attacks on Malthusians, ‘rightists’ and ‘bourgeois economists’ who championed birth control again shifted into high gear” (Orleans 1972:40). The Great Leap Forward, initiated in 1958, was designed to “involve a revolutionary struggle against nature to realize the great potential of agriculture by maximizing the advantages of the collective economy” (Aird 1972:278). It had a short life because in 1959 China suffered an economic crisis and famine (Ashton et al. 1984), resulting in the premature death of at least 30 million people. The fertility decline from the mid-1950s through the early 1960s resulted from the “national hard times” (Chen 1984:45), the famine experienced in China during and immediately after the Great Leap Forward, and

China's demographic destiny

169

the “disruption of normal married life” that occurred during this tumultuous period (Coale 1984:57). During the two years from 1961 to 1963 the total fertility rate (TFR) increased markedly, from 3.3 to 7.5 (Figure 12.1) and occurred in conjunction with the economic recovery in China (Chen 1984:45). Coale added that it also “resulted from the restoration of normal married life, from an abnormally large number of marriages, and from the unusually small fraction of married women who were infertile because of nursing a recently born infant” (1984:57). The early to mid-1960s was the period of China’s “baby boom,” in which China experienced the same kinds of growth experienced in the United States after the Second World War. China’s “baby boom” was of a shorter duration but of a significantly higher magnitude. China’s TFR peaked at 7.5 in 1963. At the height of the US “baby boom” in 1957, the TFR was but 3.7. From early 1962 to 1966, China resumed its second family planning program, mainly encouraging family size limitation. With the introduction then of the Great Proletarian Cultural Revolution, birth control work was interrupted, and the focus in China on family planning programs was minimal. The Chinese initiated in 1971 their third family planning campaign, the wan-xi-shao program, standing for the three slogans of the successful campaign: later marriages, longer intervals between children, and fewer children. Fertility fell from 5.4 in 1971 to 2.7 in 1979. However, the large numbers of children born during China’s “baby boom” in the 1960s caused Chinese leaders in the mid-to late 1970s to become concerned about demographic momentum and the concomitant growth potential of this extraordinarily large cohort. The Chinese government was “discovering the existence and usefulness of the field of demography” (Banister 1987:183), and the leaders were cognizant of the demographic momentum built into the population’s current age structure. They hence approved the “one child is best” norm and intensified their already strong family planning program by launching in 1979 the One-Child Campaign. This fourth program (which is really an extension of the third) was undertaken so that, in the words of then vice-premier Chen Muhua, “the total population of China will be controlled at about 1.2 billion by the end of the century” (Tien 1983). The principal goal of the fourth campaign is to eliminate all births over two per family, and to encourage most families to have no more than one child, especially those in the urban areas. The policy is not enforced stringently among the country’s minority populations (Poston and Shu 1987; Poston 1993), and a number of exceptions are permitted among the majority Han (Scharping 2003). The program involves a series of inducements that touch upon virtually all aspects of people’s social and economic lives, including their salaries, sustenance, health facilities, employment, and education (Sardon 1985). Between 1980 and 1982 fertility increased slightly to 2.9, and then fell back to 2.2 in 1985; by 1986 it had risen again to 2.4. The increase between 1980 and 1982 was due in part to the implementation of China’s new Marriage Law of 1981. This law raised the legal age at first marriage to 22 years for males and 20 years for females; previously, according to the Marriage Law of 1950, the legal ages for males and females were 20 and 18 years, respectively (Gu 1988). Ironically, this stipulation of the new Marriage Law of 1981 resulted in an unanticipated increase in the number of first marriages, with a corresponding increase in the fertility rate. This occurred because although the 1950

Fertility, family planning, and population policy in china

170

Marriage Law allowed males and females to marry at ages 20 and 18, most provinciallevel marriage policies required men and women to be quite a few years older, usually males at least 25, and females 23. With the passing of the new 1981 Law, many couples used these new age at first marriage stipulations as a justification for earlier marriages, and China experienced a slight marriage boom (Coale 1984; Poston 1986). By 1987 the TFR increased to 2.6. This increment was due in part to a relaxation of China’s “one child per couple” policy. The policies from the late 1980s into the 1990s were sometimes implemented more stringently, and sometimes less stringently, resulting in slight increases in the late 1980s, and then decreases in the 1990s, leading to the current TFR of around 1.8. It is the dramatic decline in the TFR since the mid-1960s and the 1970s that has resulted in many extra baby boys, compared to baby girls, being born in China every year since the 1980s. This demographic occurrence has determined China’s future. It is its destiny. As of 2001, there were still many more baby boys being born in China than baby girls, and there is no indication that this abnormal trend will end in the years of this new decade. The next section addresses the main effect of China’s fertility transition, that is, that many more baby boys are being born in China since the 1980s compared to baby girls, and its effect on China’s marriage market.

The sex ratio at birth in China Most societies have sex ratios at birth (SRBs) of around 105, that is, around 105 boys are born for every 100 girls. This so-called biologically normal level of about 105 is likely an evolutionary adaptation to the fact that females have higher survival probabilities than males. Since at every year of life males have higher age-specific death rates than females, around 105 or so males are required at birth for every 100 females for there to be about equal numbers of males and females when the groups reach the marriageable ages. Figure 12.2 shows time-series data for the SRB for China and for the United States, from 1980 to 2001. The United States shows an invariant pattern, with an SRB of about 105 for every year. This kind of stability over time at around 105 is what one should see if there were no human interventions operating to disturb the biology. In contrast, whereas in 1980 China had an SRB only slightly above 107. it began to increase in the late 1980s, reaching a value of 115 by 1990. By the year 2000, the China SRB was just under 120. The SRB reported for 2001 is 118. Since the early 1980s, China’s SRBs have been significantly above normal levels.

China's demographic destiny

171

Figure 12.2 SRB: Mainland China and the United States, 1980–2001. The sex ratio at conception (SRC) may be distinguished from the SRB. If there are no human interferences with the biological processes, the SRB depends on the factors that produce the SRC. However, if human activities such as sex selective abortion are introduced, these will influence only the SRB. What is known about these sorts of interventions? China, along with Taiwan, South Korea, and a few other countries, has been reporting abnormally high SRBs since the 1980s (Arnold and Liu 1986; Gu and Roy 1995; Goodkind 1996, 2002; Kim 1997; Poston et al. 1997; Eberstadt 2000). Research indicates that the SRBs are even higher for higher order parities (Arnold and Liu 1986; Kim 1997:20–27; Poston et al. 1997). What are the immediate causes of these abnormally high SRBs? China and the other countries just mentioned are all showing, in varying degrees, the same kinds of intervention leading to abnormally high SRBs, namely, prenatal sex identification followed by gender-specific abortion (Hull 1990; Johansson and Nygren 1991; Chu 2001; Banister 2002). But why would Chinese and some other East Asian peoples resort to an intervention that would produce higher than biologically normal SRBs? It has been mentioned above that the immediate cause is China’s dramatic fertility decline. Why would a rapid fertility reduction in China lead to abnormally high SRBs? One reason is that China (along with Taiwan and South Korea) has a Confucian patriarchal tradition where son preference is strong and pervasive (Arnold and Liu 1986; Gu and Roy 1995; Park and Cho 1995; Kim 1997; Poston et al. 1997). When birth rates are low or are on the decline, and “where a strong preference for sons over daughters is already part of the culture, SRBs tend to be higher” (Poston et al. 1997:59). As noted in the previous section, birth planning policies, as well as social, economic, and industrial transformations in China, have been responsible for the fact that the number of babies born per woman has fallen to below replacement levels, and has done so quickly (Poston 2000). Couples now have fewer children than they had just a couple of decades ago, and this is because of fertility policies and newly emerging social norms and mores. However, the deeply rooted cultural influences of son preference still make it

Fertility, family planning, and population policy in china

172

important for many families to have at least one son. Therefore, strategies and interventions are sought by many to ensure that they will have a son (Zeng et al. 1993; Gu and Roy 1995). In China since the late 1980s, ultrasound technology enabling the pre-natal determination of sex has been widely available. As of the time of this writing (January, 2005), China has proposed a ban on the practice and has launched a “pro-girl” media campaign to help mediate the strong son preference (China Daily 2004). There is little evidence of female infanticide causing the high SRBs (Zeng et al. 1993; Eberstadt 2000:228; Chu 2001; Banister 2002). The human interventions that disturb the SRB are mainly due to norms and traditions among Chinese families to have sons, within a more recent policy as well as a normative context to have fewer births. Are the reported SRBs for China the true values of the SRB? The higher than biologically normal sex ratios at birth in China are being registered in the context of a birth registration system that is far from 100 percent complete (Eberstadt 2000:228; Goodkind 2002). Given this, most of the reported SRB data for China come from other sources, including sample surveys and censuses. But the reported SRB data are often higher than SRB data for the same years collected in 945 Chinese hospitals in 29 provinces covering over 1.2 million birth records (Zeng et al. 1993; Goodkind 2002). The hospital SRB data are complete, but are they representative of the SRBs in the country as a whole? The hospitals may be reporting SRB data that are biased downwards. As Goodkind has noted, this could occur if the hospitals tended to be located in (urban) areas where son preferences are less strong and sex-selective abortion less common A second potential bias…could occur if parents choosing sex selective abortion were more likely to have their births away from hospitals. (2002:4) On the other hand, the national SRB data may be biased upwards, because of the underreporting of the births of girls. The underreporting of female births is acknowledged to occur because if parents do not report the births of girls, they are usually able to avoid the penalties that would be imposed under the one-child policies (Goodkind 2002:4). Some demographers are unsure about whether the national, often survey-based SRB data were more valid than the more limited, often urban-based hospital SRB data (Goodkind 2002). Data from the 2000 census, however, provide an SRB value for the year 2000 of 119.9 (State Council and State Statistical Bureau 2002). This value, although high, is much more consistent with the trends in the SRB based on the nationallevel data than those based on the hospital-level data. Thus it is likely that the nationallevel SRB data are more reliable. First, the SRB data based on the 2000 census, unlike the SRBs for earlier years based on surveys, are virtually 100 percent complete. Second, it is more difficult in census reporting in China to underreport (i.e. hide) a newborn baby girl than is the case in the more selective and less representative surveys. However, more research is needed from China to learn about the true levels and patterns of the SRB in China.

China's demographic destiny

173

It was noted earlier that the SRB is biologically normal at about 105 because this excess number of males pretty much guarantees near equal numbers of males and females when the groups reach the marriageable ages. For decades, the United States has had a balanced SRB of around 105 (see Figure 12.2). Figure 12.3 shows sex ratios in the United States for the year of 1999 for every five-year age group. The sex ratio is around 105 at age 0 and drops to near 100 for the ages 20–29, when men and women typically marry. Abnormally high SRBs, however, will disturb this balance. This demographic change of many more baby boys born than baby girls in China since the early 1980s is China’s demographic destiny. Starting around the year 2000, and continuing until around 2020, there will be many more extra boys of marriageable age seeking females to marry, who will be unsuccessful in their courtship pursuits. How many excess Chinese boys will there be in China who will be unable to find Chinese brides? For every year from 1978 to 2001, data on China’s total population size, the crude birth rate (CBR), and the SRB have been gathered. With these data it is easy to calculate the numbers of males and females born every year; these are shown in Figure 12.4. Using “l(x)” data from China life tables for males and for females based on 1989–1990 death data (Huang and Liu 1995: tables 2–6–1 and 2–6–2), the boys born each year are then survived to age 22, and the girls born each year to age 20. These are the minimum ages at first marriage permitted by China’s Marriage Law of 1981 (see earlier discussion). For each year starting in the year of 2000 through the year of 2021, the numbers of females survived to age 20 are

Figure 12.3 Sex ratios by age: the United States, July 1, 1999. subtracted from the number of males survived to age 22. Given an approximate 2 year difference in the ages of males and females at first marriage, each year the females age 20 will comprise the pool of potential brides for the males age 22. The numbers of marriage-age males and marriage-age females appearing in 2000 were born in 1978 (males) and in 1980 (females). It is estimated there was a surplus in 2000 of more than 341,000 males. In 2002 the male surplus is estimated to be almost 1.7 million.

Fertility, family planning, and population policy in china

174

In the years of 2011, 2012, and 2013, the numbers of excess males are estimated to be 2.3, 2.7, and 2.1 million, respectively. There are a few years (2001, 2002, 2006, and 2007) when there will actually be deficits of males. Figure 12.5 shows the numbers of excess males of marriageable age for each year from 2000 to 2021. In all, between the years of 2000 and 2021 it is estimated there will be a total surplus of males of more than 23.5 million. If the boys born in China in 1978 decide to marry when they are 22, they will seek females to marry who are aged 20 and who were born in 1980; this bride-search process will have begun in the year 2000. Based on the data shown in Figures 12.4 and 12.5, it is estimated there will be over 23.5 million surplus males looking for wives between the years of 2000 and 2021. There will not be enough Chinese women in the marriage market for them to marry.

Figure 12.4 Males and females born in Mainland China, 1980–2001.

Figure 12.5 Number of excess males at marriageable age of 22: Mainland China, 2000–2021. An alternative estimate of the number of excess marriage-age males may be obtained by using instead of the minimum legal ages for marriage, the “encouraged” ages, which are 23 for females and 25 for males (China Population Information and Research Center 2003). Extending the marriage ages by 3 years for males and females reduces minimally the number of surplus males from 23.5 million to 23.3 million. Were one to extend further the ages at first marriage to the late 20s, or even to the early 30s, these

China's demographic destiny

175

adjustments would not change appreciably the numbers of excess males because not many males, or females, die while in their 20s or early 30s. No matter to what ages the males and females survive, there will be a large excess number of marriage-age males, from 22 to more than 23.5 million, who will not be able to find women to marry who are 2 years younger than they. This will be particularly difficult because it will occur in a society where marriage is nearly universal. What will these many millions of young men do when they cannot find brides? The final section of this chapter explores some of the implications of these numbers of millions of excess males for China’s destiny.

Implications The analysis undertaken in this chapter indicates that the unbalanced sex ratios in China since the 1980s have resulted, and will result, in the existence between 2000 and 2020 of over 23 million more marriage-age males than marriage-age females. The excess males are known in Chinese as guang gun, standing for “bare branches” or “bare sticks.” Few Chinese know what will happen to the guang gun. Eberstadt (2000:230) cites an essay published in 1997 in the magazine Renmin Luntan predicting that “such sexual crimes as forced marriages, girls stolen for wives, bigamy, visiting prostitutes, rape, adultery…homosexuality…and weird sexual habits appear to be unavoidable.” Some of the implications of this excess number of young marriage-age males are now presented. Discussed first are some of the historical experiences in China and other countries at other times. What does history report about situations when earlier populations were faced with large numbers of excess males? In past centuries there are several instances of national and subnational populations having excess numbers of males. What have these countries and populations done to handle the large numbers of extra males? One result of the excess numbers of males has been a rise in the authoritarian nature of a region’s political system. This has typically occurred because of the perceived need to deal with the threat of increased violence in a society with too many unattached men. Authorities have endeavored to reduce their numbers through physical force directed at them, or instigated among them, or through some form of relocation effort (Hudson and Den Boer 2002, 2004). Authorities in the past have also recruited excess males into dangerous law enforcement and military occupations, or into large-scale public works projects, often in remote regions, both of which were characterized by higher than average mortality rates. Governments have also used extra males in the development of unexplored territories, and have encouraged surplus males to migrate to other countries. A final solution, especially relevant in those instances where the excess males were of low socioeconomic status, was for the authorities to ignore their in-group violence, and even to encourage divisions among them that led to increased violence and self-destruction (Hudson and Den Boer 2002, 2004). Among the historical cases one could cite to illustrate how instability of the social order was caused by excess males in the population are well-known episodes in China and in Portugal. During the nineteenth century in China, the Nien Rebellion in northern

Fertility, family planning, and population policy in china

176

China’s Shandong Province originated in part due to harsh social and environmental conditions. A resulting famine prompted many in the population to choose female infanticide as one way to secure their families, resulting in sex ratios averaging 129 males for every 100 females (Perry 1980). These excess males, more than 100,000 strong, finding themselves without wives and in a region entrenched in economic and social hardship, organized and turned to criminal activity. At the height of their revolution they controlled a part of Shandong Province containing 6 million people. It took the ruling Qing dynasty 17 years to overthrow them (Hudson and Den Boer 2002, 2004; Wiseman 2002:2A). Portugal, during the Middle Ages, also experienced ill effects caused by a high sex ratio influenced by primogeniture. Unlike the China case just described, Portugal’s excess males came from both low and high status groups. Government policies were affected because organized groups of these excess males generally supported uprisings against the government when they were promised valued resources. Expansionist policies were also established that relied on these males to colonize neighboring territories, with these policies sometimes being encouraged by the recruits themselves who saw opportunity in new areas. Many of the excess Portuguese males were sent off to invade North Africa (Hudson and Den Boer 2002, 2004; Wiseman 2002:2A). Evident in these and other similar initiatives was the desire to harness and/or squelch the potential energies generated by an excess number of unattached males in the society in the hope of limiting the harmful social dynamics that would have occurred. Of the various scenarios that could occur in China in the next few decades as the country’s demographic destiny caused by over 23 million excess males begins to be realized, some are more probable than others. It is not believed that China will have an easy time adapting and adjusting to their excess males, the guang gun. While it is true that throughout history, especially in Western Europe, “bachelorhood was an acceptable social role, and the incidence of never-marrying bachelors in the total population was high” (Hajnal 1965; Eberstadt 2000:230), China and its East Asian neighbors throughout their thousands of years of history have never been so characterized, and are not so today. Eberstadt may well be correct in his remark that unless it is swept by a truly radical change in cultural and social attitudes toward marriage in the next two decades,… China [is] poised to experience an increasingly intense, and perhaps desperate, competition among young men for the nation’s limited supply of brides. (2000:230) China could well turn to a more authoritarian form of government. In such a scenario, the country’s slow progress toward democracy could be stalled if not halted. If 100,000 excess males in Shandong Province were a thorn in the side of the Qing rulers for 17 years during the nineteenth century, one may only imagine the level of a rebellion that could be instigated by over 23 million Chinese bachelors. China could modify the magnitude of the potential unrest of the guang gun by dispatching them to public works projects thousands of miles away from the big cities. There are several huge construction projects currently underway in China, all of which could benefit from a young male labor force. The natural gas pipeline from the western

China's demographic destiny

177

provinces to Shanghai, the railroad to Tibet, and the Three Gorges Dam are but three projects with locations in rural and isolated areas where many of the bachelors could be sent. There is also historical precedent for dispatching Chinese to the countryside. In 1960 at the end of the Great Leap Forward (1958–1960), nearly 14 million city residents were systematically and involuntarily shipped to the rural areas and remained there for years. During the Cultural Revolution (1966–1976) many millions more young unmarried males and females, as well as intellectuals, were involuntarily sent to the countryside to work with, and “learn from,” the peasants, and their stays often lasted for a decade or more (Thurston 1987; Fairbank 1992: chapters 19 and 20). Portugal sent their extra males off to wars in North Africa. With many millions of guang gun in China’s big cities, all within 20 years of age, bellicose Chinese leaders might be tempted to “kill two birds with one stone”; they could reduce the tensions caused by the bachelors in the cities by sending this excess manpower to pick a fight with, or participate in an invasion of, another country. And what better country with which to engage in such activities than their “renegade” province, Taiwan, located less than 100 miles across the Taiwan Straits from the southern province of Fujian. There will be more bachelors in China in the next two decades than there are females in Taiwan. China is already co-opting young and poor unmarried males into the People’s Liberation Army and into the paramilitary People’s Armed Police. Indeed many of the armed personnel who participated in the crushing of the Tiananmen prodemocracy movement and rebellion in 1989 were poor, uneducated, unmarried males from isolated rural areas of China. In the next few decades there will be many millions more of such males available for these and similar kinds of activities. One solution to the problem, although deemed unlikely, would be the immigration to China of Chinese women from Hong Kong, Singapore, Indonesia, Thailand, and other countries with large numbers of Chinese people (Poston et al. 1994; Poston 2003). Chinese women from other countries would of course enlarge the pool of wives for the Chinese men in China. While there is ample evidence within China of marriage migration in which brides from provinces hundreds and thousands of miles away in-migrate to other provinces for the purpose of marriage, there is very little, if any, evidence of brides immigrating to China from other countries to marry Chinese men (Yang 1991; Xu and Ye 1992; Davin 1998; Fan and Huang 1998). Compared to the more than 30 million overseas Chinese in the diaspora, China is a poor country, and most of its guang gun are/will be poor rural workers; most of the bachelors will not be able to afford “mail order brides” (Eberstadt 2000:231). Even if this kind of marriage immigration were to occur, which is unlikely, it would need to be of a substantial magnitude to even begin to offset the gender imbalances of marriage-age males that are expected in China in the first two decades of this new century. If this immigration did occur, it would cause shortages of many millions of females in the areas of origin. Polyandry is another possibility (see Cassidy and Lee 1989). Although some might think this to be an unlikely scenario, there is limited evidence of its existence currently among some of China’s minority populations (Johnson and Zhang 1991; Zhang 1997). But it is deemed unlikely throughout China. An even less likely solution would be increases in levels of homosexuality. This is not really an alternative because most scientific evidence on the origins of homosexuality argues in favor of a biological foundation, that is, persons are born with a homosexual

Fertility, family planning, and population policy in china

178

orientation (LeVay 1991,1996; Masters et al. 1994; also see Stein 1999 and Murray 2000; Pinker 2002, for other views and arguments). It is not at all likely that when Chinese males are unable to find females to marry that as an alternative to (heterosexual) marriage they will turn to homosexual relationships. The most likely possibility is that these Chinese bachelors will never marry and will have no other choice but to develop their own lives and livelihoods. They will re-settle with one another in “bachelor ghettos” in Beijing, Shanghai, Guangzhou, Tianjin, Chengdu, and other big cities in China, where commercial sex outlets would likely be prevalent. There is also historical precedent behind this expectation. In the nineteenth century many thousands of young Chinese men immigrated to the United States to work in the gold mines and help build the railroads. When the work projects were completed, many stayed there and re-settled in Chinese bachelor ghetto areas in New York, San Francisco, and a few other large US cities (Lee 1960; Kwong 1987; Zhou 1992). The sex ratios of the Chinese in these areas were extraordinarily high. In 1850 the sex ratio of the Chinese living in San Francisco was almost 40,000 (787 males and a mere 2 females) (Tsai 1986:2). In 1900, the then very small Chinatown in New York City had a sex ratio of over 11,000 (about 4,000 men and only 36 women). In the same year, the state of New York had a Chinese sex ratio of almost 5,000 (7,028 men and only 142 women) (Zhou 1992:34, 44). If these men do not marry, research suggests that they will be more prone to crime than if they were married (Sampson and Laub 1990; Horney et al. 1995; Mazur and Michalek 1998). This possibility has alerted some to the potential increases in crime in China’s future, and perhaps political ramifications, resulting from these excess males (Hudson and Den Boer 2002, 2004). This potential for high levels of criminality is based on criminological research and historical insights. The research has shown that banditry, violence, and revolutions are likely to occur in areas with large numbers of excess males (Hudson and Den Boer 2002). This is a real implication of China’s demographic destiny. No one, of course, knows what this excess number of young Chinese males will do. Several possibilities have been entertained. The only fact known for certain is that there have already been born, in China, over 23 million more baby Chinese boys than there will be Chinese girls for them to marry. These “bare branches,” the guang gun, are China’s demographic destiny. Their presence is now beginning to be felt in China’s marriage market and will continue to be felt, at least until 2020, perhaps longer.

References Aird, J.S. 1972. “Population Policy and Demographic Prospects in the People’s Republic of China.” In People’s Republic of China: An Economic Assessment, A Compendium of Papers Submitted to the Joint Economic Committee, Congress of the United States, 18 May (92nd Congress, 2nd Session). Washington, DC: US Government Printing Office. Arnold, F. and Z.Liu. 1986. “Sex Preference, Fertility, and Family planning in China.” Population and Development Review, 12:221–246. Ashton, B., K.Hill, A.Piazza, and R.Zeitz. 1984. “Famine in China, 1958–61.” Population and Development Review, 10:613–645. Banister, J. 1987. China’s Changing Population. Stanford, CA: Stanford University Press.

China's demographic destiny

179

Banister, J. “Recent population trends and issues in China.” Paper presented at the annual meetings of the Australian Population Association. Sydney, October 2002. Cassidy, M.L. and G.R.Lee. 1989. “The Study of Polyandry: A Critique and Synthesis.” Journal of Comparative Family Studies, 20:1–11. Chen, S. 1984. “Fertility of Women during the 42-Year Period from 1940 to 1981.” pp. 32–58 in China Population Information Center, Analysis on China’s National One-per-Thousand Sampling Survey. Beijing: China Population Information Center. China Daily. 2004. “China Bans Selective Abortion to Fix Imbalance.” http://www.%20chinadaily.com.cn/english/doc/2004–07/16/content_349051.htm (accessed 1– 11–05). China Population Information and Research Center. 2003. “Encouraged Ages for First Marriages.” http://www.cpirc.oig.cn/policyfp.htm (in Chinese) (accessed 1–30–2003). Chu, J. 2001. “Prenatal Sex Determination and Sex-selective Abortion in Rural Central China,” Population and Development Review, 27:259–281. Coale, A.J. 1984. Rapid Population Change in China, 1952–1982. Washington DC: National Academy Press. Davin, D. 1998. Internal Migration in Contemporary China. Basingstoke, England: MacMillan Press. Eberstadt, N. 2000. Prosperous Paupers & Other Population Problems. New Brunswick, NJ: Transaction Publishers. Fairbank, J.K. 1992. China: A New History. Cambridge, MA: The Belknap Press of Harvard University Press. Fan, C.C. and Y.Huang. 1998. “Waves of Rural Brides: Female Marriage Migration in China,” Annals of the Association of American Geographers, 88:227–251. Goodkind, D. 1996. “On Substituting Sex Preference Strategies in East Asia: Does Prenatal Sex Selection Reduce Postnatal Discrimination?” Population Research and Policy Review, 22:111– 125. Goodkind, D. “Recent trends in the sex ratio at birth in East Asia,” Paper presented at Conference on Chinese Populations and Socioeconomic Studies: Utilizing the 2000/2001 Round Census Data, Hong Kong, June 2002. Gu, J. 1988. “Marriage and the Family.” pp. 129–149 in China Financial and Economic Publishing House, New China’s Population. New York: Macmillan. Gu, B. and K.Roy. 1995. “Sex Ratio at Birth in China, with Reference to Other Areas in East Asia: What we know.” Asia-Pacific Population Journal, 10(3):17–42. Hajnal, J. 1965. “European Marriage Patterns in Perspective.” Pp. 101–143 in D.V.Glass and D.E.C.Eversley (eds), Population in History. London: Edward Arnold. Horney, J., D.W.Osgood, and I.H.Marshall. 1995. “Criminal Careers in the Short-term: IntraIndividual Variability in Crime and its Relation to Local Life Circumstance.” American Sociological Review, 60:655–673. Huang, R. and Y.Liu. 1995. Mortality Data of China. Beijing: China Population and Information Research Center. Hudson, V.M. and A.Den Boer. 2002. “A Surplus of Men, a Deficit of Peace.” International Security, 26:5–38. Hudson, V.M. and A.M.Den Boer. 2004. Bare Branches: Security Implications of Asia ‘s Surplus Male Population. Cambridge, MA: The MIT Press. Hull, T.H. 1990. “Recent Trends in Sex Ratios at Birth in China.” Population and Development Review, 16:63–83. Johnson, N.E. and K.Zhang. 1991. “Matriarchy, Polyandry, and Fertility Amongst the Mosuos in China.” Journal of Biosocial Science, 23:499–505. Johansson, S. and O.Nygren. 1991. “The Missing Girls of China: A New Demographic Account” Population and Development Review, 17:35–51.

Fertility, family planning, and population policy in china

180

Kim, D. 1997. “The Pattern of Changing Trends and the Regional Difference in the Sex Ratio at Birth: Evidence from Korea and Jilin Province, China.” Korea Journal of Population and Development, 26:19–24. Kwong, P. 1987. The New Chinatown. New York: Hill and Wang. Lee, R.H. 1960. The Chinese in the United States of America. Hong Kong: Hong Kong University Press. LeVay, S. 1991. “A Difference in Hypothalamic Structure Between Heterosexual and Homosexual Men.” Science, 252:1034–1037. LeVay, S. 1996. Queer Science: The Use and Abuse of Research Into Homosexuality. Cambridge, MA: MIT Press. Masters, W.H., V.E.Johnson, and R.C.Kolodny. 1994. Heterosexuality. New York: HarperCollins Publishers. Mazur, A. and J.Michalek. 1998. “Marriage, Divorce, and Male Testosterone.” Social Forces, 77:315–330. Murray, S.O. 2000. Homosexualities. Chicago, IL: University of Chicago Press. Orleans, L.A. 1972. Every Fifth Child: The Population of China. Stanford, CA: Stanford University Press. Park, C.B. and N.Cho. 1995. “Consequences of Son Preference in a Low-Fertility Society: Imbalance of the Sex Ratio at Birth in Korea.” Population and Development Review, 21:59–84. Perry, E.J. 1980. Rebels and Revolutionaries in North China, 1845–1945. Stanford, CA: Stanford University Press. Pinker, S. 2002. The Blank Slate: The Modern Denial of Human Nature. New York: Viking Press. Poston, D.L., Jr. 1986. “Patterns of Contraceptive Use in China.” Studies in Family Planning, 17:217–227. Poston, D.L., Jr. 1993. “The Minority Nationalities of China.” pp. 483–495 in Proceedings of the International Population Conference, Montreal, 1993. Liege, Belgium: International Union for the Scientific Study of Population. Poston, D.L., Jr. 2000. “Social and Economic Development and the Fertility Transitions in Mainland China and China.” Population and Development Review, 26 (Supplement): 40–60. Poston, D.L., Jr. 2003. “Overseas Chinese.” pp. 130–133 in G.McNicoll and P.Demeny (eds), Encyclopedia of Population. New York: Academic Press. Poston, D.L., Jr and J.Shu. 1987. “The Demographic and Socioeconomic Composition of China’s Ethnic Minorities.” Population and Development Review, 13:703–722. Poston, D.L., Jr, B.Gu, P.Liu, and T.McDaniel. 1997. “Son Preference and the Sex Ratio at Birth in China.” Social Biology, 44:55–76. Poston, D.L., Jr, M.Mao, and M.Yu. 1994. “The Global Distribution of the Overseas Chinese Around 1990.” Population and Development Review, 20:631–645. Sampson, R.J. and J.H.Laub. 1990. “Crime and Deviance Over the Life Course: The Salience of Adult Social Bonds.” American Sociological Review, 55:609–627. Sardon, J. 1985. “Planification familiale et pratiques contraceptives en Chine.” Population, 40:774– 779. Scharping, T. 2003. Birth Control in China, 1949–2000. London: Routledge Curzon. State Council and State Statistical Bureau. 2002. Population Census Office Under the State Council and Department of Population Statistics, State Statistical Bureau. Tabulation on the 2000 Population Census of the People’s Republic of China. Beijing: China Statistical Publishing House. Stein, E. 1999. The Mismeasure of Desire: The Science, Theory, and Ethics of Sexual Orientation. New York: Oxford University Press. Thompson, K. 1976. Auguste Comte: The Foundation of Sociology. London: Thomas Nelson and Sons Ltd. Thurston, A.F. 1987. Enemies of the People: The Ordeal of the Intellectuals in China’s Great Cultural Revolution. New York: Alfred A.Knopf.

China's demographic destiny

181

Tien, H.Y. 1983. “China: Demographic Billionaire.” Population Bulletin, 38(2):1–42. Tsai, S. 1986. The Chinese Experience in America. Bloomington, IN: Indiana University Press. Wiseman, P. 2002. “China Thrown Off Balance As Boys Outnumber Girls.” USA Today, (June 19): 1A-2A. Xu, T. and Z.Ye. 1992. “Analysis of Female Immigrants in Zhejiang.” Renkou Xuekan (Population Journal), 2:45–48 (in Chinese). Yang, Q. 1991. “The Phenomenon of Southern Women Marrying to the North and its Advantages and Disadvantages.” Renkou Xuekan (Population Journal), 5:51–55 (in Chinese). Zeng, Y, P.Tu, B.Gu, Y.Xu, B.Li, and Y.Li. 1993. “Causes and Implications of the Recent Increase in the Reported Sex Ratio at Birth in China.” Population and Development Review, 19:283–302. Zhang, T. 1997. “Marriage and Family Patterns in Tibet.” China Population Today, 14: 9–12. Zhou, M. 1992. Chinatown: The Socioeconomic Potential of an Urban Enclave. Philadelphia, PA: Temple University Press.

Index Note: Page numbers in italic indicate illustrations. 1982 China One-Per-Thousand National Survey 41 1990 Census of China: micro-data 65 1990 Census of China: one percent sample of the 69 1996 Youth Sexuality Survey 80 1997 Sample Survey of Population and Reproductive Health 42, 78, 82 Abortion: gender-specific 176; hazard of having 27, 35; and miscarriage 86; patterns of 25; sex-selective 23, 26 Abortion issues in China 23 Abortion law in China 24 Abortion rates: areas 26 Actual fertility: declining pattern of 91 Adaptation hypothesis 141; fertility and migration 129; fertility of floating migrants 131 Affinity: dialect and fertility 148, 150, 153 Age at marriage: first birth 113–117, 118–119; lowering of the 163; menarche 115; premarital conception 81 Age of marriage: premarital conception 84 Altaic language family 67 Attitude: premarital conception 84 Average ideal number of children 90 Baby boom 11, 160; China’s 174 Baby-boomlet 16 Bachelor ghettos 183

Index Bare branches 180, 184 Barefoot doctor 39 Below-replacement: levels 177; rate 159 Biological predictors 114 Biological variables 125 Birth control 25; assessment of the effect 165; campaigns, government machinery in 18, short-and long-term 13; compulsory 166; compulsory, one-child policy 15; effects of 159; monitoring of compulsory 168; policy, Chinese 106, sterilization 38 Birth planning policy 42; Birth rate rebounded 169; decentralization of the 163; urban-rural differential 160 Birth-quota violation 40 Birth rates: India and China 167 Birth records: manipulation of 162 Book of Rites 25 Caregivers 54 Carolina Population Center 56 Categorical independent variable 28 CBR 178; dialects and fertility 154; good predictor for 153; see also crude birth rate CEB: effect of intermarriage 70; frequency distribution of 132, 133; Poisson regression 136, 139; question 132; sterilization hazard 47; see also children ever born Chen Muhua 10, 17 Child-bearing guerrillas 127, 131, 140, 142 Child birth plan: one-son-two-child policy 163 Childcare givers: future 61 Children ever born: floating migration 131; minority groups 69

183

Index

184

China Academy of Preventive Medicine 56 China and the United States: SRB 176 China Health and Nutrition Survey 54 China Marriage Law of 1981 79 China’s 1988 Two-Per-Thousand Fertility Survey 42 China’s ethnic minorities 66–68 Chinese bachelors 182, 183 Chinese Communist Party 9 Chinese dialectology 148 Chinese dialects 145, 149; coding of 150; grouping of 148 Chinese-Tibetan language system 67 “Coming out” parties 113, 125 Committee on Planned Births 11 Comparisons: rural and urban fertility 97 Compensatory births 11, 17 Comte, Auguste 172 Confucian: patriarchal tradition 176; tradition 105, 106 Confucius 54 Contraception: knowledge of 48; propogation of 9 Contraceptive: choice 41, counseling 39; devices 11; effect of education 42; methods 39; patterns 41; prevalence 38; use 90 Control policy 169; compulsory population 170 Control variables 31; fertility of floating migrants 134 Coresidence 59, 62, 63 County level data: dialects and fertility 153 Cox proportional hazard analysis 118–121; ever-married, Han females 122, minority females 123 Cox proportional hazard: estimates 46; models 28, 57, 59, 60 Cross-sectional data 132 Crude birth rate 11, 12, 148, 160–161 Crude death rate 12

Index Cultural Revolution 121, 182; proletarian 174; political turmoil of the 166 Daughter-only couples 40 Daughter preference 103 Deficits: of males 179 Demographic behavior 166, 167 Demographic change 178 Demographic data 27 Demographic destiny 183; China’s 172–186, 178, 183, 184 Demographic echo 16 Demographic echo-effect 18 Demographic occurrence 175 Demographic processes 172 Demographic transition 91, 172 Demography: fertility 89 Deng Xiaoping 9 Designated cities 97 Designated towns 97 Desired fertility: changing patterns of 89–109; meta-analysis 90, 92; within-region difference 93 Desired gender 92; changing patterns of 99–105; outcome 99 Desired sex ratio 103 Desired total fertility rate 90 Dialect 146 Diffusion: language and fertility 146; mass media 147 Directive: State Council 11 Disruption: effect, fertility of floating migrants 131; hypothesis, fertility and migration 129 Doctrinaire Marxism 173 Economic transformation 170 Education: hazard ratio of 47; status, distributions of 136 Educational attainment 30 Endogamous women 75; CEB 72 Enlai, Zhou 14

185

Index Ethnic assimilation 76 Ethnic fertility: intermarriage 72 Ethnic groups: fertility patterns among different 65; fertility scores 71 Ethnicity 31; premarital conception 81 Ethnic minorities 16 Ethnic relations: review of 66 Ever-married Han women 117 Ever-married women 56; 1991 survey 60; descriptive data for 71; fertility of floating migrants 134 Excess males 179, 180, 181, 182, 183 Exogamous women 75; CEB 72 Expansionist policies 181 Extended family 61, 62; structure 60; Patrilineal and Patrilocal 54 Family: modified extended 53, 56, 59, 61 Family planning: policies, sterilization 38; policy makers 48; programs 91, social issues 23, later-longer-fewer 89; quality of service 41; quality variables, sterilization hazard 46; three keys 16; voluntary 170 Family structure: measures of 59 Female infanticide 177, 181 Fertility 44; age at menarche and 115; changes, in Western countries 106; controlled 117; data, longitudinal 131; decision 62; decline 63, 174, China’s dramatic 176, policy-induced 170, Shanghai 151; the effect of floating migration on 127–144; English proficiency and 146;

186

Index Fertility 159; feminist paradigms of 81; the impact of dialect on 145–155; indicator 142; intermarriage and 65; levels, migrants’ 129; mass media effects on 150; measures 53; and migration, theories dealing with 128; modification 26; negative effects on 75; paradigms 80; patterns, city 163, European 145, declining desired 106; planning, referendum 14; practices, diffusion of 147; precipitous decline in 160; preferences 90, study of 165; rate 169, decline of the 162; schedule; studies, prior analyses 90–92; Survey 1982 One-Per-Thousand 26; theories 114; for townships 163; tradition 93; transition 18, 172, China’s 175, managed 159–171, policy induced 14, process of China’s 159, process of the 162, rural Chinese 90 Fetal loss 117 First birth: age at menarche 117, 124; hazard of 114, 119; intervals between marriage and 80; menarche and 115; transition to 56, 62; woman month risk of 61 First birth control campaign 24 First Five-Year Plan 9 First intercourse: age at 114 First marriage: minimum ages at 178 First-married women: sterilization 43 Floaters 71, 136, 137, 138;

187

Index CEB 131; childbearing of the 132; rural-to-urban 139 Floating migrants 127–128, 132, 133, 142; fertility of 138; origin and destination of 134 Four-level urban system 97 Fourth campaign 174 French fertility: Watkins’s work on 146 Gang of Four 14 Gender 30; composition of children 99; expectation pattern 101; preference 26, 105 Gonadotropin-releasing hormone 116 Goodkind, D. 176, 177 Granny police 38 Great Cultural Revolution 18, 160; Mao 14 Great Leap Forward 10, 160, 173, 174, 182 Guang gun 180, 181, 182, 184 Han Dynasty 66 Han majority 41 Han woman 56; hazard, models 31–34, of first birth 118–19, regression 121–4 Hazard: analysis 28; function 28; model 27, 31, 56, dependent variable 29–30; independent variables 30–2; ratios, semi-standardized 34; sterilization hazard 47; regressions 121 Homosexuality: levels of 183 Household: nature 134; registration 130, 133–135, 138, system 127–128; boundary of 54 Huang Hong 127 Hui 67, 69 India: China and 167;

188

Index development indices of China and 167; population growth rates of China and 168 Induced abortion 11, 23, 24; means of preventing 47 Infant mortality rates 167 Information-based choice: components of 48; program 48 Informed choices 41, 48 Inner Mongolian Autonomous Region: family planning policy 69 Interethnic marriages 67 Intergenerational ties 55, 63 Intermarriage: coefficients of 73; implications on fertility 74 Internal migrants 145 Intrauterine device 38; see IUD Jin dialect 151 Just-replacement level 164 Kaplan-Meier: Survival Estimates, Han women 119, minority women 120; Survival Curves 44; survivor function 28 Language variables: dialect and fertility 147 Late, sparse, and few 18 Life course analysis 115 Linguistic boundaries 145 Liu Shaoqi, President 9 Local-birth planning units 39 Logistic regression: premarital conception 83 Longitudinal survey 57 Low fertility: below-replacement level of 160 Lu’s quantified indices 149 Luteinizing hormone 116 Mail order brides 183 Major ethnic groups 66 Malthusians 173 Malthusian thought 11 Manchu 67, 69, 117; fertility scores 71

189

Index Mandarin 151–154; dialects and fertility 147 Mao Zedong 173; Chairman Mao 10 Marriage-age males: excess 180; gender imbalances of 183 Marriage immigration 183; Law, China’s 27, 179; market, China’s 172, 175; migration 183; minimum age of 17; policies, provincial-level 175 Mass media: index of exposure to 147 Matrilocal 59, 61, 62 Ma Yinchu 10 Media campaign: pro-girl 177 Menarche 116, 117; age at 125; average age at 114; rites of passage 113 Menstrual cycle length 117 Miao 67 Middle Ages 181 Migrant: categories, fertility 131; fertility 128; status variables 140 Migrants 71, 137; temporary and permanent 130 Migration: data, longitudinal fertility and 132; status 71, demographic variables by 131, fertility of floating migrants 133 Minorities: family planning policies for 68–69 Minority: groups: assimilation 67; fertility patterns of 65; nationalities, China’s 76; population, China’s 66, demography of China’s 65; women, contraceptive patterns 41, fertility of 65–77, hazard of first birth 118–119 hazard regression 121, six hazard models 124–125, urban 74

190

Index Model diagnostics 151 Mongolians 67, 69; fertility scores 71 Moslem 67 Motherhood marriage and 115 Mother substitute 59 Multistage cluster design 59 National census 9 National Committee on Planned Births 16 National Committee on Planned fertility 159 National hard times 174 National SRB data 177 Natural fertility 160 Nested hazard models 31 New “marriage law” 163, 175 Nien rebellion 181 Nonmarital conception 78 Nonrnigrants 71, 137, 140 Nuclearization 55 Null hypotheses 74 One-boy-two-child 165, 169 One-child fertility policy: enforcement of the 128 one-child policy 8, 14, 163; desired fertility 106; different enforcement of 117; impact of 168–169; minorities 68; number of abortions 24; period after the 89; premarital conception 81, 84; radical 18; son preference 26; survey data 82; transition to first birth 120 One-child population policy: spacing between children 25 One-vote-down campaign 40 Only-girl families 16 Ordered affinity index: dialects and fertility 149 Ordinary Least Squares regression: dialects and fertility 151 Outliers 151 Parents and children 55 Passive fecundity 129 Patrilineal 62 Patrilocal 58, 59, 61, 62

191

Index People’s Congress, fifth 15 People’s Daily 18 Per capita calorie-intake levels: India and China 167 Permanent migrants 127, 136, 138, 140 Planned-birth policy 13 Planned births: issue of 14; management of 170 Planned reproduction 11 Poisson regression: fertility of floating migrants 133, 136–41 Poisson regression model 74, 76; CEB 73; effects of intermarriage 69 Policy: effects, comparative analysis of 166–9; implications 86; violation, sterilization as a sanction 40 Polyandry 183 Population: control, directive on the issue of 10, 17, white paper on 18; control policy 16, 17, see also control policy; first official policy on 10; policies, China’s 24, development of 8; policy, hazard models 34, assignments of China’s 159, fertility and 8–19, one couple, one child 23, personal characteristics 34; reproduction 10; tripling 166 Portugal 181, 182 Post-Cold War era 19 Pregnancies: outside marriage 26; unauthorized 25 Premarital conception 86; emerging patterns of 78–88; operationalization 82–83; prior literature 79–80 Primary sex characteristics 116 Primogeniture 181 Provider’s recommendation 41 Proximate determinants paradigm 114 Pseudo R-squared statistic 34 Qing 24, 181, 182

192

Index

193

Quasi-coresidence 54, 56, 58, 59; rates, coresidence and 60 Quinceañera 113 Quinceañeras 125 Radical natural disasters 11 Reagan administration 16 Remarried women: sterilization 43 Replacement level 162 Reproduction 116 Reproductive health strategy 48 Residence of a woman: hazard model 34 Residential patterns: parents and parents-in-law 58 Residential registration 170 Respondents: similarities and differences among 103 Risk population 42 Rites of passage 113 Rural areas: desired gender of children 94, 100; premarital conception 84 Rural births: managed 161–162 Rural China 102; desired fertility 92–96; desired gender of next child 102; fertility trends in 91 Rural Chinese 15 Rural fertility: policy induced decline in 160 Rural nonmigrants 139 Rural women: intermarriage 76 Rural-to-urban floaters: fertility of 140 Russians: descendants of 69 Sample Survey of Population and Reproductive Health 27, 117 Second birth: floating migration 130 Selectivity and disruption hypotheses 139, 140, 142 Selectivity hypothesis: fertility, and migration 128–129, of floating migrants 130 Sex education 84; effects of 81 Sex identification:

Index prenatal 176 Sex preference 26 Sex ratio 183; at birth 26; see also SRB Sex ratio at birth: abnormal 23 Sex ratio at conception 176 Sex ratios: unbalanced 180 Sex ratios at birth 176 Sexual revolution 81 Shao Lize 9 Silent sexual revolution 78, 79, 80 Social demographic characteristics: premarital conception 80 Socialization and adaptation hypotheses 139, 142 Socialization hypothesis 141; fertility and migration 129 Social order: instability of the 181 Social taboos: abortion 26 Socio-demographic characteristics: sterilization hazard 45 Socioeconomic factors 167; China’s fertility 91 Song Dandan 127 Son preference 91, 101, 103, 105, 177; abortion 25 Spousal separation 129 SRB 175, 179; abnormally high 177; see also sex ratio at birth Standardized language: dialect and fertility 152 State Committee on Fertility Planning 16 Sterilization 11; discussion and implications 47; factor predicting 42; hazard of 44; patterns of 38; in urban and rural China 40–42 Steroids 116 Stratified models 28 Subfecundity 116 Subfecund periods 117 Survey design 96 Survival: analysis 27; probabilities 29

194

Index

Television: language and fertility 147 Television sets: dialects and fertility 150, 152, 154 The New Population Theory 10 Tiananmen 182 Tibetan—Myanmese language 67 Tibetans 67; family planning policy 69; fertility scores 71 Time of pregnancy 31 Total fertility rate 12, 13; see also TFR Transition process 160 Transition to first birth 120 Tujia 66 Underreported births: extent of 162 Underreported rate: rural births 169 Underreporting: of births 90 United Nations 14 United States: SRB 175 Univariate Poisson distribution 70 Urban areas: desired gender of children 104, 105; desired number of children 98 Urban nonmigrants 139; fertility 131 Urban respondents: gender preference 103 Urban—rural status 139 Urban survey: pattern of desired fertility 97–99 Urban women: intermarriage 76 Uygurs: fertility scores 71 Violations: threats of penalties 40 Voluntary client choice 41 Voluntary planning: India 166 Wage jobs 55

195

Index Wan, xi, shao 13, 24, 39, 162, 174 Wanted total fertility rate 90; see also WTFR Watkins, Susan Cotts 145 Wealth flows theory: Caldwell’s 106 White couples: Mexican-American and non-Hispanic 146 Women-months 59 Woman’s residence 31 Women’s employment status 153 Women’s labor force 55 Women’s migration status 134 World Trade Organization 19 Wu dialect 151 Xinjiang Uygur Autonomous Region 67 Yanbian Korean Autonomous Prefective 67 Yi 67; Yi-piao-fou-Jue 40 Zhao Ziyang 15 Zhuang 66

196