Gender Imbalance in China

1
Gender Imbalance in China

• Marc Feldman
• Morrison Institute for Population and Resource
  Studies
• Department of Biology
• Stanford University

For APARC October 2, 2008
2
Collaborators Professors Li Shuzhuo Jin
Xiaoyi Du Haifeng Institute for Population
and Development Studies Xian Jiaotong
University, China Together with many students
who participated in running the surveys.
3
(No Transcript)
4
(No Transcript)
5
(No Transcript)
6
(No Transcript)
7
Differences between provinces through census years
Figure 4 SRB by province in 1982, 1990, 2000 and
2005 Source Tabulation of population censuses in
1982, 1990, 2000 and 1 population survey in 2005
8
(No Transcript)
9
(No Transcript)
10
(No Transcript)
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
Excess girl child mortality
15
Excess girl child mortality
16
Missing women Percentages of missing females
during 19002000
17
DATA AND METHODS Surveys(1) Survey sites
Figure 1 The three counties in China
18
Table 1 Information about the three counties
Survey Sites Sanyuan, Shaanxi Lueyang, Shaanxi Songzi, Hubei
Location Guanzhong plain, close to Xian City the Qing mountains Jianghan plain
Products Wheat Woods and wild products Rice and cotton
Transportations Convenient Inconvenient Convenient
Patrilineal family system Strict Relaxed Relaxed
Marriage form Virilocal marriage dominant Diversified Diversified
Development Medium-developed Underdeveloped Well-developed
Population 400,000 Relatively high fertility 200,000 Medium fertility 900,000 Low fertility
SourceSanyuan County Gazette, Lueyang County
Gazette, Songzi County Gazette
19

• 1997, Cultural transmission of son preference,
  Sanyuan and Lueyang, Shaanxi province
• 2000, Marriage form and old-age support, Songzi
  county, Hubei province.

Table 2 The composition of the surveys
Survey Sites Sampling Surveys (Household questionnaire) Sampling Surveys (Community questionnaire) In-Depth Interviews Focus-Group Discussions
Sanyuan, Shaanxi Province ? ?
Lueyang, Shaanxi Province ? ? ?
Songzi, Hubei Province ? (with information about intergenerational transfers) ? ? ?
20
Figure 6. With whom do you want to live? Source
Li S. et al. (1998).
21
Figure 7. Distribution of benefits of having a
son. Source Li S. et al. (1998).
22
Figure 9. Estimated transmission rates of no son
preference by womens age group. Source Li N. et
al. (2000b).
23
STATUS OF UXORILOCAL MARRIAGE
Figure 1 Percentage of uxorilocal marriage and
its trends in the three counties() Source The
surveys conducted in Sanyuan and Lueyang,
Shaanxi, 1997, and in Songzi, Hubei, 2000.
24
Figure 10. Simulation of future Chinese sex
ratio at birth. Source Li N. et al. (2000b).
25
(No Transcript)
26
SRB (sex ratio at birth). Bias to males. EFCM
(excess female child mortality. Causes
Proximal (a) Sex-selective abortion (b)
Underreporting of female infants (c)
Infanticide Chu (2001) 12
villages in rural China (central area)
427 male foetuses 1.6
aborted 279 female
foetuses 25 aborted (d) Differences in
nutrition and medical care Conditional
Fertility reduction Poor social
security General gender discrimination
Fundamental Patrilineal family systems
producing public and
private patriarchy. Confucian culture
family name
property rights living arrangements

27

• Demographic implications
• If SRB remains at 2000 level,
• EFCM remains at current level,
• TFR continues at 1.61.7 (20012004 level).
• Then
• By 2030 population of China will be 84.2 of what
  would be expected at this TFR.
• Excess males 2021.
• Aging of population becomes very serious.
• Marriage squeeze intensifies.

28
Policies of the Central Government in Reaction to
High SRB, EFCM (a) Proximal Stronger
enforcement against The Two
Illegalities. (1) Non-medical sex
identification (2) Sex-selective abortion.
(As per laws and local regulations)
Note Average cost for (1) RMB 300600
yuan (2) RMB 500800 yuan
Stronger punishments suggested at 2008 NPC and
CPPCC.
29

• (b) Government actions aimed at conditional and
  fundamental causes
• Care for Girls program
• 2000 Chaohu (Anhui) Experimental Zone for
  Improving Girl Child
  Survival.
• Financial help for one child and two daughter
  families.
• Fees for education of girls and increased
  pension.
• Promotion of uxorilocal marriage.
• Result SRB went from 125 in 1999 to 114 in
  2002.

30
(b) Government actions (continued) 20032005 Ext
ension of Care for Girls to 24 counties in 24
provinces with high SRB. Preferential Reward
Policies Result in these counties average SRB
dropped from 133.8 in 2000 to 119.6 in
2005. January 2006July 2006. Stipulation and
Initiation of nationalCare for Girls campaign
Aim To bring SRB to normal in 15
years January 2008. NPFPC launched Care for
Girls Youth Volunteer Action pilot for
20082010. More than 1,000 recruits (mostly
university students) to engage in promotional
activities and some data collection (under
Chinese Communist Youth League).
31
Le Bin, Minister of NPFPC, 2008 NPC,
CPPCC We must firmly fight against sex
identification tests and artificial termination
of pregnancy or sex selection. We need to
intensify the construction of a new reproductive
culture and widely promote state policies
including family planning and gender equality for
the purpose of creating a favorable atmosphere of
public opinion on caring about girls and
comprehensively addressing the gender imbalance.
32

• Floating Population (rural-urban migrants)
• 2008 NPC and CPPCC Le Bin Ministry of NPFPC
• We must study the distribution of the floating
  population.
• We must come up with plans concerning registered
  permanent residence for floating populations.
• We must strengthen family planning services for
  floating population.
• 100150 million people

33
Migrants and Networks

• Statistical analysis
• Effects of social network on son preference
• Effects of care-givers out-migration on old-age
  support
• Descriptive statistics of whole networks
• Fit to models
• Small-world phenomena
• Scale-free properties
• Community structure

34
1. Statistical analysis
1.1 Effects of social network on son preference
Analysis Framework
35
Models
Dependent variables Attitude towards son preference (Cumulative logistic models) What will you do if your first child is a girl? Stop childbearing-Without son preference (Ref.) Have a second child and stop- Weak son preference Have more children until have a boy- Strong son preference 2. Behavior of son preference (Binary logistic models) The gender of second birth after migration Give birth to a girl (Ref.) Give birth to a boy
Independent variable Social network factors Overall effect of childbearing discussion network members Weak ties of childbearing discussion network
Control variables Migration experiences Age at first migration (numerical), Times back home per year Years of living in urban areas (1)Years from first migration to survey year -attitude (2)Years from first migration to year 2nd child was born- behavior
Control variables Individual characteristics Gender Marital status, Education, Sex composition of ever born children, Residence region before migration
36
Descriptive Results

• (1) Attitude of son preference after migration
• Only a minority of rural-urban migrants
  report strong son preference after migration.

Definition Percentage()
No son preference 37.1
Weak son preference 55.9
Strong son preference 7.0
Sample size 1739
37

• (2) Behavior of son preference after migration
• Sex ratio of migrants children born after
  migration

38

• Regression results
• (1) Attitude of son preference (Model 1 and 2)
• The risk of having son preference tends to
  decrease when the overall effect of network
  member is positive (without son preference) and
  in presence of weak ties
• The older the migrants at first migration, the
  higher the likelihood of having son preference
• Duration of living in cities, and education, have
  negative effects on the risk of having son
  preference.
• The odds ratio of having son preference is lower
  for migrants who only have sons.
• Those migrating from central and western regions
  are less likely to have son preference compared
  with those migrating from eastern regions.
• (2) Behavior of son preference (Model 3 and 4)
• The odds ratio of having a boy at second birth
  tends to decrease when the overall effect of
  network members is positive (without son
  preference).
• Increase in age at first migration decreases the
  risk of having a boy.
• Compared with those living in urban areas for no
  more than one year, the risk of having a boy is
  lower among those living in urban areas for 8
  years and above.
• Migrants with a higher educational level are
  more likely to have a boy at second birth.
• The likelihood of having a boy is 9.836 (e2.286)
  times greater for migrants whose first child is a
  girl than for those whose first child is a boy.

39
Conclusions

• The majority of rural migrants report some son
  preference, but the proportion of the migrants
  reporting strong son preference is very low.
• The sex ratios of the migrants children are
  significantly higher than normal, increasing with
  birth order. The childbearing behavior of these
  migrants exhibits strong son preference.
• The changes in their expression and behavior of
  son preference are partly driven by interpersonal
  influences.
• Age at first migration and sex configuration of
  ever-born children are important factors
  influencing son preference.
• With more years of living in urban areas, the
  attitude of son preference tends to be weaker, as
  does the behavior of son preference. Individuals
  attitudes and behaviors of son preference are
  influenced by the period effect.
• The change in childbearing behavior lags far
  behind the change in expressed attitudes.

40
1. Statistical analysis
1.2 Effects of care-givers out-migration on
old-age support
Analysis Framework
41

• To analyze the effect of childrens migration on
  financial support to parents and parents-in-law,
  samples were restricted to those married before
  migration
• Since married children coresiding with their
  parents usually share the same household economy
  with their parents in rural China, samples were
  also restricted to those did not coreside with
  their parents.

 
42
Models
Dependent variables Whether the amount of financial support to parents increased after migration (Logistic Regression Model) 2. Amount of financial support to parents after migration (OLS Regression Model)
Independent variable Gender of out-migrating children
Control variables Migration experiences Years since first out-migration, times back home per year
Control variables Individual characteristics Age, education, number of offspring, spouse living in home town, income, spouses income, effect of current financial help to one set of parents on other set of parents
Control variables Parents characteristics Survival status, age, physical status, co-residing with individuals children, coresiding with other married children, main source of income, current financial help to child
43

• Gender difference in the mean increment of
  financial help
• Significant in financial help to
  parents-in-law
• Not significant in financial help to
  parents
• Females increase financial help to two
  sets of parents while males only increase
  financial help to natal parents.
• Difference in financial help to parents and
  parents-in-law
• Males Financial help before migration
  is higher for parents-in-law than for natal
  parents, but it reverses after migration
• Females Financial help is always much
  higher for parents-in-law than for natal parents
  before and after migration, but it tends to be
  reduced after migration.

44

• Likelihood of increasing the amount of financial
  support after migration is affected by
• Gender
• Female migrants are more likely to
  increase the amount of financial support to their
  parents-in-law after migration
• Migration experience
• Longer duration of out-migration helps
  raise the likelihood of increasing the amount of
  financial support to parents, but no linear
  relationship
• Individual characteristics
• Age, income, giving financial help to
  other set of parents
• Parents characteristics
• Physical health status, living
  arrangements, main financial source, whether they
  provide financial help to migrant children.

45

• Actual amount of financial support after
  migration is affected by
• Gender
• Female migrants give more financial
  support to their parents-in-law after migration
• Individual characteristics
• Income, education, spouses income,
  giving financial help to other set of parents
• Parents characteristics
• Parental living status, living
  arrangements, main financial source, whether they
  provide financial help to migrant children.
• Migration experience has no significant
  effects

46
Conclusions

• Gender difference Females are likely to give
  parents-in-law more financial support
• ? Patrilineal family system is still
  dominant
• Both males and females provide more financial
  help to natal parents after migration
• ? Out-migration of females could change
  the traditional pattern of old-age support and
  might weaken son preference in rural China
• Grandparents receive more remittance when they
  take care of grandchildren
• ? Intergenerational transfer between
  parents and their migrant children is reciprocal.