Family Changes and Income Inequality under Globalization: The Case of Hong Kong

1
Family Changes and Income Inequality under
GlobalizationThe Case of Hong Kong

• Stephen WK ChiuDepartment of SociologyThe
  Chinese University of Hong Kong

Photo Alex Chung
2
The Problem
Gini Coefficient
1991 0.476
2001 0.525

• Increased by 10.3

3
The Problem
4

• Hong Kong has become one of the most unequal
  societies in the developed world.

Country/Region Year Gini Coefficient
Brazil 2001 0.59
Hong Kong 2001 0.53
Argentina 2001 0.52
Costa Rica 2000 0.47
China 2001 0.45
Singapore 1998 0.42
United States 2000 0.41
Ghana 1999 0.41
Italy 2000 0.36
United Kingdom 1999 0.36
India 2000 0.33
Korea, Rep. 1998 0.32
Czech Republic 1996 0.25
Sweden 2000 0.25
5
Economic Explanation

• Economists explain rising inequality by Human
  Capital Theory and Trade Theory
• Globalization and technological development the
  primary causes

6
Economic Explanation

• In advanced economies, globalization increases
  return to high level human capital and reduces
  return to low level human capital through trade
  with less developed labour-surplus economy.
• Technological change, especially in IT, increases
  return to skilled worker by increasing the
  demand.

7
Sociological Explanation

• Saskia Sassens Global City Thesis

8
Sociological Explanation

• Globalization led to the rise of global cities
  with dispersal of production and centralization
  of control.
• Global cities are command centres of the global
  economy with a high concentration of producer
  services (banking, advertising, accounting, law)
  controlling a dispersed network of production.
• Primary global cities New York, London, Tokyo.
• Secondary global cities Hong Kong, Singapore,
  Frankfurt, Sao Paulo, Chicago etc.

9
Sociological Explanation

• Global city development leads to social
  polarization
• 1) the growing inequality in the profit-making
  capacities of different economic sectors and in
  the earning capacities of different types of
  workers 2) the polarization tendencies embedded
  in the organization of service industries and the
  casualization of the employment relation and 3)
  the production of urban marginality, particularly
  as a result of new structural processes of
  economic growth rather than those producing
  marginality through abandonment. (Sassen 1998
  137)

10
Sociological Explanation

• Manufacturing generates a large number of
  middle-income jobs.
• Services typically consist of high income and low
  income employees with few in the middle.
  Investment bankers vs the janitors.
• Therefore globalization transforms the
  occupational structure, leading to polarization
  of income.

11
Criticisms on the Polarization Thesis

1. First, there are doubts whether the polarizing
   trend in occupational and income structure is
   really observable in global cities and that it
   ignores the process of professionalization and
   the growth of the new middle class in the global
   cities (Hamnett 1994, 1996 Baum 1999).
2. Second, the concept of polarization as used in
   the global city literature by Sassen is also
   found to be imprecise, whether it points to
   merely relative polarization in the sense of
   widening income differentials or absolute
   polarization leading to the growth in both the
   top and bottom ends but a shrinking middle
   (Hamnett 1994 Vaattovaara and Kortteinen 2003).

12
Criticisms on the Polarization Thesis

1. Hamnetts critique of the concept of polarization
   in the global city thesis boils down to the
   danger in the thesis to mask the mediating role
   of local factors in the social outcomes and urges
   the necessity to conceptually unpack the term
   polarization and to examine the extent to which
   different forms of polarization are found in
   different contexts and to theorise the reasons
   for such variations (Hamnett 1996 1408).

13

• My objective therefore is to look at local
  factors that mediating between global changes as
  well as the social factors interacting with
  economic changes.

14

• One problem in unit of analysis in income
  inequality
• Individual or household?
• Key Hypothesis
• Household formation strategies affect income
  inequality significantly
• Wives rising income and labour force
  participation increased overall household income
  inequality

15

• Data Public Use Dataset of Population Census
  1991 and 2001.
• Definition of household income primary and
  secondary income excluding domestic helpers.
• Adjusted vs unadjusted income by household size.

16

• Basic statistical tool decomposition analysis of
  coefficient of variation

• where Sk µk/ (µh µw µo), CVk is the
  coefficient of variation for income component k,
  ?kj is the correlation between a pair of income
  components, Sk is the share of total family
  income form component k, and µk is the mean of
  income from component k. In the analysis of the
  relative contribution of different sources in
  household income, the subscript h denotes
  husbands income w denotes wives income and o
  denotes residual income from other sources.

17

• In simple terms
• If
• I H W (I household income, H husband
  income, W wife income)
• Then
• The dispersal of I will be greater if
• H and W are highly correlated, and vice versa.
• Also important is the share of each source of
  income in the total household income
• as well as the level of income of each marital
  partner.

18
Table 4A Household Income by Sources, 1991 and
2001
Share () Share () Share ()
Sources 1991 2001 Change
1. Unadjusted Household Income 1. Unadjusted Household Income
Husband 58.08 55.66 -4.17
Wife 18.86 24.29 28.79
Others 23.06 20.05 -13.05
Couple Households 100 100 ---
2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size
Husband 59.65 56.49 -5.3
Wife 20.03 25.3 26.31
Others 20.32 18.21 -10.38
Couple Households 100 100 ---
19
Table 4B Household Income by Sources, 1991 and
2001
Mean (HK) Mean (HK) Mean (HK) Mean (HK) Mean (HK) Mean (HK) Correlation with Wifes Income Correlation with Wifes Income Correlation with Wifes Income
Sources 1991 1991 2001 Change Change Change 1991 2001 2001
1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income  
Husband 12985.51 15820.73 15820.73 15820.73 21.83 0.333 0.382 0.382
Wife 4217.34 6903.73 6903.73 6903.73 63.70 --- --- ---
Others 5156.35 5700.35 5700.35 5700.35 10.55 -0.085 -0.132 -0.132
Couple Households 22359.2 28424.81 28424.81 28424.81 --- 0.589 0.658 0.658
2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size 2. Adjusted Household Income by Household Size  
Husband 6844.61 8522.36 8522.36 8522.36 24.51 0.376 0.416 0.416
Wife 2298.06 3817.14 3817.14 3817.14 66.10 --- --- ---
Others 2331.29 2747.15 2747.15 2747.15 17.84 -0.121 -0.159 -0.159
Couple Households 11473.96 15086.65 15086.65 15086.65 --- 0.652 0.702 0.702
20
Table 6 Decompositions of Change in Income
Inequality by Income Source, 1991 and 2001
CV2 CV2 CV2 CV2 Relative Contribution Relative Contribution Relative Contribution Relative Contribution Contribution of Income Source to Change
Income Sources 1991 2001 change change 1991 2001 change change
1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income 1. Unadjusted Household Income
Husband 0.908 1.081 19.087 0.53 0.53 0.561 5.81 8.99 8.99
Wife 2.753 2.321 -15.712 0.237 0.237 0.297 25.501 9.166 9.166
Others 3.676 3.851 4.739 0.233 0.233 0.142 -39.038 -7.613 -7.613
Couple Households 0.608 0.672 10.541 1 1 1 --- 10.543 10.543

2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income 2. Adjusted Household Income
Husband 1.017 1.153 13.295 0.601 0.601 0.593 -1.318 4.915 4.915
Wife 2.955 2.442 -17.342 0.278 0.278 0.329 18.125 8.209 8.209
Others 3.481 3.699 6.255 0.12 0.12 0.078 -35.298 -3.503 -3.503
Couple Households 0.649 0.712 9.624 1 1 1 --- 9.621 9.621
21

• Next I assess the impact of changes in income
  sources among households by comparing the
  observed distribution with a reference
  distribution. The reference distribution is
  constructed by assuming three counterfactual
  conditions in order to evaluate whether wives
  income had a disequalizing effect on the income
  distribution
• Counterfactual 1 all wives did not work and had
  zero income.
• Counterfactual 2 the mean and dispersion of
  wives income had not changed over the period in
  question.
• Counterfactual 3 the mean, dispersion and
  correlation of wives income with other sources
  had not changed over the period in question.

22
Table 7 The Impacts of Changes in Wives
Earnings under Different Counterfactual
Conditions for Married Couples
CV2 of Couple Households Actual Counter- factual 1 Counter- factual 2 Counter- factual 3
1. Unadjusted Household Income 10.55 -2.00 -17.88 -18.88
2. Adjusted Household Income 9.62 -1.56 -12.80 -13.85
23
Table 8 The Impacts of Changes in Wives
Earnings under Different Counterfactual
Conditions for All Households
All Households Acutal Change in CV2 between 1991 and 2001 Change in CV2 between Counterfactual 2001 Actual 1991
Counterfactual 2 (mean and dispersion of wives' income hand not changed) Counterfactual 2 (mean and dispersion of wives' income hand not changed) Counterfactual 2 (mean and dispersion of wives' income hand not changed)
Unadjusted Household Income 10.55 -4.93

Counterfactual 3 (mean, dispersion and correlation of wives' income with other sources had not changed) Counterfactual 3 (mean, dispersion and correlation of wives' income with other sources had not changed) Counterfactual 3 (mean, dispersion and correlation of wives' income with other sources had not changed)
Unadjusted Household Income 10.55 -5.71
24
The Broader Picture

• Expansion in Higher Education
• Increase in female enrolment in higher education
• Girls outnumbered boys in tertiary institutions
  in 2001, accounting for 54.4 of all students in
  state-funded programmes.
• The proportion of women with tertiary education
  in the population also increased from 9.4 in
  1991 to 15.2 in 2001
• The overall labour force participation rate among
  all women rose moderately from 49.5 to 51.6
  during the same period. The proportion of wives
  with tertiary education who were working also
  rose from 68.5 in 1991 to 71.1 in 2001.

25
The Broader Picture

1. The percentage of households with an
   economically-active wife increased moderately
   from 30.9 of all households in 1991 to 32.7 in
   2001.
2. Rising income level of working women because
   rising producer services offer high income jobs
   to women
3. Other institutional factors also help foreign
   domestic helpers released middle class wives
   from childcaring.

26

• Higher correlation between husband and wives
  income, from 0.333 in 1991 to 0.382 in 2001
• The reason educational homogamy

27
(No Transcript)
28
Conclusion

• Economic factors contributed to widening
  individual inequality but socio-demographic
  factors further accentuated household inequality
• As households typically pool together resources
  in consumption and human capital investment (eg
  education for children), inter-household
  inequality is important in determining life
  chances
• Institutions and policies encouraging the working
  of women in low-income households deserved more
  consideration