Subjective well-being, comparisons and reference groups in post-apartheid South Africa

1
Subjective well-being, comparisons and reference
groups in post-apartheid South Africa

• Marisa Coetzee
• University of Stellenbosch
• South Africa
• 27 October 2011

2
Outline

• Background
• Theoretical Framework
• Data
• Results
• Conclusions

3
Background

• Subjective well-being / happiness (SWB)
• Kingdon and Knight 2007 the determinants of SWB
  in South Africa using 1993 PSLSD (SALDRU) data
• Conclusions (briefly)
• Relative income more NB than absolute income
• Relative income more NB than other relative
  measures (e.g. unemployment and education)
• Households are altruistic towards other
  households within the same residential cluster
  but compete with households in the same district
  and racial group
• South Africans reference groups divided along
  racial lines

4
Theoretical Framework

• Post apartheid, intra-race inequality has
  increased, while inter-race inequality decreased
  slightly.
• Have all attempts by the SA government to
  integrate society been successful?
• Pierre du Toit and Hennie Kotzès Liberal
  Democracy and Peace in South Africa (2011)
• Racial divide in apartheid legislation further
  entrenched by affirmative action the
  re-racialization of society
• However, signs of racial integration from WVS
  data
• Increased tolerance towards other race groups
• How does this affect South Africans reference
  group?

5
Data

• National Income Dynamics Study (NIDS)
• Sample of 7305 households (31170 individuals)
• All individuals aged gt16 were asked about their
  SWB
• First Wave (2008)

6
NIDS 2008
7
The question is

• However, no study verifying the results of
  Kingdon and Knight for post-apartheid SA
• Is race still a deciding factor in determining
  the SWB of South Africans or has democracy
  changed the reference group?
• Replicate the analysis by Kingdon and Knight
  (2007) but using NIDS data from 2008
• Issues
• NIDS has 10-point scale versus 5-point scale for
  PSLSD (SALDRU) data
• NIDS SWB question was asked at individual level,
  SALDRU at household level
• Have to therefore include individual- and
  household-level variables

8
Subjective well-being and relative income across
spatial reference groups
1 2 3 4 5 6 7
african -0.2772 -0.2670 -0.2044 -0.1977 -0.2458 -0.2310 -0.2377
coloured 0.0779 0.0828 0.1333 0.1315 0.0898 0.1021 0.1114
asian 0.0500 0.0428 0.0646 0.0774 0.0886 0.0850 0.0656
hhurate 0.1352 0.1306 0.1503 0.1424 0.1507 0.1411 0.1219
asset_index 0.0170 0.0172 0.0150 0.0146 0.0168 0.0162 0.0151
lhhinc_pc 0.1025 0.1017 0.0950 0.0978 0.1035 0.1000 0.1018
Cluster hh umepl rate 0.6048 0.4419 0.4576
District hh umepl rate 0.7728 0.0440
Cluster hh educ 0.0340 0.0436 0.0370
District hh educ -0.0637 0.0147
District log hh pc income -0.1845 -0.1826 -0.1925
Cluster log hh pc income 0.0321 0.0093
N 10444 10444 10444 10444 10449 10444 10444
Notes Reported results are coefficients from
ordered probit regressions on subjective
well-being categories. A full set of control
variables are included, but not reported.
significance at 1 level, significance at 5
level, significance at 10 level.
9
Race-specific relative income
1 2 3 4
african -0.2310 -0.1640 -0.1547 -0.1662
coloured 0.1021 0.1494 0.1544 0.1461
asian 0.0850 0.1027 0.1271 0.1084
hhurate 0.1411 0.1418 0.1531 0.1434
asset_index 0.0162 0.0160 0.0166 0.0161
lhhinc_pc 0.1000 0.0990 0.1112 0.1036
c_lnhhpci 0.0321 0.0267 0.0261
d_lnhhpci -0.1826 -0.2043 -0.2068 -0.2016
lrdm_inc 0.0469 0.0497 0.0448
rpctile 2 -0.0903 -0.0862
rpctile 3 0.0537 0.0609
rpctile 4 -0.0327 -0.0238
rpctile 5 -0.0458 -0.0254
N 10444 10444 10444 10444
Notes Reported results are coefficients from
ordered probit regressions on subjective
well-being categories. A full set of control
variables are included, but not reported.
significance at 1 level, significance at 5
level, significance at 10 level.
10
The effect of Perceived Relative Income on
subjective well-being
Specification 1 Specification 2
african -0.2779 -0.2644
coloured 0.1879 0.2016
indian/asian -0.0084 -0.0263
asset index 0.0134 0.0111
log of pc hh income 0.0751 0.0564
Relative household income to others in your village/suburb
above average inc in village/suburb -0.3836 -0.3622
average inc in village/suburb -0.6439 -0.5962
below average inc in village/suburb -1.0957 -0.9829
much below average inc in village/suburb -1.3052 -1.1155
Relative household income to others in SA
ladder rung 2 in SA 0.3885
ladder rung 3 in SA 0.5003
ladder rung 4 in SA 0.6155
ladder rung 5 in SA 0.8839
ladder rung 6 in SA 1.0199
N 9865 9831
Notes A full set of control variables are
included, but not reported. significance at
1 level, significance at 5 level,
significance at 10 level.
11
Perceived relative income and subjective
well-being above and below the poverty line
Below the R515 poverty line Above the R515 poverty line
african -0.2844 -0.2320
coloured 0.2779 0.1906
asian -0.0167 -0.1120
asset_index 0.0044 0.0195
lhhinc_pc 0.0417 0.0583
Relative household income to others in your village/suburb
above average inc in village/suburb -0.3908 -0.2218
average inc in village/suburb -0.6059 -0.4515
below average inc in village/suburb -1.0086 -0.8163
much below average inc in village/suburb -1.1815 -0.8138
Relative household income to others in SA
ladder rung 2 in SA 0.3954 0.3513
ladder rung 3 in SA 0.4421 0.5936
ladder rung 4 in SA 0.6082 0.7008
ladder rung 5 in SA 0.6672 1.1244
ladder rung 6 in SA -0.5229 1.9269
N 6197 3631
Notes A full set of control variables are
included, but not reported. significance at
1 level, significance at 5 level,
significance at 10 level.
12
Conclusions

• Results from Kingdon and Knights study remain
  unchanged in relation to spatial reference groups
  
• Altruism appears to be one possible explanation
  for the positive effect of education and
  employment levels of other households
• However, racial division of reference groups
  seems to have changed
• Appears to be pointing in the direction of a more
  racially integrated society