Community, Comparisons and Subjective Well-being in a Divided Society¶

Literature › Works

Geeta Gandhi Kingdon; John Knight (2007). Community, Comparisons and Subjective Well-being in a Divided Society. Journal of Economic Behavior & Organization 64(1). Elsevier.


Type	journal article
Scope	Comparative and methods
Method	econometric
Period studied	1993–1993
Regions	South Africa, national (9 provinces, 187 districts, 360 clusters)
Themes	inequality poverty unemployment, origins segregation, transition post apartheid, methods data sources, comparative theory
Citation key	`kingdon2007`

Argument¶

Using the 1993 SALDRU household survey, the paper tests how absolute and relative income, and comparator groups defined by space and race, shape household subjective well-being in racially divided South Africa. Higher average income among neighbours in the small residential cluster raises well-being, which the authors read as altruism or fellow-feeling. Income of more distant others (the district) and of same-race households in the district lowers it, which is consistent with relative deprivation. Relative income matters only to the non-poor, while absolute income matters to both poor and non-poor. The Africans' large well-being penalty, even after controls, is read as the legacy of apartheid disadvantage. Evidence is ordered probit and OLS regressions with robustness and IV checks.

Evidence and method¶

SALDRU 1993 national household survey (Living Standards Measurement Study pattern; c. 8,800 households, 8,279 used; 360 clusters, 187 districts, 9 provinces). Dependent variable is the household's reported satisfaction with life on a five-point scale. Poverty lines from May (1998). Cluster food price indices are used for deflation.

Findings¶

Absolute household income raises satisfaction but with modest effect sizes. (p. 12)

An increase in absolute household income (log of household per capita income, ln_hhpci) from one standard deviation below to one standard deviation above the mean raises the probability of being satisfied or very satisfied with life by 11 percentage points.
A large racial gap in satisfaction persists after controlling for income, education and employment. (p. 12)

The African probability of being satisfied or very satisfied is 21.5 percentage points lower than that of Whites, even after controlling for observed income, education, employment, and so on.
Cluster-level income of others has a positive effect on well-being, while district income does not. (p. 13)

the income of others within the cluster increases happiness, but income within the district weakly decreases happiness.
The positive neighbour effect is the same across African and White households, which is inconsistent with the mutual-insurance explanation. (p. 13)

The coefficient for Africans is 0.271 (t = 4.28) and for Whites 0.272 (t = 4.56). This result is inconsistent with the mutual insurance explanation.
Higher income of same-race households in the district sharply lowers satisfaction, consistent with race-based relative deprivation. (p. 16)

if race-specific district mean income increases by one standard deviation (0.989) from its mean (5.946), the probability of being satisfied or very satisfied with life falls by a large 21.2 percentage points.
The disadvantage of being African outweighs the advantage of being a relatively rich African in the race-specific income distribution. (p. 16)

Thus, the disadvantage of being African greatly outweighs the advantage of being a relatively rich African.
Unemployment hurts well-being more in low-unemployment areas, suggesting norm effects among the poor. (p. 17)

the coefficient and robust t-value of hhurate in high and low unemployment areas were -0.097 (t = -1.5) and -0.399 (t = -4.0) respectively
Apartheid divide-and-rule is argued to have entrenched separate reference groups that may have delayed political change. (p. 22)

The apartheid policy of divide and rule accentuated racial differences. Our findings suggest that people did not feel themselves to be part of a common society.