جامعه، مقایسه و رفاه ذهنی در یک جامعه تقسیم شده

Using South African data, the paper poses six questions about the determinants of subjective well-being. Much of the paper is concerned with the role of relative concepts. We find that comparator income, when measured as the average income of others in the local residential cluster, enters the household's utility function positively (close neighbors are ‘positives’, not ‘negatives’), but that the income of more distant others enters negatively. Race-based comparator groups are also important in racially divided South Africa. Relative income is more important to happiness at higher levels of absolute income. Potential explanations and implications of these results are considered.

In this paper we pose several questions about the determinants of subjective well-being. We do so for South Africa, a country that, because of its unusually divided society, provides a good case study of the effects of community and comparisons on subjective well-being. Each of these questions is new for South Africa. More broadly, the major contribution of the paper to the economic literature on subjective well-being concerns the role of comparisons made with others in the community. What are the reference groups against which people compare themselves, and do their comparisons give rise to fellow feeling or to feelings of relative deprivation? We highlight the roles that space and race can play. Spatially defined reference groups are shown to have a positive effect on subjective well-being, suggesting that neighbors may be ‘positives’ and not ‘negatives’. By contrast, racially defined reference groups are shown to have a negative effect, implying that relative deprivation may be experienced in relation to others of one's own race rather than to neighbors or to the larger society. There are interesting implications for welfare economics and for policy. In Section 2, we provide a framework of concepts and literature about the effects of comparisons on subjective well-being. Section 3 describes the South African context and the data, and outlines the method: the estimation of subjective well-being functions that include explanatory variables representing relevant comparisons. The empirical Section 4 presents the results, question by question. In particular, we test whether and how spatial and racial comparisons affect subjective well-being. Section 5 concludes and draws out the implications of the analysis.

The first question that we posed was to what extent is it absolute income and to what extent relative income that determines happiness. We found a good deal of evidence that both the income of the household and the income of other households influence subjective well-being. Some of our estimates suggest that the latter relative to the former is more important than the former on its own. Second, insofar as relative concepts matter, is it only relative income that counts or are comparisons made in other dimensions as well? The relative concepts that appear to have an effect are unemployment, education and income. However, unemployment and education are to some extent collinear with income. When all three are included, only income remains important. Third, if relative income matters, who are the relevant others with whom people compare themselves? We identified two types of reference group. Comparisons are made with the income of other households, the relevant others being based on space and on race. Fourth, and most importantly, does low income relative to others decrease or increase happiness (i.e. given own income, does the income of relevant others affect happiness negatively or positively)? We found that higher income of other households in a small community raises subjective well-being. This is a powerful new result. We considered four explanations: altruism, mutual insurance, a social wage, and non-causal association. Although suggestive rather than conclusive, the results of our various tests were consistent with altruism or fellow feeling, but inconsistent with the alternative explanations. This has interesting policy implications, such as for local taxation. It is not predictable whether our finding that close neighbors are positives applies also in other societies. The study for the United States that found neighbors to be negatives (Luttmer) used too broad a measure of neighborhood to be regarded as contradicting our result. Nevertheless, the peculiar nature of South African society (not only poor and unequal but also racially fractured) may help to produce unusually strong feelings of local solidarity. 10 In posing the fourth question we also examined race as the reference group. We found, in contrast to the spatial result, that higher income of households of the same race is associated with lower subjective well-being. This is consistent with perceptions of relative deprivation (aspirations ahead of achievements) possibly arising from standard setting or notions of unfairness, or envy. It appears that, whereas close spatial proximity (the same cluster) creates a sense of community, close social proximity (the same race) creates feelings of relative deprivation or sets goals and aspirations. Fifth, does the strength of the positive or negative relationship weaken as the reference group is broadened to include socially more distant people? Our evidence suggested that positive spill-overs on subjective well-being at the local (cluster) level are diluted as orbits of comparison are expanded to include strangers. At the broader (district) level we found that spill-overs are negative. Similarly, perceptions of relative deprivation in relation to persons of the same race are evident not at the local level but at higher levels of geographical aggregation. Our sixth inquiry was to discover whether the importance of relative income varies with absolute income. Indeed, the effect of relative income on happiness is strengthened as income rises. Whereas absolute income is an important determinant of the happiness of people whether they are below or above the poverty line, relative income is not relevant to the poor. For those who are not poor, however, within-race relative income is important, suggesting that perceptions of relative deprivation play a role. The finding that South Africans confined themselves to comparisons with others of their own race corresponds to the claim that English slum-dwellers a century ago confined themselves to comparisons within the slum (Roberts). It suggests that people choose their comparator groups by reference to social proximity, possibly to ease their states of mind or to protect self-esteem (in line with Festinger). South Africa in 1993 inherited a legacy of White privilege and black disadvantage. Africans would want to reduce hurt by avoiding comparisons with others whose achieved states were infeasible for them. Similarly, Whites would want to assess themselves in relation to other Whites, the community of which they felt part; comparisons with blacks would involve loss of face. Policy-making requires an understanding of reality, including the reality of people's perceptions. Nevertheless, we are hesitant to draw policy conclusions from our results. One reason is the argument of Sen, 1983 and Sen, 1999 eschewing the ‘metric of utilities’ in favour of the ‘capabilities’ approach to addressing poverty; similarly, others regard the fulfilment of ‘basic needs’ as the appropriate objective. The underlying case against the utilities approach is that, by adjusting their aspirations to reality, people are conditioned to bear hardship. In the words of Sen (1999, p. 358):