Subjective well-being poverty vs. Income poverty and capabilities poverty?

Abstract

The conventional approach of economists to the measurement of poverty is to use measures of income or consumption. This has been challenged by those who favour broader criteria, such as fulfilment of ‘basic needs’ and the ‘capabilities’ to be and to do things of intrinsic worth. This paper asks: to what extent are these different concepts measurable, to what extent are they competing or complementary, and is it possible for them to be accommodated within an encompassing framework? We conclude that it is possible to view subjective well-being as an encompassing concept, which permits us to quantify the relevance and importance of the other approaches and of their component variables. Any attempt to define poverty involves a value judgment as to what constitutes a good quality of life or a bad one. We argue that an approach which examines the individual's own perception of well-being is less imperfect, or more quantifiable, or both, as a guide to forming that value judgement than are the other potential approaches. The argument is illustrated using a South African household survey.

Acknowledgements

We would like to thank John Helliwell for stimulating discussions about the ideas in this paper. In addition to workshop participants at the GPRG workshop, Oxford, and seminar participants at the Wellbeing in Developing Countries (WeD) Research Group at Bath University, we would like to thank Alan Krueger, Allister McGregor, Wendy Olsen, John Toye and an anonymous referee for helpful comments on the paper. Any errors are ours. The support of the UK Economic and Social Research Council is gratefully acknowledged. The work was part of the programme of the ESRC Global Poverty Research Group (grant M571255001).

Notes

1. Ravallion and Lokshin (2001) and Graham and Pettinato (2002) are exceptions.

2. There are obvious issues of endogeneity and causality which will be discussed below.

3. Powdthavee (2005) and Moller (1998) have also written on perceived quality of life in South Africa.

4. In Table 2 (and throughout the paper) standard errors have been corrected for clustering.

5. An instrumentation procedure can in principle be used to address the likely endogeneity of income in a happiness equation. Empirically justifiable instruments available are the variables proportion of males in the household and household size, both of which are statistically significant in the income equation and insignificant in the happiness equation. However, there is no strong a priori theoretical justification for them. Studies using panel data and exogenous variation in income (for example a lottery win) have found that causality runs from income to happiness (for example see Gardner and Oswald, 2001).

6. We experimented with variables from the cluster questionnaire including distance from the cluster to various facilities (such as health clinic, school, shops, bank, post office, market, etc), number of such facilities within the cluster, and distance to nearest source of transport, as well as with cluster averages of household variables such as distance to nearest source of water for the household, etc.

7. The ownship, debt and urateb variables could be included under the monetary variables category, together with income and assets, and the crime variable included under the physical functionings (basic needs) category.

8. Kingdon and Knight (2007, forthcoming) explore the influence of relative income on subjective well-being in greater detail.