Inequality aversion, health inequalities and health achievement

https://doi.org/10.1016/S0167-6296(02)00006-1Get rights and content

Abstract

This paper addresses two issues. The first is how health inequalities can be measured in such a way as to take into account policymakers’ attitudes towards inequality. The Gini coefficient and the related concentration index embody one particular set of value judgements. By generalising these indices, alternative sets of value judgements can be reflected. The other issue addressed is how information on health inequality can be used together with information on the mean of the relevant distribution to obtain an overall measure of health “achievement”.

Introduction

The literature on health inequality measurement has benefited substantially from cross-fertilisation, both within the discipline of economics (principally from the literature on income inequality measurement to the literature on health inequality measurement) and between the disciplines of economics, epidemiology and public health (Wagstaff et al., 1991, Mackenbach and Kunst, 1997). This paper extends the literature on health inequality measurement in two directions, borrowing heavily on the income inequality literature.

The first is to allow for the fact that commonly used summary measures of health inequality have ethical judgements about inequality aversion built into them—albeit implicitly. This is true, for example, of the Gini coefficient, which has been used to measure pure health inequality (Le Grand, 1987, Le Grand, 1989). But it is also true of the concentration index1 (Wagstaff et al., 1991, Kakwani et al., 1997), which has been used to measure socio-economic inequalities in health—i.e. health inequalities by income or by some other measure of socio-economic status (SES).2 The implicit ethical judgements have been recognised in the measurement of pure health inequality, where Atkinson’s (1970) index has been used to allow attitudes to inequality to be varied (Le Grand, 1987, Le Grand, 1989). But varying attitudes to inequality have not been allowed for up to now in the measurement of socio-economic inequalities in health. To allow for varying attitudes to inequality aversion, this paper develops the concentration index analogue of the Yitzhaki’s (1983) extended Gini coefficient. While the aim is primarily to extend the literature on the measurement of socio-economic health inequalities, the paper also contributes to the literature on the measurement of pure inequality, since, from a formal point of view, the latter can be thought of a special case of the measurement of socio-economic inequality in health, where what matters is the individual’s rank in the health distribution rather than their rank in the income distribution. The approach suggested here, when used in the measurement of pure health inequality, is a natural alternative to Atkinson’s index.

The second direction in which the paper extends the literature on the measurement of health inequality is to recognise that policymakers are unlikely to be concerned only about health inequalities, either of the pure variety or the socio-economic. Rather they are likely to be willing to tradeoff increases in inequality against improvements in the mean of the distribution (Wagstaff, 1991). This paper shows how, as in the income inequality literature (Lambert, 1993), a single summary measure can be computed that reflects both average health and inequality in its distribution. This index is termed here an index of “achievement”, but is in effect an abbreviated social welfare function—albeit in the health domain. Again, the exposition is for the case where the interest is in socio-economic inequalities, but the application to the case of pure inequality is immediate.

The plan of the paper is as follows. Section 2.1 generalises the concentration index to allow the degree of inequality aversion to be specified. Section 2.2 proposes the achievement index that combines information on inequality with information on the average level of health. Section 3 presents some empirical illustrations of these two measurement tools using data for 44 developing countries on socio-economic inequalities in and average levels of three health indicators: under-five mortality, child malnutrition and adult fertility.

Section snippets

Measurement issues

The starting point is the measurement of health inequalities. To make the discussion more applicable to typical health indicators, it is assumed that the health variable measures ill health. It might be an index based on, say, a self-assessed health question (Wagstaff and Van Doorslaer, 1994, Gerdtham et al., 1999, Humphries and Van Doorslaer, 2000). Or it might be an anthropometric measure of malnutrition (Wagstaff and Watanabe, 2000, Wagstaff et al., 2001). Or it might be a binary variable

Empirical illustrations

In this section, these methods are illustrated for three health indicators—under-five mortality, child malnutrition and adult fertility. The computations are based on grouped data from 44 developing countries, taken from tabulations by Gwatkin et al. (2000) on data from the Demographic and Health Survey (DHS). The tabulations show average values for each of five “wealth” quintiles.

Summary and conclusions

To recap briefly, the concentration index has embedded in it a particular set of value judgements about the weights to be attached to the health of people at different points in the income distribution. The standard concentration index can be shown to be equal to the complement of a weighted sum of the health shares of the individuals in the sample. The weights decline in a stepwise fashion, starting with a weight close to 2 for the poorest person, declining by equal steps for each one person

Acknowledgements

Without wishing to incriminate them in any way, I am grateful to Eddy Van Doorslaer and two anonymous referees for comments on an earlier version of this paper. The findings, interpretations and conclusions expressed in this paper are entirely those of the author and do not necessarily represent the views of The World Bank, its Executive Directors or the countries they represent.

References (36)

  • G.A.O. Alleyne et al.

    Equality, equity: why bother?

    Bulletin of the World Health Organization

    (2000)
  • P. Braveman et al.

    World Health Report 2000: how it removes equity from the agenda for public health monitoring and policy

    British Medical Journal

    (2001)
  • Evans, T., Whitehead, M., Diderichsen, F., Bhuiya, A., Wirth, M., 2001. Introduction. In: Evans, T., Whitehead, M.,...
  • D. Filmer et al.

    The effect of household wealth on educational attainment: evidence from 35 countries

    Population and Development Review

    (1999)
  • E. Gakidou et al.

    Defining and measuring health inequality

    Bulletin of the World Health Organization

    (2000)
  • D. Gwatkin

    Health inequalities and the health of the poor: what do we know? What can we do?

    Bulletin of the World Health Organization

    (2000)
  • Gwatkin, D., Rutstein, S., Johnson, K., Pande, R., Wagstaff, A., 2000. Socio-Economic Differences in Health, Nutrition...
  • International Monetary Fund, OECD, United Nations, The World Bank, 2000. A better world for all: progress towards the...
  • Cited by (194)

    View all citing articles on Scopus
    View full text