Multidimensional Perspective on the Well-being of Older People

2.2 Measurement

We report on generic considerations concerning the definition of income to be used in measuring economic well-being. The preferred definition of income is that of the ‘Annual Household Net Income’. Thus, household is preferred over benefit-unit as an income unit, since it captures more meaningfully the sharing of economic resources that older people experienced. Annual income is preferred over current income as it provides a more reliable measure of a household’s financial situation, and similarly net income is a better proxy of a household’s disposable resources than the gross income. Income before deducting housing costs is preferred over income after housing costs, and it is desirable, when possible, that the imputed rent be included in this measure.
The British Household Panel Survey is chosen as the preferred database, mainly because it is the survey that provides annual income data in Britain, and also due to the fact that it offers detailed health data in its 1999 wave. The annual income available in the BHPS is synthetic, as it is generated based on data on receipt of different sources of income throughout the year and the information on benefit values and their uprating. The health variables and their aggregation into health indicators are derived from the SF-36 questionnaire in the 1999 wave of BHPS.

3. Assessment of Income Deprivation and Limiting Health

Given the link between advancing age and physical frailty and the onset of ill health and disability — which are debilitating as well as costly — it is essential to understand how factors such as health relate to and impinge on older people’s welfare. Multidimensional analysis sets out to capture these insights, by using health and income as the base dimensions for measuring personal well-being of older people in Great Britain. The analyses here are structured in such a way as to allow three distinct result sequences to emerge.
•    First, in subsection 3.1, we summarise the findings on the extent to which health restricts the positive functionings of older people in their normal economic and social activities. By analysing the diversity of older people with respect to their physical, mental and social functionings, we identify those characteristics that are most commonly associated with limiting health and/or disability in old age.
•    Next, in subsection 3.2, we shift the focus towards measuring the economic well-being of older people. These analyses demonstrate the relative position of older people as a whole, as well as the relative position of older people’s subgroups within the society as a whole. The income-based analyses help us identify those attributes associated with high risks of financial poverty in old age.
•    Finally, in subsection 3.3, these analyses — assessing older people’s well-being uni-dimensionally with respect to their economic and health status — are complemented by the estimation and use of the financial costs of disability. This work aims at aggregating the two dimensions of economic and health well-being for the purpose of the multidimensional measurement of older people’s well-being. This work also makes headway towards providing an ‘equivalisation’ of incomes that accounts for not only differences in household size and composition but also other sources of variation in household needs such as disability. It can therefore be argued that identifying and accounting for a fuller set of peoples’ needs through equivalisation leads us towards the operationalisation of Sen’s capability approach.2
As mentioned above, all empirical work on multidimensional analyses makes use of data from the BHPS. The empirical analyses are derived from the 1999 data, the year for which the BHPS included detailed information on health.

3.1 Health Well-being: The Unidimensional Measure

The empirical results included in this section examine older people’s well-being using the health dimension alone. These analyses are based on various individual indicators of health, as well as on aggregated indices derived from these indicators. Alternative analytical techniques based on descriptive statistics as well as on multivariate regression analyses are used in order to identify the factors associated with a high incidence of limiting health in old age. The most notable findings to emerge are:
•    Older women are worse off compared to older men in terms of physical health, and this result holds true when controlling for other attributes (such as age, marital status, living arrangements, job status and housing tenure).
•    For physical health, older age groups (aged 80 to 84, and aged 85+) are clearly worse off compared to younger age groups (aged 65 to 69). Again, this result holds true when controlling for other attributes of older people.
•    Housing tenure as a tenant is one other attribute that stands out in identifying older people with poor health.

3.2 Economic well-being: the unidimensional measure

Here, we concentrate on the economic aspects of older people’s well-being. For this purpose, we use equivalised household income as an indicator of economic well-being. First, results of relative ranking of various subgroups of older people are analysed. This is followed by statistics on the poverty incidence and the poverty severity for the same subgroups.
a. Relative ranking of subgroups
Here, we show how older people as a whole and various subgroups of older people are ranked in the decile cumulative distribution of income, which is derived on the basis of income of the overall population.3 The results in Table 1 demonstrate that older people are disproportionately represented in the bottom parts of the overall income distribution. This result is a known phenomenon in the case of Great Britain. Next, we summarise results that are further subdivided in subgroups, based on sex, marital status, living arrangements, job status, housing tenure and age group.
Results show that older women are clearly worse off as compared to older men since they are more often represented in the bottom part of the income distribution. Moreover, the cumulative share for older women are everywhere greater than that for older men, which is another indication of older women being more deprived than older men. Results also identify widow(er)s as the group that experiences low-income status more often than others: about one-third of all widow(er)s are found within the bottom one-fifth of the overall population. Those who are classified as ‘divorced, separated or never married’ are also found more often amongst the low income groups within older people.
Results also show that older people who live independently are more likely to have a lower income status. Those who live with others appear to do much better, as they are less often in the bottom part of the income distribution. As can be expected, older people who are economically active are found less often in the bottom part of the income distribution than those who have retired from the labour market. Tenants are also identified as the high-risk group. Homeowners, on the other hand, enjoy just about the same economic ranking as the whole group of older population. The oldest age group is also more often in the low-income status than younger age groups: about one-fifth of all people who are aged 85+ are found in the bottom decile group.
b.    Poverty statistics for older people’s subgroups
Poverty results for subgroups of older people are presented by making use of the 60% of median income for the overall population. This practice has now become a most accepted way of defining the poverty threshold in the context of computing financial poverty for the European countries (see Atkinson et al. 2002). The poverty incidence refers to the head-count measure (i.e. the proportion of people below the poverty line), whereas the poverty intensity is measured by the average gap between the income of the poor and the poverty threshold. Results are presented in Table 2.
Using the 60% cut-off, almost 23% of all older people can be identified as falling below the poverty line. The poverty incidence is relatively high for widow(er)s (29.6%), those who are divorced, separated or never married (26.6%), those who live independently (23.3%), those who live as tenants (25.9%), and those in the two oldest age groups (about 30%). Not surprisingly, these patterns of poverty risk groups are in line with the detailed analysis of the relative ranking of different subgroups of old age people in Table 1.
To sum up, results for both dimensions suggest deprivation in terms of health and income for the subgroup of older age groups (aged 75+) as well as for those who live as tenants. In contrast, an association with the labour market is associated with a positive outcome in terms of both health and income. Women — when compared to men, and after controlling for other important attributes (age, marital status, living arrangements, job status, and housing tenure) — are worse off in different dimensions of health. Those who live independently (e.g. single elderly persons) are more likely to be deprived in terms of income but do as well as others in terms of health.

Table 1: Relative position of subgroups of older people within the overall population: cumulative shares in income decile groups


Notes: (1) Income deciles are calculated based on net household income before housing costs for the overall population.
Source: Author’s calculations from the 1999 BHPS.

Table 2: Poverty incidence and poverty intensity amongst older people


Notes: (1) Poverty intensity refers to ‘median’ poverty gap, which is preferred over ‘mean’ poverty gap since it provides a better measure of the average income of poor people.
Source: Author’s calculations from the 1999 BHPS.

2 Sen himself makes this suggestion: “the income level of a family may be adjusted downwards by illiteracy and upwards by high levels of education, and so on, to make them ‘equivalent’ in terms of capability achievement’ (Sen 1997: p. 215).
3 The decile cumulative distribution provides the cumulative shares of the population up to (and including) each decile.

Pages: 1 2 3