Bridging the gap between the measurement of poverty and of deprivation

Abstract

One way of measuring the deprivation or poverty of persons is to use monetary-based measures: a person is regarded as ‘poor’ if his/her income (or expenditure) falls below a poverty line value. Such an approach – usually termed poverty analysis – has spawned a large literature embodying several sophisticated measures of poverty. The downside to this is that low income or expenditure may not be very good indicators of deprivation. Another way, usually termed deprivation analysis, is to define an index whose value, for each person, is the number (or proportion) of items, from a prescribed list, that he/she possesses: persons are then regarded as ‘deprived’ if their index value is below some threshold value. This offers an alternative method of identifying deprived persons. The downside of deprivation analysis is that it measures deprivation exclusively in terms of the proportion of deprived persons in the total number of persons. The purpose of this article is to bridge the gap between poverty and deprivation analysis by constructing a wider set of deprivation measures and showing, with data for Northern Ireland, how they might be applied.

Acknowledgements

The analysis contained in this article is based on the Poverty and Social Exclusion in Northern Ireland (PSENI) project's data set and I thank the project leaders – Paddy Hillyard, Eithne McLaughlin and Mike Tomlinson – for permission to use these data though, needless to say, I am entirely responsible for the contents of this article.

Notes

Zheng (1997) provides a good review of this literature.

In order to account for differences in tastes, households that do not possess an item are distinguished by whether they did not want it or they wanted, but could not afford it. This meets Veit-Wilson's (1987) criticism of Townsend's (1979) work.

For details see Hillyard et al. (2003).

It is important to emphasize three aspects of Sen's measure: it takes account of the number of deprived persons, relative to the population, through H, the headcount ratio; it takes account of the depth of their deprivation through I, the deprivation gap ratio; it takes account of relative deprivation through G^P , the Gini coefficient calculated on the achievement.

When α = 0, the FGT index is the head count ratio, H = M/N; when α = 1, FGT(y;1) = HR, where R = 1 − μP/z is the achievement gap ratio; when α = 2, the FGT index incorporates the idea of ‘relative deprivation’, as measured by outcome inequality among the deprived households.

The following items were either owned, or not owned but desired, by all of 1542 respondents: a television; a telephone; a refrigerator; a dry, damp-free home; a washing machine; a video recorder; home contents insurance; a microwave oven; fresh fruit and vegetables; deep freeze; central heating; a vacuum cleaner; new, not second-hand, clothes; a warm water proof coat; two pairs of strong shoes; a good outfit to wear on special occasions; a meal with meat, chicken or fish every other day; enough money for keeping the home in a decent state of decoration, for replacing worn out furniture, for replacing or repairing non-functioning electrical goods, and for paying heating, electricity and telephone bills on time; a small amount of money for the respondent to spend on himself/herself; regular savings of £10 a month.

The mean and median ‘possession’ scores, over the 1542 responses, were 94.2 and 100, respectively, implying that at least half the respondents possessed all the 23 items listed above. The distribution of the possession indicator was heavily skewed (with a skewness value of −2.63) with the observations bunched at the right of the distribution.

Bought cheaper cuts of meat or less than you would have liked to buy; gone without fresh fruit and vegetables; bought second-hand, instead of new, clothing; continued wearing worn out clothing because you couldn’t afford replacements; put off buying clothing for as long as possible; relied on gifts of clothing; continued wearing worn out shoes because you could not afford replacements; put up with feeling cold because you could not afford heating costs; stayed in bed longer to save heating costs; skimped on food so that others in the household would have enough to eat; postponed visits to the dentist; not picked up a prescription; gone without, or cut back on, visits to family and friends; gone without, or cut back on, telephone calls to family and friends; gone without, or cut back on, trips to the shops or local places; spent less on hobbies than you would like; not gone to a funeral you would have liked to attend because of the costs; cut back on visits to the local pub; ever used less than you needed of gas, electricity, and the telephone because you could not afford it.

The mean and median ‘economizing’ scores, over the 2407 responses, were, respectively, 85.53 and 94.74. This points to a heavily skewed distribution in which the observations are bunched to the right of the distribution and, indeed, the skewness value for the distribution was −1.67.

Anyone with an ‘economizing score’ less than 68 was regarded as deprived in terms of ‘economizing’ and any person with a ‘possession score’ less than 75 was regarded as deprived in terms of ‘possessions’.

The Sen index, as discussed earlier, is an equity sensitive deprivation index (in the sense that it takes account of inequality in scores between deprived households) and its values are shown in Table 1. The difference between the achievement gap ratio and the Sen ratio is, as shown earlier, due to the fact that the latter is equity sensitive while the former is not. This difference was greatest for single parents (a rise from 13.98 to 19.68 on the economizing indicator and a rise from 9.43 to 13.43 on the possessions indicator) suggesting that relative deprivation (as defined earlier) was greatest for this group.

The deprivation risk, ρ _k , is the ratio of a group's contribution to deprivation to its contribution to the population: ρ _k  > 1(ρ _k <1) means that it contributes more (less) to deprivation than its population share warrants. If the norm for deprivation risk is taken to be unity, then, say, ρ _k  = 1.3 means that the deprivation risk for members of group k is 30% above the norm; similarly, ρ _k  = 0.82 means that the deprivation risk for members of group k is 18% below the norm.