Project summaries

Poverty reduction indicators: a discussion paper

In January 2001, the Combat Poverty Agency commissioned the New Policy Institute to undertake research into the use of poverty reduction indicators. The study was carried out within the context of a process set up to review the targets and establish indicators under the National Anti-Poverty Strategy (NAPS). Various working groups had been established to propose new targets and indicators in 7 key thematic areas which included education, health, housing, income, unemployment, rural, and urban. 3 cross-cutting themes were also identified which each working group had to incorporate into their discussions; these included women, children and older people, giving a total of 10 subject areas in total being examined.

The objectives of the research were to provide information on relevant indicators of poverty reduction, drawing on national and international experience, to provide data requirements for the poverty reduction indicators, and to provide a structure for assessment. Rather than try to come up with definitive conclusions, the idea behind the research was to contribute to the debate through providing material to support the development by the NAPS Unit of an ‘official’ list of indicators for the future strategy. As well as supporting this process through its discussion of possible subjects for indicators, this report also aims to help subsequent monitoring through its discussion of possible definitions of indicators for particular subjects.

It takes the form of a handbook which is designed to provide practical assistance in the process of indicator selection and definition. As such, it provides:

An introduction to the use of poverty reduction indicators (Chapter 2), including a discussion of the differences between indicators and targets.
Criteria for indicator selection (Chapter 3). In effect, these proposed criteria provide a basis for excluding particular indicators, for example, on the grounds that they are not obviously to do with poverty (e.g. because they are not more prevalent for lower income groups than the rest of the population) or that they are not practical (e.g. because the relevant data is not available).
A discussion of possible indicators (Chapter 4) for each of the 10 topics identified as key to the anti-poverty strategy (e.g. income, education, children, rural, etc). This discussion aims to stimulate and focus the debate by identifying the key areas which the indicators should cover.
The indicator selection process (Chapter 5), which sets out a proposed process for indicator selection. Somewhat different processes are suggested for the subject-oriented topics identified (e.g. income, education, etc) than for the age, gender or geographic topics/groups identified (e.g. children, rural, etc).

The material is supported by a set of appendices – one for each topic area – which provide an initial assessment of the merits or otherwise of many of the possible indicators.

The use of poverty reduction indicators

An indicator is a number or set of numbers. As such, it has to be defined in precise terms. It also requires that data is available so that it can be calculated. In practice, the greatest constraint on the selection of indicators is the availability of data and a major challenge will be to find the data which allows a comprehensive set of indicators to be selected, and to avoid a situation where data limitations mean that the set of indicators becomes unsatisfactory.

Indicators are typically used for two major purposes: to monitor changes over time and to monitor differences between different groups in the population. Whereas it may be reasonably safe to draw conclusions about trends and about differences between groups, the absolute numbers must be treated with more caution.

A second major challenge is to select a manageable number of indicators which adequately cover the wide range of issues associated with poverty such that they collectively provide a reasonably comprehensive picture of the extent and nature of poverty and how it is changing.

The choice of indicators is a less politically sensitive subject than the choice of targets. Indicators can be selected relatively freely and multiple indicators on the same subject can be used if need be. In theory, the set of indicators that are used should remain stable over time.

Proposed criteria for indicator selection

It is suggested here that the total number of number of indicators should be between 50 and 100.

To be considered for inclusion as a poverty reduction indicator, a statistic should ideally meet all of the following criteria:

1. Have something to do with poverty (not just health, etc):

Is more prevalent amongst those on low incomes.
Is not entirely an issue of free choice.

2. Relate to something that one wants to monitor:

Is a matter of concern.
Is important in its own right.
Ideally, is indicative of wider conditions.

3. Be quantifiable on a regular and repeatable basis:

Is capable of precise definition.
Relevant, reliable and authoritative data is available (or at least obtainable).
Can be broken down by income group or equivalent (cf. criterion 1 above).
Data will be available on the same basis in the future.

4. Other:

Understandable, such that it cannot be mis-interpreted easily.
Robust to changes in government administrative rules.

A couple of these criteria merit further discussion. Taking “have something to do with poverty” first: one risk is that some of the selected indicators are more to do with the subject (e.g. health) than with poverty. For example, premature mortality is widely recognised as an important indicator of health outcomes. But premature mortality is on a downward trend largely because of developments in healthcare, food and non-smoking. So, premature mortality per se is arguably more to do with health than poverty. A more appropriate definition of a premature mortality poverty reduction indicator would be the difference between the levels of premature mortality amongst those on low incomes and for the population as a whole. Putting the point another way, poverty reduction indicators in the area of health should mainly be about health inequalities rather than just health per se. Generalising this point, a key criterion for any indicator to be a poverty reduction indicator is that it is more prevalent amongst those on low income than amongst the population more generally.

Second, taking “be quantifiable on a regular and repeatable basis”: the Economic and Social Research Institute’s (ESRI) Living in Ireland Survey provides sufficient data for a comprehensive range of income and deprivation indicators, with break-downs by age, occupation etc. also available. In all the other areas, however, data availability will be a major constraint, particularly given the wish for data which can a) be broken down by income group or equivalent and b) which is available on the same basis from year-to-year. So, selection of some of the indicators may well involve making the best pragmatic choice given the data that is available. It would, however, be very deleterious if important subjects were not adequately covered. Rather, we suggest that indicators in these areas still be selected, even if this means that once-off and partial data sources have to be used initially, and that these data gaps be explicitly flagged up for possible collection in the future.

Many of the criteria above will apply equally to all developed countries. In other words, the long list of possible indicators will be similar in Ireland as in other EU countries. The major exception to this is the data availability criteria, and thus it is these criterion which are in practice likely to dominate subsequent shortlisting. Views on topicality/importance will also have an impact in judging which of the possible indicators to prioritise (for example, rural poverty is clearly an issue of importance in Ireland, but is considered less so in the UK).

Possible indicators

A summary list of possible indicators is provided in the table overleaf. Although presented separately by subject, it is vital that the total set of indicators emerging from the discussion process is reviewed as a whole to ensure that it is a coherent and balanced set.

Regarding particular areas, we suggest that:

With income and deprivation, the indicators should cover both relative and absolute income measures, and should also cover lack of essential goods and services.
With unemployment and work, the indicators should cover quality of work (e.g. low pay) as well as unemployment per se.
With education and housing, the major challenge is likely to be data availability.
With health, the key will be to select indicators which adequately reflect wider problems and which focus on health inequalities rather than health per se.
With children, as well as education and health, the income, deprivation and unemployment of the household is also relevant. Another subject for possible indicator development is ‘social stability’, which would involve those things which make it difficult for children to lead a normal life.
With women, any of the subject-related indicators could be relevant, looking for those which are markedly worse for women than for men. Beyond that, it is not clear what, if any, indicators could usefully be added which are specific to women.
With older people, the income and deprivation, health and housing indicators are clearly relevant. Another broad area for possible indicator development is ‘quality of life’ covering such subjects as isolation, anxiety and support.
With urban, any of the subject-oriented indicators could be relevant, looking for those which are markedly worse in deprived communities.
With rural, any of the subject-oriented indicators could potentially be relevant, looking for those which are markedly worse in rural areas. Other broad areas for possible indicator development concern access to services and isolation.

The suggested indicator selection process

For each subject area (income, health, etc), the proposed process for selecting relevant indicators is as follows:

Identify the key topics to be covered, drawing on the discussion in chapter 4.
For each topic, identify a long list of possible indicators even those for which data might not currently exist, flagging these data gaps for possible collection in the future
Reduce each long list to a short list.
Select from the short list.
If there are no indicators, review possible data sources again.

For the age, gender and geographic topics, the proposed process is as follows:

Consider which of the subject-oriented indicators could usefully (and practically) be broken down, either by the group in question (gender, geography, etc) or by a related group of concern (female headed households, lone parents, farmers, etc).
Think about what, if any, other topics could usefully be covered and follow the same process as for the subject-oriented indicators. It is suggested that additional unique subjects only be included if there is a clear case for so doing and that, in some cases – such as women and urban – there might not be such a case.

Summary of possible indicators

Subject	Topic area	Possible indicators
Income	Relative low income (moving thresholds)	numbers below thresholds (40%/50%/60%, mean/median); income at the Nth percentile, and the ratio between this and average incomes
	Absolute low income (fixed thresholds)	as for relative income, but using thresholds fixed in time (adjusted for inflation)
	Deprivation	lacking basket of necessities; lacking particular necessities; lacking access to particular essential services
	Combination	the ESRI index (with either a fixed basket of goods and/ or one which is changed over time depending on society norms)
	Intensity	persistently on low income
	Reliance on benefits	numbers of people solely reliant on benefits; levels of basic state benefits
	Financial difficulties	numbers in debt; numbers with self-reported financial difficulties
Unemployment and work	Unemployment	ILO unemployment; numbers who want work; long-term unemployed; benefit levels
	Low pay at work	below minimum wage; below X% of average hourly earnings
	Quality of work	lack of access to training; job insecurity
	Polarisation of work	by geography, family type, housing tenure, qualifications, etc
Education and qualifications	School leavers	lacking basic qualifications, numeracy skills or literacy skills
	Younger children	lacking basic qualifications, numeracy skills or literacy skills
	Adults	lacking basic qualifications, numeracy skills or literacy skills
	Disrupted education	early leavers; excluded from school; truancy; bullying
Health	Overall mortality	concentrations of premature mortality
	Quality of life	long-standing illnesses or disabilities; other measures of morbidity
	Healthy lifestyles	obesity; drug/solvent/alcohol mis-use; smoking (or cancer rates)
	Children’s health	low birthweight babies; nutrition; accidental deaths; infant mortality; respiratory illnesses; immunisation take-up rates
	Mental health	depression / anxiety; mental illness; suicides
	Access to healthcare	unclear, but should be outcome-oriented rather than input-oriented
Housing	Quantity	homelessness; temporary accommodation; affordable housing
	Quality	damp; over-crowding
	Fuel poverty	costs of essential utilities
	Risk	re-possessions; mortgage arrears
Children	Health	the relevant health indicators, some of which are child-specific
	Education	the relevant health indicators, some of which are child-specific
	Income	the relevant income indicators for households with children
	Work	the relevant work indicators for households with children
	Social stability	in care; in institutions (criminal/non-criminal); convicted of a criminal offence; under-age pregnancies; parents divorce
Women	All	other indicators, split by gender or by group of concern (e.g. lone parents, female-headed households)
Older people	All	other indicators , split by age (e.g. income, health, housing)
	All	other indicators, split by pensioner type (singles/couples, younger/older)
	Isolation	numbers who lack social interactions with others, including relatives; numbers of undiscovered accidents; numbers who lack a telephone
	Anxiety	numbers afraid of going out; numbers who worry about being burgled; numbers who worry about paying bills, e.g. for essential services
	Support	coverage of social care; waiting times for essential operations
Urban	All	relevant indicators, split by type of district electoral division (as defined by a combination of population density and level of deprivation) by size of conurbation or by geography
Rural	All	relevant indicators, split by geography (rural/urban, large town/small town/village, remote/accessible rural, etc)
	Access to services	access to banks; access to village shops; time taken to reach essential services (e.g. health services)
	Isolation	lack of car ownership; access to public transport; levels of participation in civic society

top

New Policy Institute, 003 Coppergate House, 16 Brune Street, London E1 7NJ

Tel: 020 7721 8421 | Fax: 020 7721 8422 | info@npi.org.uk | www.npi.org.uk