Working Paper 37

Draft Status

Responding to Child Poverty in Northern Ireland:Components of an Antipoverty Strategy for Northern Ireland[1]

Eithne McLaughlin and Marina Monteith

Equality and Social Inclusion in Ireland Project

WP No 37

Introduction

This is the final working paper in a set of three addressing child poverty and social exclusion (the others, Monteith and McLaughlin 2005 and McLaughlin and Monteith 2005 have addressed issues of measurement and definition and documented the extent of child and family poverty in Northern Ireland the latter will also be reported for OFMDFM as McLaughlin and Monteith, forthcoming). It is difficult to overstate the importance of childhood poverty for both equality of opportunity & equality of condition (see also Stewart, 2005). In terms of the achievement of equality of opportunity,material inequality in childhood challenges the mythologies of equality of opportunity at a number of levels. The level playing field on which young adult members of contemporary democratic societies supposedly begin their lifelong competition for success in life does not look so level when one contrasts the state of ‘readiness for competition’ of the 18 year old from a poor family and locality with that of an 18 year old from an affluent background. The young adult from the poor background will step onto the playing field with objectively less ‘equipment’ - skills, educational qualifications, cultural and social capital, physical and mental health). H e or she is also likely to step onto the playing field with less confidence in the fairness of the competition, lower expectations of success and therefore less optimism about the outcome. On top of all this, it is likely that other players and the referee will respond differently to the poor than to the affluent young adult. Other players and the referee may even apply a different set of rules in the competition to one than the other. As Katz (1996) and McLaughlin (2005) argue it is part of the mythology of equality of opportunity and meritocracy to believe that the public educational system can and does compensate for the enormous differences in personal resources which result from the very different familial experiences and upbringings we have – the accident of our birth. The responsibility of the state to act to level out some of the effects of these accidents of birth is a responsibility across a number of domains: the public educational system; regulation of behavior in the labor market; the provision of supportive social services for families and parenting; the provision of social housing and the regulation and maintenanceof the quality of local environments; redistribution of income through the taxation and benefit systems and so on – appropriate action in all of these areas is required to reduce the extent to which our life chances remain tied to the accidents of our births rather than our own efforts and achievements. In this paper we propose action across a number of policy domains in order to reduce levels of child and family poverty in Northern Ireland. In addition we set out the targets for reductions which Northern Ireland should have in order to parallel those the UK government has committed to for the UK as a whole. Following on from working paper 36, the first part of this paper makes recommendations about the ongoing measurement and monitoring of child & family poverty.

Table 1 Summary of Child Poverty Prevalence risk) rates in Northern Ireland 2002/2003 by Poverty Measure

Measure and (Data Source) 1 / % Children poor
HBAI variants :
<60% (PSENI)Mc Clements scale NI median BHC
<60% (FRS )Mc Clements scale NI median BHC / 24
18
<60% (FRS) OECD scale NI median BHC
<60% (PSENI ) OECD scale NI median BHC
NEW UK child poverty measure
Absolute low income (PSENI)
Relative low income NI median(PSENI)
Relative low income GB median (PSENI) / 19
25
14
19
23
DWP mixed tier various combinations / 8 - 40
DWP mixed tier guestimate / 25
Consensual poverty (PSE-NI)
Children Lacking 3+ Necessities (PSENI) / 38
28

1 Note the unweighted FRS dataset for 2002/03 contains a sample bias problem

Table 1 shows how risk rates for child poverty in Northern Ireland vary from a low of 8 to a high of 40 per cent depending on the measure adopted and within that the choices made on a range of technical matters such as the equivalisation scale and the median income threshold adopted (see also McLaughlin and Monteith (2005b). Broadly speaking between a quarter and a third of Northern Ireland’s children and young people are currently growing up in poverty. The prevalence or risk rate of poverty and therefore the measure adopted matters because the scale of a social problem affects how both governments and the general public respond to it; the urgency or lack of it in that response and the resources and weight governments are prepared to put behind anti poverty measures. While much of the UK’s anti-poverty strategy is Westminster driven, it is important that the Northern Ireland Civil and Public Service and local politicians are fully informed about and representing NI interests and NI children in the development of UK wide antipoverty strategies. The consensual poverty measure at 38% is not an especially high out turn measure but it does indicate a higher level of child poverty than the traditional UK government’s poverty measures have done. The UK government’s new child poverty measures produce rates of between 14 and 40 per cent depending on the tier concerned. The wide range noted for the UK’s new mixed tier emphasizes the importance of the choices yet to be made by DWP and/or the EU-SILC team on the methods by which the new mixed sub-measure will be created. The headline child poverty rate is important politically, but it is also important to also bear in mind that measurement over time is a more reliable indication of the success or failure of government and social policies than cross-sectional headline counts of poverty. In measurement over time what is important is not so much the measure used but rather its consistent application across time points. Similarly, in spatial analysis what matters most is consistent application of measures across populations and territories, So for example, whatever combination method is chosen for the new EUSILC/DWP mixed measure, if it is applied consistently across all member states the result will be a major step forward in understanding and evaluating the performance of governments across Europe in eradicating poverty and promoting social inclusion.

How much does the choice of poverty measure matter?

As noted above the choice of poverty measure is not so significant where the objective is measurement of change over time rather than prevalence of the phenomenon at a single point in time. In the former case, validity is about the reliability and validity of the measurement of the time trend rather than the validity of measurement of the underlying phenomenon of interest. It is sometimes argued that it is best to report and utilize a number of measures of poverty. The argument against such pluralism however is that reporting of multiple measures may confuse the public and politicians and be a tactic to divert attention away from the ‘big picture’ and the need for this social problem to be addressed urgently. The alternative to pluralism – the creation of an expert consensus has been substantially progressed at European level by the EU-SILC Working Party on Income and Poverty Measurement. The mixed measure which will now be adopted by EU-SILC does not however appear to have complete working party unanimity behind it see Nolan (2005). Nonetheless it must be emphasized that much expert consensus on most of the key aspects of poverty measurement does now exist. For example it is widely accepted that mixed measures are better than income only measures. This is because the former take into account the impact of poverty durations and differential purchasing power on standard of living. Beyond that on purely scientific grounds there can be little doubt that the consensual poverty measure is the best measure of poverty available. This is because the method removes human judgment (and hence arbitrariness) from decision-making about how to combine the subcomponents of the mixed measure. It also removes the scope for political interference in measure construction. Within the consensual poverty measure statistical tests of differentiation produce an integrated combined income threshold with a number of indicators of deprivation. The best poverty measure politically however may be determined by other criteria. The political criterion for a best measure are probably not those of good science as set out by Gordon, 2000); but rather based on such pragmatic and instrumental factors as simplicity; congruence with existing public opinion; political impact; clarity of implications for antipoverty strategies and policies and so on. In other words criteria of ‘do-ability’ and ‘communication’.

Conclusions and recommendations on the Measurement of Poverty and Child Poverty

Poverty should be measured in terms of both indicators of deprivation and cash income thatis a mixed rather than an income only measure of poverty should be used. This position was previously argued by Dignan and McLaughlin 2002; Gordon et al, 2000; Hillyard et al 2003; and Monteith and McLaughlin, 2005). It has now been endorsed in the EU-SILC process and the UK’s new child poverty measure discussed in McLaughlin and Monteith, 2005. The consensual poverty measure developed by Gordon et al and first applied in Northern Ireland by Hillyard et al (2003) is the measure which best meets the standards of good science (Popper 1968, Gordon 2000). In the pragmatic world of policy and politics the best measure however may be that which is most fit for purpose rather than that which is most scientific in strict terms. In the context of Northern Ireland and poverty for example, the benefits of using measures which are comparable with those produced elsewhere e.g. at EU or UK levels may outweigh the purely scientific merits of one particular poverty measure. For the purposes of comparing population data across territories it is also the case that the most important factor is ensuring that like is being compared with like (apples with apples, pears with pears). This means that data on the prevalence or extent of poverty in one territory must apply exactly the same technical methods as in the other territory for the comparison between the two populations to be meaningful and legitimate, for example the same equivalisation scale, the same definitions of gross, net, pre or post transfer incomes and so on. If an income threshold is being used and the results compared across territories and populations it is arguable as to which median threshold should be adopted, that is whether the median from population 1 should be applied to population 2 or vice versa or whether a new median formed by merger of populations 1 and 2 into a population 3 should be applied to both. Current UK practice is to apply the GB median to UK sub- populations such as Northern Ireland and Scotland. Chapter 4 showed that this produces a higher level of poverty in Northern Ireland than would application of the Northern Ireland median as the poverty threshold in the case of child poverty yielding a 23 as against a 19% child poverty rate. The third factor to be considered in comparative poverty statistics is whether to compare income data pre or post social transfers and pre or post housing costs. It is generally agreed that the focus in poverty measurement should be at least on disposable rather than gross income, i.e. on post transfer income (post income transfer is income after tax deductions and/or tax-benefit additions to earnings). The removal of housing costs from net income is more contentious. Housing costs are not inherently more necessary or even more fixed than other expenditures e.g. those of food, fuel and care. The latter vary between NI and GB as much as do the former. Accordingly there is no abstract right answer as to whether income statistics and income poverty comparisons within the UK should be before or after housing costs. Not surprisingly prices are generally lower in lower than higher income areas within the UK. Northern Ireland however has been an exception to this rule, being both a low income and a high price area. This fact accentuates the desirability of adoption and use of mixed poverty measures (that is poverty measures which measure the standard of living achieved as well as cash income where ever and whenever possible

We recommend that analysis of poverty within Northern Ireland should utilize the consensual poverty measure wherever this is consistent with the validity of the comparison being undertaken.

(i)Analyses over time and between social categories within the Northern Ireland population where comparability of statistics between territories is not an issue should utilise the consensual poverty measure adopted in Hillyard et al (2003). The indicators of deprivation element of the consensual poverty measure will need to be periodically updated, replicating the original methodology in order to maintain their scientific validity.

Our conclusions and recommendations as to the analysis of poverty risk rates between territories are that analyses using the relative low income component or tier of the DWP’s mixed poverty measure should apply GB median income as the income poverty threshold and should be reported as post transfer risk rates both before and after Housing Costs.

(ii)We note that North-South as well as East-West analyses were recognized as desirable by The 1998 Belfast Agreement. Their desirability is further underlined by recognition that such analyses help us to better understand the relationship between economic development, economic growth and trends in poverty over time. Under the new EU-SILC practices North-South of Ireland comparisons will be easier in future as all data will use the modified OECD equivalence scale, there will be a common income only measure of 60% median income and a common mixed measure (full details of which were not available at the time of writing).

(iii)Our final recommendation on analysis and reporting on poverty in Northern Ireland is that an agreed language and set of definitions should be adopted and enforced in all government publications in order to ensure more transparency in and accessibility of policy debate and to enable the participation of the third sector and the public more generally. We recommend that the NICS restrict the word poverty to refer to statistics based on mixed measures and low income or income inequality to statistics based on income data alone. The word deprivation should be used to refer to the enforced lack of those material items, activities and services regarded as customary and necessary parts of our society’s ‘normal’ way of life but which a person or family cannot enjoy because of a lack of money. The term social exclusion refers to both the effects and the causes of poverty and deprivation.

The Measurement of Child Poverty

The discussion and recommendations above refer and apply to research, policymaking and reporting about poverty in general.

We now turn the focus to research and policy making about child poverty specifically.

Given the special urgency which attaches to child poverty:

we recommend that the NICS should measure and monitor both child and family poverty in general, themost extreme or deepest forms of it and child deprivation. In addition monitoring should be undertaken of and targets set in relation to the unusually highly spatially concentrated distribution of child poverty within Northern Ireland.

As outlined in Chapter 4 there are a number of measures of extreme poverty available and competing for adoption. These are the new UK mixed income/deprivation measure (combining adult and child deprivation with a debt indicator and low income as shown in chapter 4); the new UK absolute low income measure; the RoI’s consistent poverty measure; the new EU-SILC mixed measure confusingly this will also be called consistent poverty but will be differently calculated to the present RoI consistent poverty measure). Save the Children’s measure of severe poverty; Bradshaw and Finch’s core poverty measure. As Table 2 shows the prevalence of THE deepest child poverty in Northern Ireland will range from 8% to 11% depending on which of these measures is adopted.

Table 2 Prevalence of deepest child poverty in Northern Ireland by measure.

%

(i)RoI Consistent Poverty 9

(ii)DWP mixed measure 8

(iii)DWP absolute low income measure14

(iv)Bradshaw and Finch core poverty, all 3 dimensions11

(iv)We recommend additional monitoring over time of extreme child poverty and child deprivation firstly through periodic use of the Save the Children severe poverty measure applied to the Poverty and Social Exclusion dataset. This measure uses a combination of a very low income level (less than 40% of equivalised median income) and a full complement of consensually agreed adult and child deprivation indicators. In comparison, the EU-SILC, ROI and DWP measures use indicators selected to some extent arbitrarily by experts. The Bradshaw and Finch core poverty methodology also uses consensually derived child deprivation measures but was designed to examine the degree of overlap between measures rather than the prevalence of extreme poverty per se. The Save the Children Severe Child Poverty measure combines the lowest acceptable level of relative income with indicators of both adult and child deprivation and is in our view the most suitable measure for monitoring progress in tackling the most extreme child poverty over time.