On Measuring Child Poverty

Paul Allin and John Veit-Wilson

At their annual conference in September, the Royal Statistical Society organised a session on the government’s consultation on child poverty. With the next announcement on consultation now expected before Christmas, Paul Allin and John Veit-Wilson summarise the presentations and discussion.

It seems something of a pattern that governments consult, a community of interest gets all stirred up ... and then things go quiet for a long time. This looked to be the case with the consultation on better measures of child poverty launched in November 2012 and which closed in February 2013. Many poverty campaigners and researchers were worried by this consultation. It appeared to be presented as something on which the Secretary of State for Work and Pensions had already made up his mind. It favoured a multidimensional measure of poverty that seemed to dilute the reality of a family’s lack of income by introducing other aspects of deprivation, ignoring long-standing social science evidence that factors contributing to poverty are best treated separately from measures of outcomes that are the consequence of poverty. But then nothing more was heard.

So it was timely that the Royal Statistical Society (RSS) asked us to organise a session at their annual conference in Newcastle in early September. As Hetan Shah, the executive director of the RSS who chaired the session, said, “this topic is in the DNA of the RSS, which has always been concerned with how to measure the state of society, especially those who are less well off”.

The session was arranged as a symposium with four speakers, three presenting perspectives (government, political lobby group, research centre) on the case for multidimensional descriptions as being better than income inequality measures for identifying and counting families in poverty, plus a discussant of responses submitted to the consultation document. The presentations from the session are to be published by the RSS. This summary of what was said is relevant to understanding and responding to the continuing arguments between politicians and professionals.

First, Peter Matejic from the DWP/DfE Child Poverty Unit thanked respondents to the consultation and reassured the audience that no decisions had yet been made. The government view was that some common themes were emerging, including that there are no perfect measures, that the measures in the Child Poverty Act 2010 are supported but that poverty needs to be understood as wider than those measures, and “income should continue to be a key feature of any measures of child poverty”. Besides these themes, clarity about the causes and consequences especially of intergenerational and entrenched poverty was essential, and there should be “real action to tackle child poverty” including life chances. The potential causes of poverty were identified in the personal and social conditions listed in the consultation document, including the devastating effect parental addiction can have on child poverty. New measurement methods might be needed since this affects a relatively small number of families and so is not amenable to measurement through social surveys, including longitudinal surveys such as ‘Understanding Society’ that can reveal much about patterns of poverty over time. What is most important is to indentify the drivers of child poverty.

Matejic also explained that these drivers will be crucial for developing the new Child Poverty Strategy, which will be open to consultation later in 2013 for publication in Spring 2014. The government’s specific plans on measuring child poverty are likely to be folded into the strategy, with the implication that there may not be a separate statement responding to the 2012/13 consultation (although all contributions to the consultation should be published by the government). While not explicitly stated, it looks like the next General Election is already starting to focus thinking within government in this area.

Matt Tinsley from the right-of-centre Policy Exchange, which had been identified by insiders as the source of many of the ideas in the consultation document, denied it was the author and emphasised the need for accurate measures of “what we want to call poverty”, about which the government must be specific. This requires a measure which “does not discriminate”, meaning criteria which identify “the situation of all individuals (children) regardless of irrelevant factors”. Yet this did not include income, and much of his presentation was devoted to tables showing why a variety of household income inequality and other statistics concerned with material deprivation, hours of work and ‘worklessness’, over time and in different regions, would not meet what is needed. This is because existing statistics showed clear distortions and had little overlap with comparable measures of living standards, exaggerating some and concealing others. In particular, existing income inequality measures do not measure living standards directly. What is needed instead to measure whatever government approach to poverty is adopted probably demands statistically robust multiple measures of material circumstances.

Matt Barnes from NatCen, the National Centre for Social Research, reported on the study which it and Demos did into what it called poverty types. Like the Policy Exchange, the “impetus was to understand poverty from a multidimensional perspective”, so this presentation was not about counting children in poverty but describing their families. The aim was to develop “a toolkit to guide policy makers and practitioners” by offering “solutions for each poverty type”. Twenty normative indicators across a wide field were chosen, one of which was low income, and applied to households with incomes below 70% of the median (so omitting higher income households who might also be affected by these indicators). Five ‘child poverty types’ were described, though in fact these were typologies of how the households and in particular mothers managed or failed to do so.

The final presentation was by John Veit-Wilson. In summarising the chief issues raised in professional (as opposed to other public) responses to the government’s consultation he inevitably touched on many of the points made or overlooked by the previous three speakers. He reviewed well-established definitions of poverty which they had ignored. These ranged from the World Bank’s definition of poverty as a ‘level of well-being deemed to constitute a reasonable minimum by the standards of that society’, to the United Nations’ 2001 version and the consensual definition of poverty ‘as an enforced lack of socially perceived necessities`(Ravallion; Mack and Lansley, 1985).

He also reported widespread agreement even with the previous speakers that household income inequality is not in itself a poverty measure — though that is because the speakers think income is largely irrelevant, while social science respondents see an arbitrary percentile of income inequality as not reflecting the income that evidence shows is needed for social inclusion. Many respondents had drawn attention to the government incomprehension (if that is what it is) that descriptions of personal characteristics and problems shared by others among the non-poor population and even the very rich (like personal and family problems and parental addiction mentioned by the previous speakers) cannot be valid criteria of poverty by which to distinguish it from non-poverty. Children in poverty cannot be counted by describing their parents’ personal and social problems, however comprehensively, or the consequences of their lack of enough money to buy their way out of such commonly experienced problems. He described the government’s descriptive evasion of adequate incomes as a zombie argument about stereotypes instead of an attempt to find a better statistical measure of a serious human rights problem.

It was a lively session. Peter Matejic claimed that the current statistical definition of child poverty does not resonate with public opinion. But he failed to make it clear that this is not the case with the consensual method of defining poverty in relation to publicly chosen necessities – a method which is incorporated in the Child Poverty Act - which shows strong public support across widely different groups for what constitutes a minimum contemporary living standard.

Moreover, the government’s dependence on subjective stereotypical descriptions of who ‘most needs help’ risks losing the core idea of poverty. The nod to the social science community in the ‘common themes’ acknowledges the ambivalences in what he referred to as “an evolving field”, so while we await clearer signs of the government plans for measuring child poverty later this year, there is still much for social scientists and poverty policy campaigners to debate with both politicians and officials.

Paul Allin is a former member of the Government Statistical Service and now Visiting Professor in the Department of Mathematics at Imperial College London.

John Veit-Wilson is Emeritus Professor of Social Policy of Northumbria University, Visiting Professor in Sociology at Newcastle University and Vice-Chair of the Child Poverty Action Group (of which he was a founder member in 1965).


Martin Ravallion, Poverty comparison: A guide to concepts and method, Washington DC: The World Bank (1992, p 4);  Statement adopted by the UN Committee on Economic, Social and Cultural Rights (E/C.12/2001/10), para. 8; J Mack and S Lansley, Poor Britain, 1985, p 39.


Publication date: 
Oct 23 2013


NB: You need to be Logged in as a registered user of this site to post comments.

pallin's picture

On re-reading this, I would just like to clarify a couple of things we wrote in relation to Policy Exchange (PE). First, we had heard before the RSS event that PE had been involved in preparing the consultation document but Matt Tinsley did say at the RSS that PE were not involved, inndeed were slightly surprised by its contents and would have favoured something very different. Second, PE do not completely reject income as important but want to highlight that it is "open to a range of distortions which other measures are not". Hope this clarification is helpful, and apologies for any confusion caused. Paul

Thank you for the clarifications Paul. Sasha, Web Moderator

I very much doubt whether the distortions in earned income would have a substantive affect on the multi-dimensional method of measuring poverty. It is by far more 'scientific' and therefore more accurate than the 'consensual' method advocated by the writers. Perhaps we could look again at the American approach of a 'minimum necessary standard'?
pallin's picture

We still seem to be some way from a resolution to the issue of measuring child poverty. The statement in the link below from Work & Pensions Secretary Iain Duncan Smith and Chancellor George Osborne seems to indicate that the present measure continues until "better" measure of child poverty are developed: http://www.theguardian.com/commentisfree/2014/feb/26/conservative-child-poverty-strategy-george-osborne-iain-duncan-smith The government ministers identify a number of "root causes" of poverty although "we need to develop a better understanding of the dynamics that drive these deep causes of poverty ... [then] we need to start collecting much better data so we are measuring what we want to target instead of targeting what we happen to measure". This is presented as a Conservative Party commitment and part of their "long-term economic plan". Any bets on the timing of further work on developing measures? Not until after the 2015 general election?