Written evidence from the Centre for Analysis of Social Exclusion (CASE) at LSE (CPM0043)


Our response draws on the work of current and former CASE researchers: Tania Burchardt, Kerris Cooper, John Hills, Rana Khazbak, Ruth Lupton, Abigail McKnight, Polina Obolenskaya, Ilona Pinter, Mary Reader, Nick Roberts, Tom Sefton, Isabel Shutes, Kitty Stewart, Polly Vizard and Jane Waldfogel.


The main headline measure of child poverty should be the share of children living in households with less than 60% of median equivalised income after housing costs.

No single poverty measure can give a full picture of child poverty, so a series of additional measures should also be used, along with multi-dimensional indicators of children’s wider well-being.

Targets for poverty reduction have proved very effective in keeping a clear policy focus on child poverty. Concerns that they distort decision-making are not supported by the evidence of the last two decades. But targets do need to be treated reflectively within a broad child poverty strategy, with attention to wider indicators and to groups of children who are not counted or who are hidden in existing statistics. 

There is very strong evidence that poverty has a negative impact on children’s well-being and development. This means the measurement of child poverty and the adoption of a serious child poverty reduction strategy are of critical importance, for children themselves and in the interests of the economy and society.


The Centre for Analysis of Social Exclusion (CASE) is a multi-disciplinary research centre at the London School of Economics and Political Science. Established in 1997, our research examines social disadvantage and the role of social and public policies in preventing, mitigating or exacerbating it. We understand social disadvantage to be multidimensional, and to be best understood in a dynamic or lifecourse perspective, with individual, family, local, national and international aspects.


We have published extensively on multiple aspects of poverty and disadvantage in the UK over the last two decades, including causes, consequences, the effectiveness of policy responses and questions of measurement. This collective response draws on research evidence from publications and accumulated wisdom from across the centre.

How should child poverty be measured and defined?

Poverty is a lack of sufficient material resources to meet human needs. We stand by Peter Townsend’s 1979 definition:


"Individuals, families and groups in the population can be said to be in poverty when they lack the resources to obtain the type of diet, participate in the activities and have the living conditions and the amenities which are customary, or at least widely encouraged or approved in the societies to which they belong. Their resources are so seriously below those commanded by the average family that they are in effect excluded from the ordinary living patterns, customs, and activities.” (Townsend, 1979, p.31)


Operationalising this definition of poverty is not straightforward, and it is difficult to capture all aspects using a single poverty measure. Nonetheless, if a single measure is required, the approach used widely across Europe remains our preferred one: child poverty should be measured as the share of children living in households with income below 60% of median income, equivalised for household size.


In the UK context, where housing costs are large, variable and largely reflect factors beyond household control, the question of how housing costs are treated is important. As a single headline measure, our preferred measure would be an After Housing Costs one. Some groups of children are particularly affected by high housing costs, which can put them at a greater risk of after housing cost poverty.[1] However, for a fuller picture (and to aid international comparisons) a Before Housing Costs poverty measure should also be tracked.

There is also good reason to keep track of an ‘anchored’ poverty measure, where household income is compared to an income poverty line that is fixed over time, though adjusted for inflation. These measures are sometimes called ‘absolute’ though there is nothing absolute about them; they take the relative line at an arbitrary point in time and show how living standards at the bottom compare to this line. This can be a very useful indicator, particularly in times such as boom or recession when median incomes are changing rapidly. However, it should not be the single or main measure of child poverty. In times of economic growth we can expect progress against this measure without any government action, and yet families may still be left increasingly further behind and excluded from ordinary living patterns. We know that our colleague John Hills would have emphasised this point: if an anchored measure is getting worse, that is really worrying. If the anchored measure is improving but the relative measure is getting worse, there is more work to do. Only if poverty is falling against both a relative and an anchored measure should we feel that progress is being made.

Direct measures of deprivation such as indicators of material deprivation are also useful as they are able to avoid the problem that equivalisation for household needs is inevitably partial and imperfect. We know for example that income poverty measures do not take account of the extra costs associated with disability and therefore overestimate living standards in households with a disabled member (Zaidi and Burchardt, 2005). Material deprivation can be seen to capture living standards more directly. Material deprivation indicators can also be useful in focusing on children’s own experiences; if income is not evenly shared within the household income poverty measures may underestimate or overestimate children’s living standards. For example, we know that parents seek to protect children from the impact of poverty by disproportionately absorbing the income shortfall (e.g. by eating less themselves). As a third type of indicator, then, a combined income and material deprivation indicator (e.g. living in a household below a certain income line and falling below a set material deprivation standard) is useful. However, material deprivation indicators have their own imperfections. For one thing, results will be influenced by the specific indicators included. Further, like the anchored measure, we would expect improvement in these indicators over time almost by default if the indicators remain unchanged; it is a low bar for policy. The indicators can (and should, in the medium- and long-term) be updated – but this interferes with our ability to assess progress.

For a fuller picture, it is also very useful to have measures of poverty persistence and of poverty depth.

To sum up, we recommend that child poverty is measured using a series of indicators, with the first one used as the main ‘headline’ indicator:

Our conclusions here come out of research, discussion and reflection on this topic within CASE over two decades and more. In addition, they are supported by our analysis of responses to a 2012-13 government consultation on child poverty measurement (Stewart and Roberts, 2018). This consultation revealed very strong support for income-based measures across those working in local authorities, children’s charities and frontline services as well as academic researchers. Of the 251 responses, just two advocated removing income from poverty measurement, while only 10 in total would have liked to see income downgraded within the government’s approach to measurement – i.e. treated as just one indicator among a number of others in a multidimensional framework, rather than given special or headline status. Responses also overwhelmingly reveal a relative understanding of poverty, with an understanding of household and children’s needs that is strongly linked to social norms.

The Social Metrics Commission’s recent proposals

Recent work done by the Social Metrics Commission has been useful and important in underlining the imperfections in the standard approach to measurement. On the resources side, SMC points to the impact on household disposable income of large costs which are beyond the household’s control, such as childcare. They also argue that the resources measure should take liquid assets into account, as these can be drawn down when incomes are low. In terms of household needs, their work highlights the challenges of adjusting effectively for the extra costs of disability. They also argue that a moving average of median income over the last few years provides a better way to capture social norms than median income in the contemporary year.

These are all important issues. The downsides to incorporating the recommendations into national headline poverty measures are threefold (and linked): measurement limitations, consistency, and a hidden complexity in the final measure in which certain assumptions are crucial but not visible. First, some of the things we might like to adjust for, such as disability costs, are hard to measure using existing data (though we would support investment in survey capacity to improve our ability to do it). Measurement issues in turn lead to challenges of consistency. For example, in relation to the deduction of unavoidable costs from the income measure, the SMC acknowledge that social care costs should in principle be deducted along with childcare costs, but current data does not allow for this. This means that the SMC measure ends up deducting childcare costs, including high spending on childcare by better off families, while not deducting essential costs of personal care for pensioner households. This gives us an overall measure which seems no more accurate than household income at capturing a household’s ‘true’ level of disposable income, while having sacrificed simplicity and transparency.

Similarly, because we cannot actually equivalise for the extra costs and needs associated with disability, the SMC propose as a middle ground that benefits aimed at compensating for extra costs (e.g. PIP and DLA) should at least be deducted from the income measure. This makes sense if we wish to compare the living standards of households with and without a disabled member: extra costs benefits are aimed at compensating for extra costs of disability, and including them in the income measure will overestimate the living standard of the disabled household relative to the non-disabled household. However, excluding extra costs benefits from a measure used to track income over time is deeply problematic. These benefits are the main lever the government has for influencing living standards in households with a disabled member, especially where disability is severe. Excluding these benefits means we see a flat line for poverty rates in households with a disabled child during a period in which the real value of disability benefits first rose and then fell, with obvious material impact on households affected.

Finally, in relation to the incorporation of liquid assets, we have serious reservations about an assumption that households with low income should be expected to draw down all their savings in the course of a year, which is the assumption the SMC makes (adding 1/52 of liquid assets into the measure of weekly income).[2] Households that are struggling to make ends meet week-to-week on their income, while also seeking to protect a level of savings as a buffer against future or unforeseen expenses, should be considered to be living in poverty. Even means-tested universal credit ignores £6,000 in savings before starting to assume that some savings can be spent down. Other measures such as material deprivation indicators can be (and are) used to identify the number of households that are both income poor and unable to save.

Our strong preference is therefore to continue with the suite of measures highlighted above, while also continuing to consider additional costs and challenges, and developing supplementary indicators which cover them.

The importance of multi-dimensionality

Finally, we should underline our very strong belief there are many other important dimensions to disadvantage affecting children beyond their material living standards. These include (but are not limited to) their educational opportunities, their access to health services, and their wider environment, including safe streets and parks and clean air. We think it is essential that wider dimensions of disadvantage are also measured and monitored, and we have done extensive work in CASE along these lines (Holder et al, 2011; McKnight et al, 2017). We also think it is important to look at measures of children’s well-being as defined by children themselves, and to improve understanding of how poverty links to children’s wider opportunities, and their capability to achieve what they themselves see as their valuable beings and doings. (The Children’s Society has a valuable wellbeing index based on children’s definitions of their well-being.)

However, we think these wider dimensions should be measured alongside measures of child poverty as set out above, and not in replacement of them. 

The measures of child poverty changed in 2016. What has the impact of those changes been?

It is clear that the policy changes announced in the Welfare Reform and Work Act (WRWA) of 2016 and introduced in late 2016 and spring 2017 contributed to a sharp increase in child poverty, which had already been rising from 2013/14 as a result of earlier cuts and reforms to social security benefits (Cooper and Hills, 2021; Stewart and Reader, 2021).

Have the changes to the child poverty measures made a difference? While cuts to social security support are behind the increases in child poverty, rather than the change in measurement per se, the measurement changes have made it much more difficult to hold the government to account on these increases. It is fortunate that the House of Lords passed an amendment to the WRWA which ensured that the income measures continue to be published annually, so the data are still available to researchers and the general public. It would have been even more disastrous for the ability to hold government to account had this amendment not been included. But the WRWA removed the ‘teeth’ that the Child Poverty Act contained, which would have required government to report annually on progress towards the targets, to explain any deviation, and to set out a strategy for getting back on track. The changes to the measures have let the government off the hook.

The new measures that the WRWA introduced were indicators of worklessness and educational attainment. It is useful to have these available but they are NOT measures of child poverty. We write more on the relationship between these measures and poverty in a further question below.

What were the advantages and disadvantages of having a set of targets for reducing child poverty?

The advantages of the targets are clear. They focused policy attention on the problem and ensured that child poverty had to be given political priority. They also operated as a signal, making it clear that child poverty was considered an important social goal with cross party support. This may have increased public support for action to tackle child poverty (though we have no direct evidence of this).

The main disadvantage of targets is that they can potentially have distorting effects, leading to a focus on the targeted measure at the expense of other important outcomes. One example is that the chosen targets were based on income Before Housing Costs, meaning less attention was paid to After Housing Costs poverty, and there was indeed much less progress in reducing poverty AHC than BHC (Cooper and Hills, 2021). In some sense this is a testament to the effectiveness of the target – while also showing its limitations.

Similarly, having targets for headcount and not poverty depth may have meant less spotlight on areas where the social security system is failing most severely. One example of a group whose needs are missed is children in asylum seeking families and more generally those affected by ‘no recourse to public funds’ restrictions and therefore outside of the mainstream social security system (Pinter et al, 2020). A headcount child poverty target fails to take any account of their situation. The lowering of the benefit cap, or the application of sanctions, would also not show up in headcount figures if families affected were already below the poverty line.

We think these are not reasons to abandon targets, but instead to have a series of targets, alongside a strategy which is open and explicit in thinking about the limitations of a target-driven approach, and how to supplement it. Alongside publication of headline poverty measures, we would recommend annual analysis of statistics analysing child poverty disaggregated by children’s characteristics, including family size, ethnicity and lone parenthood. We also strongly advocate attention to tackling “data exclusion”, which means that some groups of children remain missing or invisible in available statistics (Vizard et al, 2018). This includes young carers, children in recent migrant families and Roma, Gypsy and Traveller children. These characteristics are not currently identified in the Family Resources Survey and are therefore not captured in the annual published HBAI child poverty statistics.

Finally, in principle, there is also a broader problem with income poverty targets: they may lead to policy efforts being focused on income alone, meaning other important dimensions of disadvantage are neglected. In practice there is no evidence that this was the case in the UK. Our assessment of the Labour Government’s record pointed to extensive action on wider aspects of inequality and deprivation, including investment in early years services such as Sure Start, substantial boosts for health and education funding, and additional interventions targeting extra support to disadvantaged areas, alongside increased financial support for households with children (Hills et al, 2009; Stewart, 2013). In contrast, since the targets have been scrapped, rising child poverty has taken place alongside stagnation or cuts (sometimes severe) in funding for services for children (Stewart and Reaer, 2021; Vizard and Hills, 2021). But in principle such a distortion is possible, lending further support to the importance of monitoring a range of indicators of different dimensions of disadvantage and well-being, alongside indicators of poverty.

What has been the effect of removing from law the targets in place between 2010 and 2016?

It is hard to say directly what the effect of removing the targets has been. We know that child poverty has risen sharply, especially in larger families and families with a young child. Removing the targets did not directly cause this, but it has meant less policy attention on the topic. With the legal targets, the government would have been required to explain to Parliament why the target had not been met and to set out a strategy for progress going forward. It would surely have been much more difficult politically for the government to take decisions such as the introduction of the two-child limit, with direct and unambiguous negative effects on poverty, had the legal targets been in place. 

Perhaps even more important than the targets themselves is the loss of the structures around them, including the requirement to have national and local child poverty strategies, and for local authorities and other “delivery partners” in England to conduct a local needs assessment and to work together to tackle child poverty.

What is the impact of child poverty and how can it best be measured?

The evidence is very clear that child poverty has negative impacts on children’s outcomes. We answer this question from a position of authority in CASE, having conducted a systematic review of this evidence, including only studies which use situations and techniques which allow researchers to examine causal relationships with reasonable confidence, through quasi-experimental approaches and use of longitudinal data (Cooper and Stewart, 2013; Cooper and Stewart, 2020). We find strong evidence that household income has causal effects on early child health; on children’s cognitive and schooling outcomes; on their social, emotional and behavioural development; and on intermediate outcomes which help explain the relationship between income and child outcomes, including maternal mental health and the physical and emotional home environment. Increases in income, including through changes in the benefit system, lead to positive improvements in child development, with a given amount of money having greater impact on households that have lower income to begin with.

In short, it is clear that children growing up in low income households do less well in part because they are poor, and while money is not the only factor explaining differences in outcomes for children from different socio-economic backgrounds, it is an important part of the story.

More recent work in CASE continues to explore the ways in which income changes affect outcomes for children. Cooper (2017) looked at the way that economic hardship affects parenting, showing that worsening levels of material deprivation are associated with a worsening of a number of parenting behaviours, including how close the mother feels to the child, play activities and discipline. Changes in experiences of hardship are also related to changes in mothers’ mental health, which appears to be an important mechanism explaining the link between poverty and parenting.

Reeves et al’s (2020) study of the benefit cap (part of an ongoing project on Welfare Reform and Larger Families) finds further evidence of the link between low income and mental health, showing significant decreases in adult mental health in households at risk of being capped when the benefit cap was lowered in November 2016, with worrying implications in turn for the children in those households. Reader’s (2021) examination of the shortlived universal Health in Pregnancy Grant (2009-2011) finds evidence that the policy led to improvements in average birthweight, particularly for young mothers and those living in deprived areas.

What links can be established for children between financial hardship, educational under-achievement, family breakdown and worklessness?

We know that financial hardship is linked to educational underachievement. The evidence cited in the previous answer supports this conclusion. In addition, work done by Reader at the Education Policy Institute shows how the widening attainment gap has been driven by worsening outcomes for the most persistently disadvantaged pupils – indicating that rising poverty in households with children is starting to show up in attainment data (Hutchinson et al, 2020).


Khazbak’s PhD research (forthcoming) investigates the ways families’ financial hardship can have a negative impact on adolescent’s emotional and mental wellbeing, sometimes translated into disruptive behaviour at school (possibly leading to exclusions), or being emotionally vulnerable, taking risks, etc. which can impact on educational attainment. School exclusions (especially permanent ones) can be detrimental for future outcomes including worklessness. Khazbak’s findings show that teenagers may carry a particularly heavy burden of families’ financial hardship, with young people often feeling a responsibility to ameliorate it by looking for ways to earn money themselves. Employment at this age does not have to be a negative outcome but if there are difficulties in finding work or the payment is not enough (which is often the case), this can push some towards seeking illegal ways of making money, eventually affecting their educational attainment.  

There is also evidence that financial hardship increases the risk of family breakdown, though this is not something we have looked at in CASE directly. And we would underline the links between financial hardship and other important outcomes not included in the question, including early child health outcomes and parental mental health, as discussed earlier in this submission.

Looking at causal relationships running the other way, we know that educational achievement is strongly protective from financial hardship in later life; just one of a number of reasons to prioritise children’s access to high quality education. We also know that family breakdown can result in higher rates of poverty – but this is very much dependent on policy decisions about the financial support available to families with children, including lone parent families, as well factors like the cost of childcare and protection of part-time and flexible working, which facilitate lone parent employment (e.g. Brady et al, 2017).

Finally, it is clear that risks of poverty are higher in households with no paid work. However, two points are important. First, it is increasingly the case in the UK that work in itself is not sufficient to protect a household from poverty: the risk of poverty is rising in working households. This is true even in households where all parents work full-time, as forthcoming work as part of the Nuffield-funded Welfare Reform and Larger Families project will show.

Second, the fact that households in which no one is in paid employment face a very high risk of poverty is entirely a political decision, and reflects decisions about the generosity of the social security system. At least prior to the pandemic recession, the vast majority of households with children where no adult worked had very young children, were lone parent households, and/or were affected by disability. Rates of poverty in these households came down substantially between 1997 and 2012, as social security benefits for families with children were increased in real terms, before rising sharply thereafter, as successive rounds of cuts took effect.

Joint working: What would be the merits of having a cross-government child poverty strategy? How well has this worked in the past?

A cross-government child poverty strategy seems to have proved very effective in the Labour years in ensuring the focus on child poverty was shared across government departments at national and local level. The establishment of the Department for Children, Schools and Families, in which the Child Poverty Unit was based, ensured that while child poverty was measured in terms of income and material deprivation, reducing poverty was not seen as solely the remit of the Department of Work and Pensions, as is the case today.

Nonetheless, there were significant gaps in attempts to work cross-government, and a future strategy should seek to do better. Notably, the Child Poverty Unit never considered how Home Office policies contributed to child poverty.

We would also emphasise the positive lessons from the focus on joined-up working at local level, which plausibly improved the effectiveness of government spending in an era of multiple initiatives, by reducing duplication. Service integration was an important part of the thinking behind Sure Start, and the Every Child Matters framework picked up this approach and developed it further, ensuring clearer lines of accountability for children’s welfare. While fully joined-up working was difficult to achieve in practice (see Eisenstadt, 2011), there is evidence that the development of Sure Start Children’s Centres led to wider reconfiguration of local services for young people and to the creation of effective multi-agency teams (House of Commons Children, Schools and Families Committee, 2010, cited in Stewart, 2013).


Brady, R Finnigan and S Hübgen (2017) ‘Rethinking the risks of poverty: A framework for analysing prevalences and penalties,’ American Journal of Sociology, 123 (3), 740-786.

Cooper, K (2017) Poverty and Parenting. LSE PhD Thesis. http://etheses.lse.ac.uk/3633/ 

Cooper, K and J Hills (2021) ‘The Conservative Governments’ Record on Social Security from May 2015 – pre-Covid 2020: Policies, Spending and Outcomes,’ SPDO Research Paper

Cooper, K and K Stewart (2013) Does money affect children’s outcomes? A systematic review, York: Joseph Rowntree Foundation.

Cooper, K and K Stewart (2020) ‘Does household income affect children’s outcomes? Evidence from a systematic review,’ Child Indicators Research.

Eisenstadt, E (2011) Providing a Sure Start: How government discovered early childhood. Bristol: The Policy Press.

Hills, J, T Sefton and K Stewart (eds) (2009) Towards a More Equal Society? Poverty, Inequality and Policy Since 1997 Bristol: The Policy Press.

Holder, H, T Tsang and P Vizard (2011) Developing the Children’s Measurement Framework: Selecting the Indicators. Equality and Human Rights Commission Research Report 76.

Hutchinson, J, M Reader and A Akhal (2020) Education in England – Annual Report 2020 London: Education Policy Institute. https://epi.org.uk/publications-and-research/education-in-england-annual-report-2020/ 

McKnight, A, P Mendes Loureiro, P Vizard, A Prats, A Claver, C Kumar, P Espinoza-Lucero, C Leavey and L Mompelat (2017) Multi-Dimensional Inequality Framework. LSE: Atlantic Fellows for Social and Economic Equity. https://sticerd.lse.ac.uk/inequality/the-framework/media/mif-framework.pdf

Obolenskaya, P, T Burchardt, I Shutes and P Vizard (forthcoming) Poverty and deprivation among children in recently-arrived migrant families in the UK, unpublished manuscript.

Pinter, I., Compton, S., Parhar, R. & Majid, H (2020) A Lifeline for All: Children and Families with No Recourse to Public Funds. London: The Children’s Society. Available at: https://www.childrenssociety.org.uk/information/professionals/resources/lifeline-for-all

Reader, M (2021) ‘The birthweight effects of universal child benefits in pregnancy: quasi-experimental evidence from England and Wales,’ CASEpaper 222. London: Centre for Analysis of Social Exclusion.

Reeves, A, M Fransham, K Stewart and R Patrick (2020) ‘Did the introduction of the benefit cap in Britain harm mental health? A natural experiment approach,’ CASEpaper 221. London: Centre for Analysis of Social Exclusion.

Stewart, K (2013) Labour’s Record on the Under Fives: Policy, Spending and Outcomes 1997-2010. Social Policy in a Cold Climate Working Paper 4, July 2013.

Stewart, K and M Reader (2021) ‘The Conservative Governments’ Record on Early Childhood from May 2015 – pre-Covid 2020: Policies, Spending and Outcomes,’ SPDO Research Paper

Stewart, K and N Roberts (2018) ‘Child poverty measurement in the UK: Assessing support for the downgrading of income-based poverty measures,’ Social Indicators Research, 142 (2), 523-542.

Townsend, P (1979) Poverty in the United Kingdom. London: Allen Lane and Penguin Books.

Vizard, P and J Hills (eds) (2021) The Conservative Governments’ Record on Social Policy from May 2015 – pre-Covid 2020: Policies, Spending and Outcomes. An assessment of social policies and social inequalities on the eve of the Covid-19 pandemic.

Vizard, P, T Burchardt, P Obolenskaya, I Shutes and M Battaglini (2018) Child poverty and multidimensional disadvantage: Tackling “data exclusion” and extending the evidence base on “missing” and “invisible” children Overview Report, CASEreport 114.


Waldfogel, J (2010) Britain’s War on Poverty. New York: Russell Sage Foundation.

Zaidi, A and T Burchardt (2005) ‘Comparing incomes when needs differ: Equivalization for the extra costs of disability in the UK,’ The Review of Income and Wealth, 51(1) 89-114.


March 21





[1] For example, in recent work we found that nearly half of children in recently arrived non-EEA migrant families were living in poverty (2013/14 – 2015/16) after taking housing costs into account, with a slightly lower proportion of children in recently arrived EEA migrant families. These rates are nearly twice as high as for children in UK-born/long-term resident families. Moreover, we also found that children in recently arrived migrant families who are (after housing costs) poor experience a greater depth of poverty, on average, compared to children in UK-born/long-term resident families (Obolenskaya et al, forthcoming; Vizard et al, 2018).

[2] No justification is given for the choice of one year, or for the simple split by 52. Asset-augmented income measures usually use more comprehensive annuitization methods.