A wider range of subjects?


Today Justine Greening is giving a speech about education and specifically the attainment of children from low and middle income families. The reaction to the speech will focus on the rights (not many) and wrongs (many) of grammar schools, but that should not wholly drown out some very welcome data work released by the Department for Education alongside the speech to help us better understand the relationship between education outcomes and family income.

By linking together data collected by DWP, HMRC and the Department for Education, the work seeks to set out a new educational attainment measure for children from ‘ordinary working families’ on modest incomes. This marks the latest step by the government to live up to their rhetoric to better support the just about managing. It represents a welcome move to assess future educational – and perhaps wider – outcomes, embedding a policy focus on children from households who are in work but with relatively low income.

A new measure that acts as a proxy for relative disadvantage, and sits alongside rather than replacing the traditional indicator based on Free School Meal entitlement, makes a lot of sense. Families are eligible for Free School Meals (FSMs) if they are entitled to one of the income-related out-of-work benefits (like Jobseekers Allowance or Income Support) or if they work but do not qualify for Working Tax Credit (which kicks in when someone is working at least 16 hours a week). But with the workless household count at a record low and two thirds of those in relative poverty living in working households it’s well worth building on the existing narrow measure to ensure we are monitoring the outcomes for those on low and middle incomes, alongside the most disadvantaged.

So how best to measure this group of children from households with income above the very lowest but less than the median? The Department for Education has attempted to build a definition that chimes closely with the Resolution Foundation’s own measure of Low to Middle Income households, covering those children in families below typical incomes but above the incomes of families on FSMs. In doing so they have the challenge of basing it on actual administrative data held within government of individual households and children rather than the representative surveys of the population that most work on family incomes rely upon.

That they have attempted to overcome this challenge is very welcome indeed despite the inevitable compromises when bringing together different data sources and producing something relevant for policy development purposes. As the chart below – taken from the DfE’s consultation document – shows, there appear to be some important differences in the outcomes of pupils within this group.


Here are our five key reflections on the suggested indicator:

  • A measure of household income – Sensibly the Department for Education propose to account for total income of all people in a child’s household. This is a better measure of the total resources they have available than individual earnings. They are also right to take account of household size (via a process called equivalisation) in considering how far those resources stretch. A household measure also avoids the problem seen in, for example, the Child Benefit means test where taking the income of only one parent may simplify administration but results in income inequity: a single earning family on £60,000 a year receives no child benefit, whereas dual earning parents each on £49,999 receive the full entitlement.
  • Ease of data collection – Its clearly not feasible for all families to provide detail of their household composition and income to schools every year to create this measure. Instead administrative data that the government already collects on income, through the benefit, tax credit and tax system, is being matched with data on pupil attainment. This approach may not create a perfect measure of total household income, given gaps in the information government collects and the difficulty of matching child records to those of all members of a household, but this is a reasonable trade-off to make the indicator implementable. Taking a step back the approach is an important step forward in making better use of administrative data and will hopefully promote more use and improvements in data quality. The matching process should be taken further to focus on the earnings of children after they leave education. That’s because important as qualifications are we do also want to know how parental income affects the living standards of children once they enter the world of work.
  • Stability of the target population – So far the government’s measure is a snapshot of the target population at a single point in time (the academic year 2015/16), but this group is not a stable one. Children will move in and out of the definition, perhaps because of a fall in income because parents separate or an income boost from that rare thing – a large pay rise. Alternatively, households may have remained at this income level for an extended period of time, Resolution Foundation analysis has shown that two-fifths of low to middle income households had the same status 15 years later. Clearly the life experience of the children in these examples will differ, and those differences can in turn impact on their educational attainment with different policy implications. Such fluctuations need to be taken into account and their impact understood, not least because shifts will probably also be related to the age of children. For example more households have second earners – which can push them out of the government’s definition – when children are in secondary than primary school.
  • Size of the target population: The downside of accounting for a fluctuating population, perhaps by flagging if a child is ever in the group or a member for a minimum period of time, is that it increases the size of the target population. The group above the Free School Meal threshold and below typical income already accounts for a third of all children on the government’s initial measure. Keeping tabs on those moving in and out of the group risks expanding that population further, risking the relevance of the indicator. Mitigation could come via some form of income threshold, determined by further investigation of historic outcomes at different levels of income, or introducing multiple measures. There is a simple trade-off here between simplicity and meaningfulness of the measure.
  • Longevity: Drawing lines to put families into different groups is inevitably an arbitrary business but ultimately where it lies should come down to ensuring the indicator both reflects the policy intent and, given the need to track outcomes over a lengthy period of years, can stand the test of time. That’s also why the suggested label for this group of ‘Ordinary Working Families’ presents a risk to the measures long term viability. Future governments will inevitably prefer a different label – we’ve already seen this government move from JAM to OWF, themselves an evolution of previous political tags. A more technical phrase would better suit dusty statistical annals, and may reduce the risk of a future government striking the definition from official records, leaving more flowery language to policy publications.

A broader measure of educational attainment for lower income households is a good starting point if the need for it to sit alongside existing measures of more severe disadvantage. The Department for Education deserves credit for putting in the work to make such a measure possible but a robust strategy to support the living standards of this group needs to go further. At the level of the children concerned that means tracking outcomes beyond education and into earnings, or even future household income.

But to understand what drives outcomes the government will need to understand how the level of income for this group compares to past periods and other parts of the income distribution. Assessing how the incomes of this group fare is vital to understand their living standards. And doing so will be ever more important in the remainder of this parliament with a new squeeze on pay underway and the coming £12 billion of cuts set to make half of working age households with less than typical income – the very focus of this new measure – worse off by the end of the parliament.