26th April 2018

EPI Statement on the Data Protection Bill’s immigration exemption

Risks to children and to educational research from the ‘immigration’ exemption at Schedule 2, Part 1 Paragraph 4


As set out below, we believe this exemption poses serious risks to the welfare of children and to the future of education research in England. Furthermore, we believe the use of education data for immigration enforcement purposes is likely to be a self-defeating exercise, which risks damaging the integrity of the data available to government for the development of evidence-based education policies without any realistic prospect of furthering the government’s aims to reduce net migration in the medium or long term.

Background: Use of the current (DPA 1998) ‘crime’ exemption for immigration enforcement

Under the current law, government departments can and do already use the exemption for data processing for the purpose of the ‘prevention or detection of crime’ and the ‘apprehension or prosecution of offenders’, in the DPA 1998’s Section 29, in order to exempt themselves from data protection rules when pursuing immigration enforcement policies. This means that, since 2015, the Department for Education (DfE) has passed datasets to the Home Office (HO) on a monthly basis containing potential address and school details for up to 1,500 children per month. In practice the numbers of records requested and supplied have been lower than this according to ‘data share transparency’ information published by the DfE, ranging from zero to 507 requests per month, resulting in zero to 94 educational records per month shared with the HO.[1]

The Memorandum of Understanding (MOU) between the DfE and the HO in respect of these data transfers describes the purpose of the HO’s processing of the data as ‘support[ing] the government’s commitment to reduce net migration’.[2] The MOU goes on to cite several ‘strategic aims’ for the data sharing, including ‘to protect the interest and safety of the child’ in cases where the HO has lost contact with children and their family members, but also the aim to ‘create a hostile environment for those who seek to benefit from the abuse of immigration control’. This latter aim has no focus on the interests of those children who are caught up in immigration enforcement through the processing of their school records.

Proposed ‘immigration’ exemption in the Data Protection Bill 2017

It is unclear why the government requires an exemption specifically focused on immigration control (in addition to the continuing exemption for the prevention or detection of crime) unless it intends to increase the volume of data processed for this purpose substantially, and/or to reduce the very limited transparency around its processing that currently exists as a result of a programme of Freedom of Information Act requests by the defenddigitalme group. A possible effect of the proposed exemption would be to enable the government to undertake large-scale processing of National Pupil Database (NPD) records for non-education and non-child-welfare purposes, to undertake this processing with a greater degree of secrecy, and to undertake it with no right of objection by parents, children, or schools.

Currently the HO processes school records of children whose parents or relatives it suspects of being in breach of immigration rules. There is no independent or external oversight of the decision to request a particular child’s data, and the only check on the proportionality of the data processing is the agreed limit of 1,500 records per month. It is not possible from the information released so far to establish on how many occasions the child records that have been shared with the HO related to the intended child, and on how many occasions they related to another child with similar personal details. Nor is it publicly known on how many occasions the data provided actually resulted in the effective enforcement of any law.

We do know, however, that fewer than a quarter of the 4,296 requests made by the HO between July 2015 and September 2017 resulted in an apparent match to a child record, which was then shared with the HO. With such a low match rate it could be tempting to process the data for increasing numbers of children, casting an ever-wider net based on increasingly limited evidence, or even no suspicion at all, in the hope of finding cases that could be pursued by immigration enforcement officials. We believe that the Committee should consider carefully the likely consequences of this scenario for children and for education research.

Risk of harm to children

There are various ways in which the use of educational data for immigration enforcement could cause harm to individual children. A very real risk is that parents, teachers and schools, once aware of this data processing, will take steps to avoid the involvement of children known to them in immigration proceedings that could destabilise their family and home life. This could take the form of boycotting aspects of statutory data collections, and/or supplying inaccurate or incomplete information about children to their school.[3],[4]

If the scale of the use of education data for immigration enforcement increases, an even worse scenario could arise, whereby increasing numbers of children are withdrawn from (or never access) school and other public services, in an attempt to avoid immigration sanctions. Migrant children are already more likely to be missing form education than other children.[5] The question of whether an individual has the right to remain in the country is not always straightforward and well-understood by those concerned, and the fear of removal could effectively deny education, healthcare and social integration to children who have not committed any crime, and who may not even be related to anyone who has committed an offence. In some cases, children who are not in education and do not access NHS services may be at heightened risk of abuse and/or neglect.[6]

In this sense, the NPD places those children who are accurately recorded within it on the official radar, and enables national and local government to undertake responsibility for their educational participation and outcomes more effectively, as well as their safety and wellbeing. This is only possible if parents feel that they can submit personal data to schools safely and without risks to their interests or those of their children.

Risk to education research capability

Incomplete or inaccurate education data, as a result of parental and professional fears about how those data will be used, are not solely a problem for migrant families, or those that fear they may be mistaken for migrant families. There is an additional risk to the long-term interests of all children in current and future school cohorts if the quality of the data in the NPD becomes degraded. As we found in our recent research report on the educational outcomes of children with English as an Additional Language (EAL), whenever data about pupils collected from schools is used for non-education purposes this raises ethical questions, and even the perception of data processing that is adverse to the child’s interests presents a threat to the future accuracy and completeness of education data.[7]

It is important not to underestimate the contribution to good government made by the integrity of data in the NPD, and not to imperil this for insufficient cause. To take EPI as one example of the approximately 130 organisations (mainly engaged in education and health research and analysis) that accessed NPD data during the spring of 2017, last year we published seventeen reports, twenty policy analyses and four government consultation responses. All of these relied on accurate and complete educational data to some degree, and most would have been entirely impossible without access to the high-quality data recorded in the NPD. We used this research to give evidence or contribute expertise in Parliament five times and it was cited by others in Parliament thirty-two times. Furthermore, 2017 was just one of fifteen years for which NPD data have been available.

In addition to external research organisations, the government itself relies on NPD data to allocate funding to schools, to forecast how many teachers need to be trained, and to hold itself and schools to account for the educational progress made by children. These are just three examples of many, illustrating how vital the integrity of the NPD data is to government itself. It is worth noting that the accuracy of pupil data is of most importance to groups of children who are educationally vulnerable – be those children with EAL, with Special Educational Needs and Disabilities, those eligible for free school meals due to economic deprivation, or those in care or on the edges of the care system. These are the children for whom accurate statistics are most vital to promoting improved outcomes. It is important that positive steps the government has taken to improve its understanding of children’s needs through the NPD, such as the collection of data on the English proficiency of children whose first language is not English, continue to be available to support children’s learning and provide insight into the effectiveness of education.

Envisaged policy goals of the proposed exemption unlikely to be furthered

The final point which we wish to draw to your attention is the likely futility, over the course of time, of attempting to use educational data to further attempts at immigration control. It isn’t credible to expect parents and school staff to continue to provide personal data that could place children’s interests at risk in the context of this type of data processing. Various campaigns protesting against this use of data are already underway, and have received media coverage. The more that government extends its processing of school data for these purposes, or seeks to give itself powers to do so in greater secrecy, the greater the danger to the completeness and integrity of the data will become. That not only means that the data would lose their value to the DfE and to education researchers who seek to enhance children’s outcomes, it means they would rapidly become useless for immigration control purposes too. This self-defeating nature of the purposes for which the immigration exemption is designed means that beginning with children and their families, then later the DfE, and eventually the HO, nobody would gain from this attempt to alter the purposes for which the data are collected.

[1] Extracted from: https://www.gov.uk/government/publications/dfe-external-data-shares

[2] Released on 22.02.2017 via Freedom of Information Act Request: https://www.whatdotheyknow.com/request/pupil_data_off_register_back_off

[3] https://www.thetimes.co.uk/article/boycott-school-data-trawl-parents-urged-30p9j06vj

[4] https://www.theguardian.com/uk-news/2016/sep/26/parents-boycott-requests-childrens-country-of-birth-information

[5] https://www.childrenscommissioner.gov.uk/wp-content/uploads/2017/11/BRIEFING-Falling-through-the-gaps-in-education-CCO.pdf

[6] https://www.ncb.org.uk/sites/default/files/field/attachment/Not%20Present%20What%20Future.pdf

[7] https://epi.org.uk/publications-and-research/educational-outcomes-children-english-additional-language/