Written evidence from the Cabinet Office (GOS 14)
Public Administration and Constitutional Affairs Committee
Governance of Statistics inquiry
The Statistics and Registration Service Act 2007 received Royal Assent on 26 July 2007 and the majority of its provisions came into force on 1 April 2008. This included the creation of the UK Statistics Authority (UKSA), as an independent body with a statutory objective to promote and safeguard the production and publication of official statistics that serve the public good.
Residual ministerial responsibilities in respect of UKSA fall to the Cabinet Office under the provisions of the Statistics and Registration Service Act 2007.
Marking a decade since the establishment of the UK Statistics Authority, it is timely to consider the work and achievements of the Authority over the last ten years.
The implementation of the Statistics and Registration Service Act 2007 (relates to Q1 and Q3 of the inquiry terms of reference)
The Statistics and Registration Service Act 2007 brought a number of reforms to the previous statistical system. This included the creation of the ‘Statistics Board’ and its executive office, and setting in statute the responsibilities and functions of the Board. The UK Statistics Authority was established as a non-ministerial department in 2008, with the Office for National Statistics (ONS) as its executive office. The National Statistician is both the Chief Executive of the UKSA and its principal adviser on the quality of official statistics, good practice in relation to official statistics, and the comprehensiveness of official statistics. The National Statistician is also head of the cross-departmental Government Statistical Service (GSS), although this function is not defined in the legislation.
Post-legislative scrutiny of the Statistics and Registration Service Act 2007 by the Cabinet Office in 2011[1] considered how far the legislation had achieved its agreed aims and objectives in terms of six underlying principles which had guided the reform of the statistical system:
Stakeholder consultation informing the 2011 post-legislative scrutiny revealed high levels of agreement that the reforms to the statistical system have strengthened independence; that official statistics are of high quality and integrity; and that there is clarity on roles and responsibilities.
There was, however, considerably less agreement that the new system was sufficiently flexible to respond to changing needs, and whether it was efficient and provided value for money. In particular, the requirements of the, then, Code of Practice were considered to hamper the ability of statistical producers to be responsive to changing needs, and the assessment process was considered resource intensive. There was also a view that there was a focus on compliance with the Code which did not sufficiently recognise existing good practice.
The refreshed Code of Practice, published in 2018, and the current, more principles-based approach to compliance with the Code, has responded to stakeholder views. The new Code promotes the wider values of statistical professionalism, and we are aware of a number of organisations[2], who are not designated producers of official statistics, who have opted to voluntarily comply with the Code of Practice. The establishment of the GSS Best Practice Team[3] achieves a balance between the careful regulation of statistics by the Office for Statistics Regulation (OSR) and the provision of support to departments to encourage and foster new approaches to producing and disseminating statistics.
Additional issues highlighted in the 2011 post-legislative scrutiny consultation have also been subject to subsequent reforms. In particular the perceived conflict between the Authority’s responsibility for the production of ONS statistics and their regulation, which was considered in Sir Charles Bean’s Independent Review of UK Economic Statistics[4]. The Bean Review recommended that the Authority’s regulatory function should be strengthened, and considered the arguments for and against it being situated within the UKSA. In response, the Office for Statistics Regulation (OSR) was launched in 2016, providing greater clarity on the respective roles of the regulator and executive functions of UKSA. The OSR reports publicly on its activities, as recommended by the Bean Review. These reports can be found alongside the business plans, strategies and reports on the wider work of the Authority[5].
Low levels of public trust in official statistics were a primary driver for statistical reform.
Research into public confidence in official statistics in 2009[6] asked the extent to which respondents agreed or disagreed that ‘official statistics are produced without political interference’ to which a majority (59%) disagreed (17% agreed, 23% neither agreed or disagreed, or did not express an opinion).
More recent research, conducted in 2016[7], showed:
Statistical innovation and development (relates to Q2, Q7, Q8, Q9 and Q10 of the inquiry terms of reference)
Sir Charles Bean’s independent review of UK economic statistics set out a number of strategic recommendations to respond to both current and future challenges associated with measuring the modern economy. This included the importance of unlocking the value of administrative data for more timely and accurate economic statistics, through improved access to data and to the capabilities needed - both technological and analytical - to exploit the value of data.
The amount and range of data now available allows statistics and analytics to exploit data in new ways to inform policies and decisions. To help achieve the Bean Review recommendations Government provided additional investment in the 2016 Budget[8] to establish an ONS data science hub and centre of excellence in economic measurement. The ONS Data Science Campus[9] was launched in 2017, with the aim to investigate the use of new data sources for public good and to help build data science capability. The Economic Statistics Centre of Excellence[10] is an academic partnership focused on analysis of emerging and future issues in measuring the modern economy. Achieving the right balance between traditional data collections and more novel sources of data will be important for informed policy-making.
UKSA and ONS are working with the Government Digital Service (GDS) to increase access to novel technologies, software and analytical tools, in a way which is both secure, and scalable. Collaborative work in this space includes the Reproducible Analytical Pipelines project[11] which can increase efficiency in the statistical production process, and the Data Science Accelerator programme[12] which aims to build data science skills among analysts working in the public sector.
Provisions in the Digital Economy Act 2017 have been designed to support the greater reuse of administrative data held by government and improve UKSA’s ability to access data for the purpose of producing official and national statistics. Part 5 of the Digital Economy Act 2017 introduced a number of new powers to share information to help make the digital delivery of public services more efficient and effective. The UKSA is responsible for the powers under chapters 5 and 7 of Part 5 of the Digital Economy Act 2017 which deal with research and statistics respectively. The powers covered by this legislation have only been fully operational since the accompanying codes of practice were published in July 2018. The early signs in the six months to date have been largely positive but more time will be needed to realise the full benefits that the legislation can provide in terms of removing barriers and enabling the effective exploitation of administrative data.
The Digital Economy Act provisions for data sharing for statistical purposes apply to the ONS and its functions in the Statistics and Registration Service Act 2007. Data held by health and adult social care bodies is exempted from other provisions within Part 5 of the Act in response to public concerns about confidentiality, and the need to protect the privacy of health records.
The UK Statistics Authority has established a Research Accreditation Panel to oversee the governance and accreditation of processors, researchers and projects wishing to use the research powers in chapter 5 of the Digital Economy Act 2017. The panel has an independent chair and members, alongside representatives of government departments, the devolved administrations and UK Research and Innovation, and will meet monthly to accommodate the anticipated demand. ONS has further developed its technical systems to support scaled-up access to data for the research community and has expanded its secure researcher training programme.
Since the statistics powers (chapter 7) came into force, amending the Statistics and Registration Service Act 2007, ONS has continued to work with departments across Government, and with other data providers, to identify where data sharing agreements can be put in place to give ONS access to data sources to support better statistics and statistical research for better decision making for the public good.
In parallel with these enhanced powers of access to data, ONS is transforming its capability to make the best use of administrative data. The Census and Data Collection Transformation Programme is moving ONS data collection towards greater integration of administrative and other non-survey data sources, rationalising survey data collections, and moving survey data collection online. The 2021 Census will, for the first time, be primarily online so people can take part when and how they want. The option to complete a paper questionnaire will still be available.
UKSA and ONS engage well with the parts of government responsible for data policy including data ethics in the public sector. John Pullinger, as National Statistician, sits on the cross government Data Advisory Board which has the role of setting strategic direction for the better use of data within the public sector. UKSA are actively involved in the various work streams reporting to the Board, particularly on leading work to build capability. We expect the newly created Centre of Data Ethics and Innovation to make use of the lessons already learned on data ethics, for example, through the UKSA’s National Statistician’s Data Advisory Committee.
The terms of reference for this inquiry ask whether the UKSA, following the shortcomings in economic statistics highlighted in the Bean Review, is doing enough to tackle problems with other areas of official statistics. This is a core aim of the OSR systemic review programme. The OSR report ‘Joining up data for better statistics’[13] highlights the insight and value for research and decision-making from sharing and linking public sector datasets which is not being fully realised. This report recognises that robust safeguards for protecting privacy are essential and sets out a vision for a safe and effective data linkage system to better exploit this value across the official statistics system. Through their health and social care systemic review[14] OSR raised a number of challenges around coherence and presentation of health and care data and the need for deeper analytical insight in public presentation of the data, and leading effective strategic engagement to improving and developing systems.
Building and maintaining cross-government analytical capability (relates to Q4, Q5 and Q11 of the inquiry terms of reference)
It can be considered a strength of our statistical system that professional statisticians, with a strong professional ethos but also deep subject matter expertise, work within ministerial departments to help inform policy and operational decision-making. The proper use of statistics flows from a combination of technical experts who are able to communicate statistical evidence effectively, and in a way which helps inform decision-making, and decision-makers who are, in turn, able to act as intelligent customers for statistical evidence. Close engagement between statisticians and policy-makers supports this and Statistical Heads of Profession have an important role in ensuring departmental statistical needs are met, and in safeguarding the professional integrity of official statistics within their department[15].
The Analysis Function, led by John Pullinger, is a collaboration between the analytical professions. The vision for the Analysis Function, set out in the Analysis Function Strategy[16], is for analysis to be integral to policy and operational decision-making. Enhancing capability across the function and more widely across the Civil Service is a core strand of this work. This will be delivered through setting and embedding analytical standards across government; building analytical and critical thinking capability across the Civil Service, and; owning a consistent cross-departmental career framework, to drive attraction, develop members, and increase talent and diversity.
The Learning Academy, based at the Office for National Statistics, offers a broad statistics and data science learning curriculum for GSS statisticians. This includes MSc level qualifications - delivered through partner Universities - for those who wish to pursue a formal qualification designed to meet the needs of statisticians and data analysts working in government roles.
GSS statisticians also help deliver courses offered more widely across the Civil Service, with a particular focus on helping to equip policy-makers to be effective intelligent customers for analysis and evidence. This includes the Fast Stream Policy Basecamp, a core component of the Policy Profession Fast Stream curriculum, which includes a session run by senior experts from the GSS on the uses, and misuses, of statistics in policy-making.
The Fast Stream[17] is the government’s flagship graduate programme to support and accelerate the development of talented people to become future leaders of the Civil Service. It is an investment in talent, intended to ensure that present and future governments are supported by an efficient and effective Civil Service with the right skill set and the versatility to respond to changing priorities.
The Statistician Fast Stream is centrally managed by Civil Service HR, within the Cabinet Office, with the Government statistics profession having a leading role in setting specialist recruitment and assessment standards, and for the specialist learning and development for those on the scheme.
Statistics Fast Stream learning and development opportunities include:
Independent, high quality statistics are essential for democratic debate, decision-making and accountability. Statistics help to formulate and evaluate policy; inform operational decisions, including managing the provision of services; encourage and inform public debate; and help hold Governments and decision-makers to account. Statistics must therefore be - and be seen to be - of the highest professional standards of quality and integrity. The professionalism of the cadre of statisticians working across the GSS, supported by the Code of Practice and structures established within the Statistics and Registration Service Act 2007, has helped embed these values.
1
[1]https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/62065/PublicAdministration_PostLegislativeScrutiny_acc.pdf
[2] https://www.statisticsauthority.gov.uk/code-of-practice/voluntary-application-of-the-code/list-of-voluntary-application/
[3] https://gss.civilservice.gov.uk/about-us/support-for-the-gss/
[4] https://www.gov.uk/government/publications/independent-review-of-uk-economic-statistics-final-report
[5] https://www.statisticsauthority.gov.uk/about-the-authority/strategy-and-business-plan/
[6] http://natcen.ac.uk/media/25384/public-confidence-official-statistics-2009.pdf
[7] http://natcen.ac.uk/our-research/research/public-confidence-in-official-statistics/
[8] https://www.gov.uk/government/publications/budget-2016-documents/budget-2016
[9] https://datasciencecampus.ons.gov.uk/
[10] https://www.escoe.ac.uk/about-escoe/
[11] https://dataingovernment.blog.gov.uk/2017/11/27/transforming-the-process-of-producing-official-statistics/
[12] https://www.gov.uk/government/publications/data-science-accelerator-programme/introduction-to-the-data-science-accelerator
[13] https://www.statisticsauthority.gov.uk/publication/joining-up-data/
[14]https://www.statisticsauthority.gov.uk/wp-content/uploads/2018/03/March-2018-Phase-1-Systemic-review-health-and-care-summary.pdf
[15]https://www.statisticsauthority.gov.uk/wp-content/uploads/2018/02/Roles-and-Responsibilities-of-Heads-of-Profession.pdf
[16]https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/742625/P19650_Analysis_Function_Document_landscape_HI_RES.pdf