Guidance

Independent Review of the UK Statistics Authority by Professor Denise Lievesley CBE (HTML)

Updated 12 March 2024

Lead Reviewer Foreword

1. I am delighted to present the findings and recommendations resulting from my Review of the UK Statistics Authority (UKSA). The Review has been carried out within the framework provided by the Cabinet Office Public Bodies Review Programme and this has provided a helpful structure to my work and the resultant findings.

2. The UKSA sits at the helm of a complex statistical structure which incorporates multiple producers of statistics, ranging in size from the Office for National Statistics (ONS) with over five thousand staff ^{[footnote 1]} (including field force) through to units of only a handful of statisticians based in government departments and arm’s length bodies. But importantly the UKSA also includes the Office for Statistics Regulation (OSR), the regulatory body for statistics. This juxtaposition of production and regulation is unusual and is a large focus of the review. (The structure of the organisation is detailed in Annex A.)

3. The quality of official statistics matters to us all because decisions and research both inside government and beyond (in businesses, academia, civic society etc.) are dependent upon relevant, timely and impartial data. Many of these data cannot be collected except by official agencies (under the auspices of the UKSA) due to, for example, restrictions on access to administrative data and the need for mandatory compliance with some data collection. This brings an additional responsibility upon official statisticians to ensure that they are meeting the data requirements of a large range of organisations and individuals. 

4. I am honoured to lead this Review because of my long held belief in the critical importance of official statistics which are fundamental for evidence-based policy and decision making. Statistics we can trust are essential for a healthy society as they help to ensure well-informed decisions by putting the best available evidence at the heart of policy development and implementation. They also enlighten through making explicit what is known through scientific evidence and, importantly, what is not known. 

5. It is critical to appreciate that official statistics have an even broader role: they also serve to empower, enabling citizens to call governments to account and providing a window on society. As such they are an indispensable part of a democratic society.

6. The UN Fundamental Principles of National Official Statistics describes this context well: “Official statistics provide an indispensable element in the information system of a democratic society, serving the Government, the economy and the public with data about the economic, demographic, social and environmental situation^{[footnote 2]}. To this end, official statistics that meet the test of practical utility are to be compiled and made available on an impartial basis by official statistical agencies to honour citizens’ entitlement to public information.”^{[footnote 3]}

7. The Review has examined the structure of the statistical system, in order to see whether it facilitates or inhibits the production of quality data. It was important not to duplicate the more detailed reviews and consultations which are taking place, for example in relation to the Census and the Code of Practice for Statistics, and so this Review attempts to complement where possible the ongoing evaluations conducted by the OSR. Given the extensive number of statistical releases published by the ONS each year^{[footnote 4]}, each with their own purpose and users, this Review has not sought to assess the quality of individual publications but rather to contribute to the framework by which outputs can be assessed. Having said this, I do wish to acknowledge the detailed comments on individual statistical programmes provided to the Review by a number of users – these have provided useful context and will be fed back to the UKSA. 

8. I write at a time of potential change for statistical production and analysis through the development of technologies such as Artificial Intelligence (AI). Such developments are fuelled by the exponential growth in computer power thus facilitating access to massive amounts of data in a more timely way than hitherto. It will be essential for official statisticians to work in collaboration with digital, data and technology experts  to employ sound methodologies ensuring the validity of conclusions which are robust and reproducible. An ethical and secure framework will be needed in order to manage the risks associated with the use of AI. 

9. This Review is examining a truly complex environment, as even some of the most experienced users of the statistical system told me. Thus I have attempted to throw light on the parts that are particularly controversial, or unusual, or where I have heard misapprehensions about the UK’s approach. In doing so I have been mindful of the work of others; in particular, the 2016 independent review of economic statistics by Professor Sir Charles Bean^{[footnote 5]} and the work of the Public Administration and Constitutional Affairs Select Committee (PACAC) examining the UKSA, which I have found most helpful. 

10. In the time I have had, there are many areas where I have only just touched the surface of the issues being addressed by the UKSA. Sometimes this is because the UKSA is operating in an environment where they only have partial control – examples are the division of statistical authority between the four administrations of the Union; the career development, terms and conditions for specialist civil servants; the need for data sharing across departments; the funding arrangements for official statistics – and in these cases I have tried to provide some commentary even though recommendations might not be appropriate.  

11. It is a sign of the importance of the UKSA that so many users of statistics, other providers of data, data intermediaries, and experts in the organisation of the public service gave up their time to write and/or talk to me and my team (listed at Annex D). I hope they appreciate that it is impossible to write a review which reflects the wide range of views expressed, but perhaps this Review will be the starting point for some fruitful discussions. Of particular note too are the members of the Challenge Panel (listed in Annex B) who contributed enormously by making me look at my recommendations through different lenses. I am also especially grateful to senior staff in many other national statistical offices for the frank exchange of views on the challenges facing statistical offices and for their willingness to share ideas for improvements with me. The international standing of the UKSA is remarkably high and was enhanced by its boldness in the face of the need for just-in-time data during the Covid-19 Pandemic. 

12. It has been a great pleasure to spend time with non-executive and executive members of the UKSA, ONS, OSR and the wider Governmental Statistical Service (GSS). They demonstrated commendable professionalism to contribute to UKSA’s strategic objectives. In particular, they valued the opportunities to work on topics that are relevant to peoples’ lives, policy making and current affairs. This enthusiasm and engagement to help statistics users better understand the world in which they live is central to the organisation’s work and will continue to play an important part in ensuring that the UKSA continues to rise to the opportunities and challenges facing a modern Statistics Office. I am very grateful for their openness and contribution to the review. My thanks must be extended to Sir Robert Chote, Professor Sir Ian Diamond and Ed Humpherson. 

13. Finally, I have been helped in this endeavour throughout by the commitment and energy of my Review team (members listed in Annex D). I am immensely grateful to them for their support and advice to me throughout the Review, and for their willingness to engage constructively on so many aspects of an official statistics agency’s role.

Professor Denise Lievesley CBE

Lead Reviewer

January 2024

Executive Summary

14. The findings of this Review are set out under the four quadrants of the Public Sector Bodies Review guidance, Governance, Accountability, Efficacy and Efficiency.  

Governance

15. The Governance section considers the formal constitutional arrangements that underpin the establishment of the Authority in statute, and how this works in practice. The Review also explores themes including how the statistical agenda is set, how the UKSA Board operates, arrangements for financial and risk management, the ONS executive, the relationship between the ONS and OSR, the relationship with the Devolved Administrations and the international position of the UKSA. 

16. The central conclusion is that it is time for the Board to move into a more visible, ambitious space, primarily through establishing a Triennial Statistical Assembly which will consult widely with statistics users and producers to understand the range of views regarding the priorities and data needs for the UK. This will lead to a more transparent and robust setting of the statistical agenda with a greater emphasis on user needs.  

17. While the housing of both the production (ONS) and regulation (OSR) under the UKSA Board is unusual and is the source of concern amongst some in the user community, the Review ultimately concludes that the current set up is fit for purpose. However the UKSA, ONS and OSR should take additional steps in order to better clarify how the organisations work together in practice and how the OSR holds ONS to account. The Statistics and Registration Service Act 2007 should be updated to better reflect the separation of powers that exist in practice.

18. In July 2017 pre-release access (the practice of sharing statistics and their commentary pre-publication, most commonly to Government Ministers and their advisers) to ONS statistics was removed in all but exceptional circumstances. The Review was surprised to learn that this is not the case with other departments, including across the Devolved Administrations. This is out of step with international good practice and pre-release access should be removed across all government departments.  

19. The UKSA has responsibility for the statistical system across the four administrations of the UK. All stakeholders agree that the current system is complex which can be time consuming and cause points of friction. This is most acute, but not limited to, tensions where administrative data is not comparable between nations (e.g. on health). Additionally, significant resource disparities exist between the ONS and the statistical teams in Devolved Administrations which impacts on UK wide data collection. A change in approach is required, led by UKSA at a strategic level working with the Devolved Administrations and other key stakeholders such as the Department for Levelling Up, Housing and Communities. 

20. Given the above findings, the recommendations on governance are: 

21. Recommendation 1: The UK Statistics Authority should lead the establishment and delivery of a Triennial Statistical Assembly. This Assembly should involve key organisations inside and outside Government and across the four Nations, with the remit of determining the UK’s needs for statistics through a wide consultative process. This should include the private sector, government departments, local government, academia, think tanks and media representatives. 

22. The UKSA will then respond to this by producing a proposal for the statistical priorities for the next three years, thus identifying data gaps and ensuring that users can hold the statistical system to account on the delivery of the programme of work. It will also enable other producers of statistics to complement the work of the official statistical system and factor this work into annual budget allocation processes. 

23. To respond to the Statistical Assembly and to supplement its findings, an annual public lecture from the Chair of the UKSA should be delivered to provide an update on the work of the statistical system and priorities for the year ahead. This should build on the OSR’s annual State of the Statistical System report. The lecture would raise the profile of the Board, further user engagement and establish the UKSA’s leadership role in the statistical space. 

24. Recommendation 2: The Review recommends that the expertise of the senior staff of the ONS would be greatly enhanced by the appointment of a Director General for Methodology who would be a focal point for the improvement and communication of data quality, and who would foster engagement with senior methodologists in other national statistical offices and in academia.

25. Recommendation 3: The Government should amend the statistical legislation so that the Act reflects current practice, taking the opportunity to make clearer the practical operation of the UKSA. For example, that the OSR reports separately to UKSA, not via the National Statistician, and that the Director General of the OSR is an Accounting Officer and is expected to report separately to the Public Administration and Constitutional Affairs Committee (PACAC).

26. Recommendation 4: The inconsistent application of Pre-Release Access to official statistics across the UK Statistical system has the potential to undermine trust. The Cabinet Office and devolved legislatures should amend the relevant secondary legislation for each nation at the earliest opportunity to follow the approach to Pre-Release Access taken by the ONS in line with the Code of Practice for Statistics.

27. Recommendation 5: The UKSA should build on existing work and lead discussions between the four nations and strengthen the Concordat of Statistics to encourage more UK wide data by creating common standards and improving harmonisation where appropriate and mutually agreed. HM Treasury should ensure that funding is available to support the harmonisation of key data.

Efficacy

28. The efficacy chapter focuses on the outcomes for users of the UKSA’s work by exploring data sharing across government, the role of the National Statistician, innovation and legacy technology, data ethics, the Integrated Data Service (IDS), Government Statistical Service (GSS), international engagement, communication of statistics, user engagement, the ONS website and regulating the use of statistics. 

29. The Review concludes that the UKSA’s efficacy is hampered by the systemic and cultural barriers to responsible data sharing between government departments – this is most keenly apparent, but not limited to, the development of the Integrated Data Service (IDS). It is not for the UKSA to resolve this issue alone. Indeed the benefits of data sharing extend beyond statistical work and will result in more efficient delivery of public services. The responsibility therefore must lie with the Cabinet Office, with the support of HM Treasury, to bring departments together to remove the barriers to data sharing and to hold departments to account when they are not sharing data as required by law. The medium and long term cost to government of not establishing a standard, cross government approach needs to be better understood and owned at the most senior levels of government.   

30. While broader data sharing barriers are slowing the progress of IDS, many of those consulted by the Review shared concerns about the programme itself. The ONS must listen to these concerns and take urgent steps to ensure the successful delivery of the IDS, which is a vital programme for government. 

31. The UK has been lucky to have had a succession of talented and passionate National Statisticians, and the incumbent continues this trend. However, the role is extremely broad, requiring both strong statistical expertise together with the people skills and business acumen to lead a large organisation. This undermines the resilience of the organisation each time a new leader is sought and the Cabinet Office, working with UKSA, should review the role description well ahead of the next campaign and examine how it is nurturing the talent pipeline within government.

32. UKSA’s role in safeguarding and promoting the production and publication of official statistics is central to all that it does. While there are many excellent examples of the effective communication of official statistics, some users felt that the ONS’s overall approach to communicating statistics is not as effective as it could be and that it can lack a coherent approach. The website was cited repeatedly as not meeting user needs. Additionally, high profile events such as the Gross Domestic Product (Blue Book) revisions in September 2023 could have been less surprising for users if there was greater understanding of the limitations of data amongst key stakeholders. UKSA may wish to work with expert partners to improve their messaging, and ensure that the Board membership includes a communications expert.

33. UKSA expends considerable effort to meet the needs of the wide range of users of statistics, both within central government and beyond. The success of this engagement varies across theme or topic areas. To truly deliver the UKSA strategy Statistics for the Public Good (2020-2025), it is important that genuine user engagement is embedded across all theme areas and that the obligation in the Code of Practice to consult users before making changes to statistics should be reinstated. Additionally, the informal, sometimes opaque, process of appointments to the National Statistician’s Advisory Committees has caused concern amongst some users and thus this process should be more formalised and transparent to increase trust and accountability with users.

34. Many of those interviewed by the Review shared their deep concern about the rise of misinformation and the misuse of statistics in the public arena. The UKSA and the OSR play a critical role in monitoring the use of statistics in public debate and intervening where necessary. As the current and former Chairs of the UKSA and PACAC have noted, there is scope to strengthen the Ministerial Code to mandate adherence to the UKSA Code of Practice for Statistics. This Review concurs.

35. Following the UK’s departure from the European Union (EU), the UK is no longer a member of Eurostat, the European Statistical Office. The departure from Eurostat reduces the influence which the UK can have on the development of methodology and standards for statistics across Europe, and also means that the UK has to replace the benefits of cross-national cooperation through working with other agencies and through bilateral cooperation. Steps should therefore be taken to strengthen and clarify the UKSA’s approach to international engagement.

36. Falling response rates to surveys are having a detrimental impact on the quality of outputs – most notably on the Labour Force Survey. One potential mitigation is to mandate responses, as some other countries do such as Australia. The possibility of mandating responses for some key surveys should be explored in the UK. 

37. Given the above findings, the recommendations to improve UKSA’s efficacy are:

38. Recommendation 6:  The centre of government, led by Cabinet Office, HM Treasury and No. 10 must actively work to resolve the systemic, often cultural, barriers to data sharing between departments. All government departments, particularly those who own significant amounts of data, must prioritise data sharing for statistics and research purposes and support the development of programmes such as the Integrated Data Service to enable greater sharing of data across government for statistical and research purposes.   

39. Recommendation 7: The Integrated Data Service (IDS) is critical in facilitating the greater use of administrative data and bringing greater efficiencies to statistical analysis and decision making across government and academia. The Review therefore recommends that the ONS takes action to ensure that the purpose, scope and requirements of the IDS are clearly communicated and that the needs and concerns of departmental data owners are sufficiently understood. 

40. Recommendation 8:  Regarding the National Statistician role, Cabinet Office, working with the UKSA Board should: 

• Commission a review of the role of the National Statistician well ahead of the next recruitment campaign, examining the many component parts of the National Statistician role in order to decide whether to propose changes to the role and what this may look like. This should also identify where changes to the Act may be required to facilitate the delegation or sharing of the National Statistician’s responsibilities; and
• Examine the talent pipeline and talent development structures it has in place across the GSS to ensure that those with potential to be future applicants for the National Statistician post and other senior roles in the statistical system are identified and nurtured. 

41. Recommendation 9: Within ONS a suite of actions relating to communications should be adopted, including:

• Urgently improving the website so that it meets user requirements more effectively;
• Ensuring that there is a better understanding of the levels of uncertainty around specific official statistics, particularly economic, to reduce public (and government) surprise to revisions; and 
• Building partnerships with organisations that foster relevant communication expertise to improve engagement with the wider needs of users.

42. Recommendation 10:  The Authority Board should look to appoint a Non-Executive Director with relevant communications experience to advise and support the UKSA.

43. Recommendation 11: The Advisory Groups working with the National Statistician should become more formal: recruitment should be open and be clearly advertised to encourage applications. The style, design and attendance at the meetings should also be reviewed to ensure that they facilitate frank constructive sharing of views and feedback. 

44. Recommendation 12: To demonstrate the commitment to user engagement and remind producers of its importance, the reference to ‘consult users before making changes that affect statistics or publications’ in the UKSA Code of Practice should be reinstated so that users are consulted before producers make substantial changes to statistical data collection or outputs. 

45. Recommendation 13: Internationally, the UKSA should:

• In consultation with His Majesty’s Government, prioritise the establishment and signing of a Memorandum of Understanding with Eurostat in line with the option provided in the Trade and Cooperation Agreement; and 
• Update its international strategy, Statistics for the Global Good, to provide more detail on how it will engage and lead within key global organisations, including the Organisation for Economic Co-operation and Development (OECD) and International Monetary Fund (IMF).  

46. Recommendation 14: The UKSA should engage with the Cabinet Office to explore the consequences of mandatory completion of the Labour Force Survey.

Accountability

47. The accountability chapter explores whether ministers have enough assurance to meet their duties, how accountable and effective the UKSA relationship is to Parliament and the appropriateness of the relationship between the UKSA and the Cabinet Office. The Review focuses on the UKSA’s relationship with the UK Parliament, Cabinet Office sponsorship, the strategy and business plan and functional standards. 

48. As a Non-Ministerial Department the UKSA is accountable to the Parliaments of the UK, Scotland and Wales, and the Northern Ireland Assembly. In the UK Parliament, this function is led by PACAC. The Review found this relationship to be working well, though PACAC could engage further on matters of regulation (i.e. OSR). UKSA should work where appropriate with other Select Committees and continue to champion the use of ONS’s data and statistics across both Houses. In the devolved legislatures, where the engagement is less mature and more ad hoc, further engagement is encouraged where appropriate. 

49. Within government the UKSA is sponsored by the Cabinet Office, under the auspices of Baroness Neville Rolfe, Minister of State. The consensus from those working in both the UKSA and Cabinet Office is that the current relationship is effective, though the sponsor team is under-resourced. The current Memorandum of Understanding (otherwise known as a Framework Document), which underpins the sponsoring relationship as set out in Managing Public Money, is now almost four years old so in line with good practice should be updated. 

50. In light of the above, the recommendations on accountability are:  

51. Recommendation 15: The UKSA should build on its work engaging with Parliamentarians and Select Committees of both Houses and devolved legislatures further. It should continue to seek out opportunities to proactively add insights and value to the work and interests of the Select Committees. Equally, Select Committees should actively seek to use ONS data and analysis. 

52. Recommendation 16: The Cabinet Office should look to supplement the existing sponsor team with resource to give the support and resilience required. 

53. Recommendation 17: The Memorandum of Understanding between the UKSA and Cabinet Office as Sponsor Body should be reviewed by both organisations to ensure that it reflects the requirements in the HM Treasury Framework Document Guidance for Arm’s Length Bodies mindful of the UKSA’s statutory independence. 

Efficiency

54. The final chapter of the Review discusses the theme of Efficiency. As detailed in the terms of reference, the UKSA has already committed with HM Treasury to achieving efficiencies over the current spending review period and so consideration of further efficiencies is outside the scope of this Review. However, the Review has examined evidence regarding talent planning and retention and adherence to existing financial guidelines. 

55. In the Spending Review 2021 settlement, UKSA agreed with HM Treasury to deliver cash releasing efficiencies worth 10 percent of spending on ongoing baseline activities (circa £21m). In November 2022 the UKSA agreed to further efficiencies in response to the Chancellor’s Efficiency and Savings Review. As a result, this Review has not sought to identify additional savings.  

56. Beyond HM Treasury, the ONS has for some time delivered outputs partially or wholly funded by other government departments who have sought out their capability and expertise. Within the Prioritisation Framework these projects may be seen as high priority but discretionary, but the funding attached to some of the work could risk providing such a financial incentive as to distort the ONS’s overall priorities, and impact on other, ‘business as usual’ deliverables. A more formal framework would help balance the risks and rewards of the discretionary work more transparently.  

57. While a significant amount of collaboration between the ONS and external organisations, most significantly universities, already exists there is scope to develop further relationships, in order to build skills, extend the pipeline of talent and bring fresh perspectives, expertise and ways of working into government.

58. On efficiency, the recommendations are:

59. Recommendation 18: The UKSA should develop a framework to follow when considering ad-hoc commissions for statistics in order to be open about the opportunity costs of such work. 

60. Recommendation 19: The UKSA should step up efforts to build partnerships outside of government, particularly with universities and think tanks, given the clear economic and social benefits to this collaboration.

Approach to the Review 

61. Public bodies, which include those known as arm’s length bodies, play a key role in delivering public services across a range of sectors in the United Kingdom including health, the environment and statistics.  

62. The Public Bodies Review Programme was launched in 2022 to ensure that public bodies across the UK are operating effectively and delivering the best possible outcomes for citizens.

63. The reviews have three objectives:

• To ensure arm’s length bodies are effectively delivering services that meet the needs of citizens; 
• That they are doing so as efficiently as possible; and
• That they have clear governance and lines of accountability.

64. As part of this wider programme, this Review of UKSA was commissioned by Baroness Neville Rolfe, Minister of State for Cabinet Office, and launched on 5 June 2023. 

65. The terms of reference (Annex C) for the Review follow the four themes of the Public Bodies Review Programme: governance, efficacy, accountability and efficiency. The Review was asked to explore: 

66. Governance: The UKSA’s current governance arrangements, including the adequacy of the independence of the UKSA, the production and regulatory functions and whether there is any conflict between the respective parts of the UKSA and the UKSA’s approach to UK-wide data. Additionally, the Review will establish whether the UKSA board has a clear purpose, processes for the appointment of Board members, arrangements for financial and risk management and how well it communicates with stakeholders. Furthermore, it will examine how the UKSA works to support policy development across government and its wider role with respect to the whole UK statistics system, including the UKSA’s role post EU-exit. 

67. Efficacy: The extent to which the UKSA is producing statistics that respond to changing user needs and its ability to anticipate and prioritise without compromising on quality. The Review will also examine how the UKSA communicates statistics to different users and determine the extent to which the ONS is able to take advantage of developments in data science capabilities. Furthermore, this review will explore whether the UKSA and its constituent parts have clear remits and mandates, have the correct delivery model and clear, measurable performance indicators. 

68. Accountability: Whether there is an effective relationship between the UKSA and the Cabinet Office as sponsor department, if the UKSA’s accountability to Parliament is understood, specifically focusing on the relationship with PACAC and whether the Committee is content with its relationship with the UKSA. Additionally, the review will cover the UKSA’s compliance with relevant functional standards, including Data, Digital and Technology (DdaT) and analysis. 

69. Efficiency: The Review will examine the UKSA’s adherence to existing financial guidelines while identifying any possibilities to improve the productivity of the UKSA’s workforce. It should be noted that the UKSA has recently agreed significant efficiencies with HM Treasury. As such, the primary purpose of this review is not to seek to identify additional efficiencies on top of the existing plans for the spending review period. However, where relevant, observations or recommendations have been made that should further improve efficiency. Finally, this review will explore the UKSA’s talent planning and management.

70. This Review has been conducted alongside ongoing UKSA programmes and reviews which have their own oversight arrangements. These include: 

• The ONS consultation on the future of population and migration statistics (considering the future of the England and Wales Census);
• The development of the Integrated Data Service (a new cross-Government platform accessible to accredited researchers to access to data and analytical tools) led by the ONS;
• The OSR led review of the Code of Practice for Statistics; 
• A UK-wide review announced by the Secretary of State for Science, Innovation and Technology in October 2023, regarding data collection on biological sex in public bodies and due to launch in early 2024. 

71. The Review occasionally references these areas of work and makes some relevant observations, however it has sought wherever possible to avoid duplication. 

72. The Review has taken a comprehensive approach to evidence gathering. The call for evidence (June–July 2023) resulted in 38 responses, which has been supplemented with meetings with around 100 key stakeholders from all sectors including the GSS, other Government stakeholders, business, media, civil society and international statistics offices. The Review has also drawn upon desk research and in-depth interviews with relevant colleagues from across the UKSA and is grateful to all those who have engaged and contributed to the Review since it began.

Statistical Quality

73. Given that quality is central to the reliability, usability and trustworthiness of statistics it is helpful to set out some observations regarding statistical quality early in the Review. 

74. Statistical quality is a multi-dimensional, relative and dynamic concept. It refers to the combination of activities and choices made by the producer of statistics in the whole process, namely in defining what data are to be collected, in collecting, analysing and interpreting the data and in making them available to others. The Review employed the definition of the quality of statistics as used across the European Statistical system.  Six criteria are considered: relevance; accuracy; timeliness and punctuality; accessibility and clarity; comparability and coherence. 

75. Relevance is the degree to which the statistics, the ways they are produced and the concepts which underlie them, meet the needs of users.  

76. Accuracy denotes the closeness of estimates to the truth, and is impacted by many aspects of the methodology of data collection and analysis. Users tend to focus on the sample size, which is important as it impacts on the precision of the estimates. But many other aspects of the design and implementation of the statistical programme are also critical, particularly those which might result in bias such as failure to achieve a high response rate or a response rate which varies greatly over different parts of the population of interest. This underlies some of the recent concern about the results from the Labour Force Survey.

77. Timeliness and punctuality: The timeliness of information reflects the length of time between the dates to which the statistics apply and their availability. Punctuality is the time lag between the release data and the target date when it should have been delivered, for instance, with reference to dates such as those announced in the official publication calendar. A further aspect relating to the timing of data is the frequency of release.  

78. Accessibility and clarity: There are multiple aspects to the issue of accessibility of data including how to locate and identify the data one needs, whether it is easily accessible, whether there is a pricing policy, the format and delivery mechanisms for the data etc. Clarity refers to information (methodological reports, metadata, quality assessments, interpretation, data descriptions etc.) accompanying the data and the responsiveness of the official statistical staff to queries and requests for assistance.

79. Comparability is important whenever users wish to make comparisons over time, space, or non-geographical domains. Ensuring comparability matters so that users can have reasonable confidence that changes over time, differences across geographical areas or differences between groups in the population are a reflection of reality and not an artefact of different methodologies being used. A tension which often exists is between specificity (ie relevance) and comparability. The Review raises this particularly in relation to the balance to be achieved between UK data versus data specific to the policies in the four different administrations, but also in relation to the focus on collecting data in the same way over time to maximise comparability versus making changes in response to different demands or changing circumstances, or to make improvements in the quality of the data.   

80. Coherence of statistics relates to their adequacy to be reliably combined in different ways and for various uses, often in an attempt to mitigate different biases in different data sources. Sometimes data are combined which have different frequencies. For example, there is an increasing interest in combining data from administrative records which may be available in a more timely way with the results from surveys which can have the advantage of using statistically determined questions.  

81. As is obvious from the descriptions above some aspects of quality can be traded against other aspects. One of the most common trade-offs occurs between timeliness and accuracy. Accurate data too late to input into decision making are of limited value – a fact which underpinned some of the decisions regarding statistical priorities during the Covid Pandemic. The implications of quality being multidimensional is that there is no such thing as Absolute Quality. Data have to be considered in relation to their use and hence the term ‘fitness for purpose’ is used.  The OSR writes in the introduction to the Code of Practice for Statistics^{[footnote 6]} that “statistical quality means that statistics fit their intended uses, are based on appropriate data and methods, and are not materially misleading. Quality requires skilled professional judgement about collecting, preparing, analysing, and publishing data and statistics in ways that meet the needs of people who want to use the statistics.” 

82. This explains the emphasis in this Review on building a UKSA which has a good understanding of the needs of the wide range of current users and has good intelligence systems to be able to judge what data will be valuable in the future, combined with the confidence to make decisions about statistical trade-offs and priorities. The proposal in this Review of a Triennial Statistics Assembly is intended to support UKSA in this role.  

83. Although not a measure of quality, the costs involved in the production of statistics inevitably act as a constraint on quality. It is also vital for the UKSA to take into account the impact of response burden, particularly on businesses which bear some of the costs of data collection. UKSA also needs to be aware of the plans of other agencies which collect data – deliberate replication may be acceptable but ignorant duplication has costs for the UK.

84. Underpinning the achievement of quality data is the fact that we must ensure the integrity and honesty of the statisticians. The Professional Guidelines of the International Statistical Institute includes the following principle – ‘While statisticians operate within the value systems of their societies, they should attempt to uphold their professional integrity without fear or favour. They should also not engage or collude in selecting methods designed to produce misleading results, or in misrepresenting statistical findings by commission or omission’^{[footnote 7]}.  

85. One of the challenges to the integrity of official statistics is that there is an increasing focus on statistics used in performance measures. This has the problem that Ministers are being held to account by the statistics being gathered by statisticians within their departments.

86. The integrity of the data is particularly at risk if the indicators are used for naming and shaming in league tables.  It is vital, therefore, that statisticians feel that they are supported to uphold their professional principles if they are put under pressure to manipulate data.  We should do well to remember Goodhart’s Law “When a measure becomes a target, it ceases to be a good measure” or David Boyle’s Paradox ‘If we don’t count something it gets ignored If we do count it, it gets perverted’^{[footnote 8]}.

87. Whilst it is important to promote wider use of data in decision making, the danger with a measurement culture should be recognised in that sometimes excessive attention is given to what can be easily measured, at the expense of what is difficult to measure quantitatively even though this may be fundamental.

88. This Review has been prepared mindful of these observations and the importance that statistics are produced with independence and integrity.

The UK statistical system 

89. A defining feature of the UK’s statistical system is that it is decentralised: multiple public bodies across all four UK nations are responsible for producing and publishing the statistics that underpin our lives. This Review uses the UKSA’s definition of ‘the UK Statistical System’ which incorporates the UKSA itself, its respective production (ONS) and regulation arm (OSR) and the wider network of statisticians embedded in departments across government (GSS). 

90. Beyond government and public bodies, the UK is lucky to have a rich and vibrant civil society of statisticians and statistics users. This community and strength of relationship was commended in evidence from other national systems, some of whom have a more distant relationship between official statistics and academia. While the Review will discuss these wider groups at length – especially how the UKSA can best engage and harness their expertise – for the purpose of this Review any reference to the UK statistical system is defined as the following components: 

a. The UKSA: This is an independent body at arm’s length from government with the statutory objective of promoting and safeguarding the production and publication of official statistics that ‘serve the public good’;

b. The ONS: Perhaps the most recognised brand within the UKSA, the ONS is the UKSA’s production function and UK’s largest independent producer of official statistics including GDP, prices and the decennial Census amongst others. The Permanent Secretary of the ONS is the National Statistician, currently Professor Sir Ian Diamond, who also represents the UK on the international stage, most notably the United Nations Statistical Commission, as well as regional bodies such as the UN Economic Commission for Europe (UNECE);

c. The OSR: Formerly known as the Monitoring and Assessment function, the OSR identity was greatly enhanced following the 2016 Bean Review. It is the UKSA’s regulation function providing independent regulation of official statistics produced in the UK. It is led by the Director General of Regulation^{[footnote 9]}, currently Ed Humpherson.

d. The GSS: The GSS is a network of all those who are involved in the production of official statistics in the UK including other government departments and some arm’s length bodies. Importantly, official statisticians in Scotland, Wales and Northern Ireland are part of the GSS. The GSS is also part of the cross-government Analysis Function, which brings together seven different professions including economists and actuaries together to deliver analysis across government.

Figure 1: Overview of the UK statistical system.

91. The production of statistics began to be centralised in the UK during World War II at the behest of Winston Churchill. In 1941 the Central Statistical Office was established in the Cabinet Office, the UK government’s central coordinating department, in order to improve harmonisation and standardisation of the collection and production of some official statistics which hitherto had been entirely decentralised with individual departments housing their own statistics production facilities.  

92. In the eighty years since how the UK collects and produces its statistics has continued to evolve and integrate across the system, with each wave of reform seeking to improve the efficiency and quality of the system. Most significantly, in 1996 the ONS was established, merging the Central Statistical Office with the Office for Population and Census Surveys (which operated in England and Wales only) and created for the first time a large, central statistics body.

93. The next major reform took place in 2007 with the creation of the UKSA through the Statistics and Registration Service Act 2007 (‘the Act’). This created an independent, non-ministerial, statutory body, which reports directly to the UK Parliament, and, recognising the constitutional changes of the late 1990s, the Scottish Parliament, the Welsh Parliament and the Northern Ireland Assembly. 

94. The Act (which refers to UKSA as the Statistics Board) establishes the objectives of the organisation and its membership. The work of the Authority is further defined in secondary legislation created under the Act and other legislation including the Census Act 1920. Section 7(1) of the Act states that the Authority has the objective of promoting and safeguarding the production and publication of official statistics that ‘serve the public good’. The public good includes informing the public about social and economic matters; assisting in the development and evaluation of public policy; regulating quality and publicly challenging the misuse of statistics. 

95. The importance of statistics in today’s society cannot be understated and the UKSA plays a key role across the Statistical System. It fulfils its statutory responsibilities predominantly, – though not solely, – through the work of the ONS, which informs decision makers across society and helps citizens better understand the nation, and the OSR which regulates the quality of statistics and helps increase their value, coherence and accessibility.

Introduction 

96. The science of statistics helps us to understand the world and our place in it. From measuring what contributes to our economic prosperity and the impact of global economic shocks that have affected the whole of society in recent times, to helping us identify and track changing patterns of human life and behaviour. The power of statistics shapes our understanding of the world in which we live.  

97. Strong governance, oversight and methods to maximise the value of statistics and their use is crucial as new technologies, uses and requirements arise. Recent years have seen changes in the sources, uses and delivery of statistics. For example, better understanding of consumer spending patterns by using debit and credit card transactions data, making Census data more accessible to users through interactive maps to better understand the areas in which they live, and capitalising on the power of artificial intelligence and understanding public sentiment towards this. Effective communication of statistics is also critical. Often produced with levels of uncertainty and a ‘best estimate’ of the subject being measured, users of statistics need to be provided with information not only on the uses but also on the limitations of the data. 

98. The statistical system is multifaceted, comprising a range of organisations, governance arrangements and expertise, which combine to provide an environment of statistical production and standards across the UK. It is part of a global community of statistics producers, users and organisations that work together to set standards, identify user needs and address common opportunities and challenges impacting society. 

99. Given the importance of learning from other countries as well as ensuring that UK data has credibility in international and regional agencies, the Review consulted many national statistics offices including Australia, Canada, France, Ireland, the Netherlands, New Zealand and Sweden who highlighted the opportunities and challenges they are facing. Many of these resonate in the UK environment. The opportunities centred on the growth of digital data sources, alongside the development of technologies to enable the management and use of large data.  Taking advantage of these developments is seen to be essential in the context of falling response rates to conventional surveys, growing demands for more timely statistics, and pressures on public sector budgets, but is not at all straightforward when many statistical systems are tied up in legacy IT systems, and when there is often poor public understanding of the importance of data sharing. All the official statistical offices are concerned to build and retain trust in their outputs which is threatened by the rise of misinformation.  

100. In addition to these challenges, there are unique domestic factors that the UKSA must navigate. In particular ensuring data quality across a decentralised system, understanding the impact of devolution upon data collection and statistical coherence for the UK, and establishing fruitful relationships within the international statistics landscape following the UK’s exit from the European Union and Eurostat, the EU statistical agency. 

Review Findings 

101. The Review findings are described under each of the four themes of the Public Bodies’ Review guidance^{[footnote 10]}: governance, efficacy, accountability and efficiency as one way of navigating the inherent complexity of assessing a governance model that is inherently broad and deep in its construction. However, some considerations are naturally relevant in more than one theme. As such, material has been grouped accordingly and referenced in multiple areas. 

Governance

102. This section considers the current governance arrangements, including exploring the adequacy of the independence of the UKSA, determining if there is any conflict between its production and regulatory functions while also considering the Authority’s approach to UK-wide data, board effectiveness and diversity. Additionally, the Review explores whether the UKSA board has a clear purpose, processes for the appointment of Board members, the arrangements for financial and risk management and how it communicates with stakeholders. This also includes ensuring that it is operating with the appropriate degree of separation. Furthermore, it examines how the UKSA works to support policy development across government and its wider role with respect to the whole UK statistics system, including the UKSA’s role post EU-exit and its relationships with the Devolved Administrations.

103. The main conclusion is that it is time for the Board to step into a more visible, ambitious space. As is the case with many systems which have evolved over decades, there are ongoing perennial, structural issues that require strategic leadership to resolve. The UKSA Board is best placed to provide this leadership – both across government and with the sizable, active user community within the UK and beyond.

Recommendation 1: The main recommendation is that every three years the UK Statistics Authority will lead the establishment and delivery of a Triennial Statistical Assembly. This Assembly should involve key organisations inside and outside Government and across the four nations, with the remit of determining the UK’s needs for statistics through a wide consultative process. This should include the private sector, government departments, academia, think tanks and media representatives. The list of organisations involved should be reviewed ahead of each Assembly to ensure it captures a broad range of user and producer views. 

The UKSA will then respond to this by producing a proposal for the statistical priorities for the next three years, thus identifying data gaps and ensuring that users can hold the statistical system to account on the delivery of the programme of work. It will also enable other producers of statistics to complement the work of the official statistical system and factor this work into annual budget allocation processes.

To respond to the Statistical Assembly and to supplement its findings, an annual public lecture from the Chair of the UKSA should be delivered to provide an update on the work of the statistical system and priorities for the year ahead. This should build on the OSR’s annual State of the Statistical System report. The lecture would raise the profile of the Board, further user engagement and establish the UKSA’s leadership role in the statistical space.

104. In addition, recommendations include:

Recommendation 2: The Review recommends that the expertise of the senior staff of the ONS would be greatly enhanced by the appointment of a Director General for Methodology who would be a focal point for the improvement and communication of data quality, and who would foster engagement with senior methodologists in other national statistical offices and in academia.

Recommendation 3:  The Government should amend the statistical legislation so that the Act reflects current practice, taking the opportunity to make clearer the practical operation of the UKSA. For example, that the OSR reports separately to UKSA, not via the National Statistician, and that the Director General of the OSR is an Accounting Officer and is expected to report separately to PACAC.

Recommendation 4: The inconsistent application of Pre-Release Access to official statistics across the UK statistical system has the potential to undermine trust. The Cabinet Office and devolved legislatures should amend the relevant secondary legislation for each nation at the earliest opportunity to follow the approach to Pre-Release Access taken by the ONS in line with the Code of Practice for Statistics.

Recommendation 5: The UKSA should build on existing work and lead discussions between the four nations and strengthen the Concordat of Statistics to encourage more UK wide data by creating common standards and improving harmonisation where appropriate and mutually agreed. HM Treasury should ensure that funding is available to support the harmonisation of key data.

Triennial Statistical Assembly

105. The UKSA’s unique statutory responsibility to promote and safeguard the protection and publication of official statistics that serve the public good is increasingly valuable in today’s society which has a greater reliance on data. Many users, both inside and outside of government, feel that there is too little transparency in the way the statistical agenda is set, and are unclear how they might get their needs to be recognised. Further, the consultations which do take place are sometimes perceived to be somewhat tokenistic and not open to new suggestions and ideas from the user community. The UKSA should take a more prominent role in identifying data needs and using this to set the programme for UK statistical priorities thus demonstrating its interests in ensuring that the UK has the data needed for sound decision making across society.   

106. At the same time UKSA should feel empowered to make statements when they feel that the statistical agenda is not appropriately taking account of the needs of society. This is particularly the case when important data collection is stopped or constrained for financial reasons.  

107. This issue is illustrated by a recent example from March 2023: the Coronavirus Infection Survey, established in 2020 and reporting estimates of Covid-19 prevalence across the UK, was stopped after the UK Health Security Agency removed funding. Whilst the survey has now been replaced with a smaller alternative, this decision to remove a key health surveillance tool was met with concern by many, including those providing evidence at the ongoing independent UK Covid-19 Inquiry. Professor Sir John Edmunds of the London School of Hygiene and Tropical Medicine was recently quoted as saying “our surveillance is no better than it was in the beginning of 2020. During the pandemic we set up an amazing surveillance system for tracking the disease and the virus variants that were causing it. Then in 2022 as the pandemic was ending the shutters came down and that stopped. Everything has gone back to where it was before 2020”. 

108. The UKSA Board has the legitimacy, credibility and expertise to offer proactive leadership to resolve challenges and corral support to ensure that the UK has the data it needs to understand and address the issues of the day, even when funded elsewhere in government. As such, the UKSA should use this to its advantage and ensure that it is empowered to raise concerns about stopping data collection for products it does not fund at the highest levels of government. This is in compliance with the first of the UN Fundamental Principles of Official Statistics, namely: “official statistics that meet the test of practical utility are to be compiled and made available on an impartial basis by official statistical agencies to honour citizens’ entitlement to public information”^{[footnote 11]}.

109. The Statistics Assembly should be carried out in consultation with other organisations - such as the UK Research and Innovation (UKRI), the Chambers of Commerce, OBR, the Royal Statistical Society (RSS), the Federation of Small Businesses and the Academy for the Social Sciences - and should engage widely with statistics users and producers to understand their views regarding the priorities and data needs for the UK over the forthcoming period. This should include the UK Government and Parliament, the Devolved Administrations and Parliaments/Assembly, academia, private sector statistics producers and users, wider statistics users (including local government, businesses, media and civil society) and international stakeholders. 

110. The consultation should conclude in a public event and accompanying report where the Chair of the Statistics Authority can set out the findings of the consultation and react with a plan for the statistical programme. This will enable other organisations to plan complementary work, thus reducing data duplication and increasing the coherence of the various outputs. The Statistics Assembly output will be able to inform annual budgetary planning and departmental spending bids. The Assembly would also encourage departments with policy responsibility for specific data to respond to the proposals relevant to them. 

111. The Review heard feedback from a range of stakeholders that the current UKSA structure is not always clear to stakeholders, in particular there is a sense of limited separation between ONS and OSR, which has the potential to damage trust in official statistics. The suggested process and event would have the additional benefit of providing an opportunity to set out and explain the statistical system in a way which will benefit users and the broader public. 

112. It is critical that the statistical programme achieves an appropriate balance between the continuation of important time series and responsiveness to emerging issues.  Thus the Assembly should be cognizant of the need for continuity - over time, an Assembly will adjust and update the previous Assembly’s proposals rather than beginning from scratch. Similarly, users must resist using it as a way to make totally unrealistic requests for data. It needs to be conducted in a constructive way respectful of the impact of multiple demands on the system. The Assembly could also be seen as an international facing opportunity for the UKSA to play a part in the broader international conversation exploring common themes and challenges facing national statistics offices.

UKSA Board 

113. The UKSA Board is the overarching body to which both ONS and OSR are directly accountable, and the GSS is nominally accountable through the National Statistician. The Board is unitary, made up of a Non-Executive Chair and a minimum of five Non-Executive members, plus the National Statistician and Permanent Secretary of the ONS and the Director General for Regulation (who leads the OSR) and one of the three Deputy National Statisticians, who serve on rotation. Currently there are eight Non Executives on the Board including the Chair, Sir Robert Chote, who was appointed for a five year term in 2022. 

114. The Act states that Board members are appointed by the Minister for Cabinet Office, in consultation with the Chair of the UKSA Board, Scottish and Welsh Ministers and the Department for Finance and Personnel for Northern Ireland. All appointments are regulated by the Commissioner for Public Appointments, who has spoken with regret about frequent delays in the appointments process across government. 

115. Current Board members have a wide range of relevant skills and expertise and, crucially for the integrity of the body, are politically independent. Recommendation 10 supports the appointment of a NED with communications experience at the next opportunity. The new appointments to the Board expected in 2024 should further bolster the operational and political independence of the UKSA. The length of appointments should also be extended beyond three years to enable appointees to build their understanding of the complex statistical system and to have greater impact. 

116. The Board meets 11 times a year. As per good governance practice, a Board Effectiveness Evaluation is undertaken internally each year. The 2022 internal evaluation concluded that the Board is working well, but highlighted that: 

• The frequency of Board meetings may need to be reconsidered; 
• Board papers are good but could more clearly highlight the issues as well as the successes; and
• Papers should be clearer on what is worrying the Executive and what is required from the Board as a source of support.

117. The Review met with each of the members of the UKSA Board and the Board Secretary, as well as observing meetings of the full Board and the Regulation Committee, which acts independently as the OSR’s ‘board’. Consistent with the feedback in the Board Effectiveness Evaluation, members spoke positively about their experience of working on the Board, the management information (including Key Performance Indicators) that were provided and the professional, transparent way in which it operates. Non Executives also welcomed the induction process when they joined the Board and the encouragement they received to engage with specific teams across the organisation to better understand the work and priorities. 

118. There is concern that the frequency of Board meetings puts pressure on the Executive team and means that the papers they receive can feel like slightly amended management papers and, as a result of a discussion of some topics too frequently, they are not engaging at an appropriately strategic level. It was also suggested that time freed up by less frequent meetings might enable the Non Executive Board members to help the statistical system in more focussed ways. 

119. The peripatetic nature of Board meetings was also seen as positive, particularly when they are held in Scotland, Wales and Northern Ireland which provides an opportunity for the Board to engage substantively with the Devolved Administrations in person. However, it was clear that more can be done to exploit these opportunities, for example, guest attendees from across the GSS and Devolved Administrations should feel empowered to contribute to discussion beyond the items they are specifically presenting in order to provide reflections and expertise. 

120. In this respect, there was a general recognition among members that their focus on some of the more persistent features, challenges and risks of the statistical system - such as improving data cohesion, the barriers to data sharing across government or the international strategy post EU exit - could be stronger, and that engagement in some areas, for example with the Devolved Administrations, could be seen as ‘tokenistic’.  

121. There was also a recognition amongst some members that the way in which issues and papers are presented to the Board do not represent issues in the most frank manner or seek steers and advice from members. This can have the effect of masking the true extent of persistent challenges and risks facing the organisation and the statistical system as a whole. The Review is reassured to note that those submitting Board papers are now required to ask the Board specific questions to gather advice and insight and highlight areas of concern, and feels that this approach should continue in order to bring a greater level of constructive challenge into the UKSA meetings.  

122. The Review also examined the UKSA’s arrangements for financial and risk management. The UKSA Board has three sub committees: Remuneration; Regulation; and Audit and Risk. The Audit and Risk Committee provides scrutiny and advice to the Board and National Statistician in their risk management responsibilities and membership includes four Non-Executive members and two independent members. The Review heard evidence of an improved approach to risk management following changes at Board level, which led to a change from an annual review of risks by the Board to a rolling programme which involved deep dives into risks in different areas of the UKSA. The Review encourages the Board to keep this under review.

123. When discussing UKSA’s governance with external stakeholders, concerns were identified which can be categorised as follows:

a. The relationship between ONS and OSR. There is a perception amongst some of ‘cosiness’ between the UKSA’s production arm the ONS and regulatory arm, the OSR;

b. The role and remit of the National Statistician and the talent pipeline to ensure continued good leadership for the future;

c. How the UKSA leads on statistical matters across the whole of the UK, in its engagement with Devolved Administrations; 

d. The ambiguity of the UKSA relationship with, or responsibilities for, the wider statistical profession and system across government (the GSS); and

e. Whether the UKSA Board has a clear international vision and strategy post EU-exit. 

124. To these themes a sixth has been added, which is the visibility of the Chair and the UKSA Board across government and with stakeholders nationally and internationally. 

125. The UKSA’s current strategy for the GSS - Statistics for the Public Good (2020 - 2025) - is guided by four principles: Radical, Ambitious, Inclusive and Sustainable. To fully tackle the challenges and embrace the opportunities it faces across the GSS, the UKSA must draw on these principles effectively. It must be Radical in the way it delivers new data insights, Ambitious in the way it collaborates with partners, Inclusive in the way it describes society and Sustainable in the way it uses resources and infrastructure. The Review explores the work of the UKSA mindful of these principles and the importance of quality official statistics to enable effective decision making across society. 

ONS Executive Committee

126. Reporting to the UKSA Board within the ONS is the ONS executive team led by the National Statistician. The executive meets regularly through two core executive committees, the National Statistics Executive Group (NSEG) which includes cross-GSS representation and the Executive Committee. The Executive Committee is responsible for ONS business delivery. To be fully effective Committee members must avoid siloed working and work collaboratively to meet the collective goals of the ONS. 

127. The Review recommends that the expertise of the senior staff of the ONS would be greatly enhanced by the appointment of a Director General for Methodology who would be a focal point for the improvement and communication of data quality, and who would foster engagement with senior methodologists in other national statistical offices and in academia (Recommendation 2). Some of the most trusted statistical offices worldwide have very senior methodologists (often Deputy National Statisticians). This person could also assist the organisation in the better communication of quality in outputs.

The relationship between ONS and OSR

128. The UKSA Board oversees both a production arm, the ONS led by the National Statistician, and the regulatory arm, the OSR led by the Director General for Regulation. The latter has responsibility for setting the Code of Practice for Statistics (the Code), assessing compliance with the Code (including awarding National Statistics designation to official statistics that meet the criteria) and reporting concerns regarding official statistics for example, relating to quality. The Statutory Code of Practice provides guidance to ensure statistical best practice, centred around the principles of trustworthiness, quality and value.  

129. The existence of a regulatory office that upholds the principles of the Code and can report concerns about the quality and use of official statistics is unique to the United Kingdom’s decentralised statistical system. Representatives from international statistics offices observed with admiration the value of having such an office. Speaking at an event in September 2023, Walter Radermacher, former Director General of Eurostat and Chief Statistician of the European Union, commented that this is a strength of the UK statistical system. 

130. The co-existence of both a statistics producer (ONS) and regulator (OSR) within the UKSA structure was viewed very differently by those who engaged with the Review. Those working closely within the structure felt that it worked effectively, though acknowledged it did look ‘strange on paper’. Those less close to the structure observed that it could undermine accountability expressing concern about an apparent ‘cosiness’ between the two functions.    

131. This debate is not new, indeed since the creation of the UKSA in 2008, the pros and cons of such a structure have been well explored. Including in the Bean Review and the PACAC report on the Governance of official statistics in 2019. 

132. Over time steps have been taken by the UKSA to clarify and further formalise the separation and distinction between the regulatory and production functions. For example;

• The OSR was formally launched in November 2016 to establish a distinct brand for the regulatory function as recommended by the Bean Review. 
• A separate website and social media accounts for the OSR exist, distinct from the ONS and UKSA websites. 
• The OSR manages its own consultations and publishes its own vision, annual report and business plan focused on regulatory priorities. 
• The UKSA Board agreed in 2019 to change the Terms of Reference for the Regulation Committee so that it has a role in allocating the OSR’s budget. The Terms of Reference were amended so that the Regulation Committee ‘recommend the budget of the OSR to the Authority Board for approval and monitor progress against the OSR work programme and business plan’.
• The Director General for Regulation was confirmed as an Additional Accounting Officer within responsibility for OSR in 2020. This carries with it the responsibility for ensuring that resources approved by the UKSA Board for the OSR are used for the purposes intended and means that the National Statistician does not control the regulatory arm’s budget. 
• It should also be noted that the Director General for Regulation reports to the Chair of the Authority and not the National Statistician. 

133. These steps have helped to build a clearer separation between production and regulation functions that on balance the Review considers satisfactory. The Chair of PACAC, William Wragg MP observed in correspondence with the Review that ‘It is the view of the Committee that the steps taken by the Authority to more visibly separate its production and regulatory functions in the period since [the Report by the Committee’s predecessor in 2019] have been helpful and sufficient to address previous concerns’. 

134. The UKSA should continue to actively seek to clarify the organisational structure in its communications to avoid misunderstanding or confusion regarding the work and relationship of the ONS and OSR which can ultimately damage trust in official statistics. It was suggested that one positive step would be to introduce OSR email addresses. Currently both UKSA and OSR colleagues share a ‘@statistics.gov.uk’ email domain making it unclear as to who works in which arm of the organisation in email communications. 

135. The Review also observed a meeting of the Regulation Committee, who in practice act as the OSR’s board. The agenda included consideration of sensitive topics including the ONS’s handling of the sex and gender identity question in the 2021 Census. The discussion and challenge in the Committee indicated a healthy level of scrutiny and constructive debate that members of the Committee noted is a consistent feature of meetings.

136. The concerns of those who consider the relationship between the two parts of the organisation to be too close or unclear are understandable given the unique organisational structure. Others also raised questions as to whether the balance of the UKSA Board examining ONS issues is equivalent to that of the Regulation Committee examining regulatory matters. The existence of a small number of misunderstandings by users also appear to perpetuate, such as that the Director General for Regulation is line managed by the National Statistician (he is not) or that the National Statistician controls the budget of the OSR (he does not). Nor does the National Statistician attend Regulation Committee meetings. 

137. Having reviewed the organisation thoroughly, this Review is satisfied that there is sufficient operational independence between the ONS and OSR. The Review could find no tangible evidence to support assertions that the two organisations are too cosy or that a fundamental, unmanageable conflict of interest exists between the two that undermines the integrity or quality of the statistics produced by ONS, though it is important to pay attention to the perception of independent scrutiny.

138. Greater assurance could be provided to concerned users about the value of the OSR and the robust relationship it has with ONS. For ONS, this should be by way of an annual update to the UKSA Board (which should be publicly available) setting out where and how they have addressed the OSR’s recommendations that year, and broader changes that have been made within ONS as a result of OSR interventions. For OSR they should continue to implement the recommendations of the Sturgis Report (2023), particularly on monitoring and reporting on recommendations made, including the recommendations relating to the ONS. 

139. Whilst the UKSA is operating in compliance with the existing legislation the current descriptions of the governance arrangements in the act contribute to misconceptions and so the act should be updated at the earliest opportunity. The language regarding governance in the Act should be reviewed to ensure clarity on how the system operates in practice (Recommendation 3).

Devolved Administrations 

140. The UKSA has responsibility for the statistical system, including its regulation, for the whole of the UK and thus has significant interaction with producers and users in the Devolved Administrations.

141. Devolution has given Scotland, Wales and Northern Ireland the authority to collect and publish data in devolved policy areas, such as health and education. In reserved matters, such as the economy or work and pensions, data collection remains under UK government control and these statistics are produced by UK Ministerial departments or the ONS. Consequently, depending on the statistic, the level of geography that ONS is delivering varies, from a UK wide level (e.g. economic statistics like GDP); a Great Britain level (e.g. Opinions and Lifestyle Survey); an England and Wales level (e.g. the Census) to an England only level (e.g. health statistics). 

142. The working relationship between the four nations is underpinned by a Concordat which sets out the agreed framework for co-operation between the UK Government (including UKSA and ONS) and Devolved Administrations in relation to the production of statistics -  for and within the UK - and statistical standards. Last signed by all parties in October 2021, its focus is on data coherence; international standards and obligation; data sharing; consultation and wider collaboration and professional standards. 

143. The Concordat is supplemented by the Inter Administration Committee (IAC) which meets quarterly. It is chaired by the National Statistician and membership includes the Chief Statisticians of the Devolved Administrations and the Registrar General for Scotland amongst others. Its role involves considering statistical matters across the UK, promoting approaches to the production of coherent, harmonised data and considering UK government statistics needed for devolved purposes and devolved statistics needed for UK purposes.

144. The benefit of such a system is clear: Devolved Administrations have the ability to collect locally relevant data to better inform policy making and ensure accountability. This underscores the purpose of devolution, and such statistics are a vital tool for Devolved Administrations to make effective decisions. 

145. Nevertheless, stakeholders agree that the current system is complex and time consuming, with multiple points of friction. Despite good working relationships between individuals (which all sides were keen to stress are in place) the structural challenges within the system do have a major impact on output. 

146. These challenges are multi-faceted. The UK Government’s chief concern is the lack of comparable data on devolved matters: where the data is collected as a by-product of administration, differences in policies, definitions, timing and other aspects of methodologies can obstruct valid comparisons, potentially obscuring trends and disparities. This is most obvious in health, where different approaches mean it is not possible to compare the performance of the health systems of Scotland, England and Wales, thus reducing some of the value of the data which are collected. Comparable data are highly valuable and can be delivered through bespoke surveys, though these are expensive to establish. 

147. The additional concerns from UK Ministers and government are:

a. Current legislation grants Devolved Administration Ministers greater rights over DA statistics than UK Ministers over British (or English) statistics due to the provisions in the legislation requiring ONS to seek approval from DA Ministers before they produce or publish statistics relating to devolved matters. No such provision exists regarding UK Ministers on reserved and/or English, statistics; 

b. The Chief Statisticians of Scotland and Wales do not work out of arm’s length bodies, but rather out of Ministerial departments with potential ramifications for independence of statistics. This differs to the UK-wide (and English) setup where the National Statistician works in an independent arm’s length body;  

c. Since 2017 the ONS no longer provides Pre-Release Access to its statistics except for very exceptional circumstances (for example, during the Covid-19 pandemic^{[footnote 12]}). This is increasingly recognised as an international standard of the integrity of official statistics. However for statistics produced by other government departments and the Devolved Administrations Pre-Release Access remains common practice and varies in implementation: for example, in Scotland access is provided to Scottish Ministers up to 5 days before publication. 

d. Commissions of data from ONS (or other UK departments) to Devolved Administrations are not always fulfilled because the statistics teams in Scotland, Wales and Northern Ireland have to prioritise their work given their resources. This can lead to gaps in data. The Review was concerned to learn that this has resulted in examples of UK-wide forecasting being done using only English data because data from the Devolved Administrations has not been received. This assumes that everyone in the UK has the same characteristics and circumstances as in England - which is clearly the opposite of what was intended when devolution was introduced.  

148. From the perspective of statisticians working in the Devolved Administrations their plea was that there needs to be an understanding of the significant resource disparities across the UK. The statistics teams in the various parts of the Scottish, Welsh and Northern Irish governments are very small compared to the ONS. They expressed concern that there is insufficient appreciation that there may be different policy demands and different user communities in the four nations. 

149. It is clear that a change in approach is required to ensure that the statistical system is effective across all four administrations of the UK. The UKSA should consider how it can harness the ability of the leadership of statistics in the Devolved Administrations to input into strategic, as well as technical, issues and how it can help to strengthen the UK wide strategy. The Review learnt from those working in the Devolved Administrations that formal engagement on UK data is currently felt to be led at an operational, ONS level, rather than a strategic, UKSA level.

150. Additionally, insufficient advantage is being taken of positive developments in one part of the UK. Examples provided included the Scottish National Performance Framework^{[footnote 13]}, which was launched in 2007 to measure wellbeing outcomes agreed with the Scottish Parliament, which is a good example of democratising statistics from which other nations in the UK could learn. As ONS seeks to develop better methods of data sharing for research purposes, the successful SAIL Databank^{[footnote 14]} in Wales should be seen as an exemplary model. The Northern Ireland Statistics and Research Agency’s recent statistical prioritisation exercise is also a leading example of how to effectively communicate and consult with users ahead of making changes to outputs^{[footnote 15]}.

151. A solution that is acceptable to all parties must be found on data comparability. There was a generalised assumption amongst colleagues in the Devolved Administrations that any move towards ‘data comparability’ essentially meant that Scotland, Wales and Northern Ireland would need to adopt what England does. That approach, and mentality, goes against the spirit of devolution and will undermine or delay progress. The politicisation of data through, for example, league tables should also be resisted.  

152. The UKSA should lead discussions between the four nations and, ideally, set out reporting requirements for each nation, akin to how Eurostat operates within the EU. At the very least the Concordat should be strengthened to enable better UK-wide data, establishing common standards and improving harmonisation where possible. The UKSA should ensure that the proposed Triennial Statistical Assembly takes account of the needs of different parts of the UK, and that there are constructive discussions about the data which might be harmonised. 

153. These discussions should also review the funding arrangements in place for statistics across the UK. While it is beyond the remit of this Review to comment on devolved arrangements, a more holistic approach to funding seems warranted. Ultimately, if Devolved Administrations cannot provide the data required for UK-wide analysis to be made, this impacts UK policy makers. It is in everyone’s interest that a better end-to-end funding approach be established by HM Treasury at the next Spending Review to meet UK wide, as well as Devolved Administrations’ needs. 

154. Finally, the Cabinet Office and devolved legislatures should amend the relevant secondary legislation relating to Pre-Release Access to Official Statistics for each nation at the earliest opportunity to follow the approach to Pre-Release Access taken by the ONS in line with the Code of Practice for Statistics (Recommendation 4).

Efficacy

155. This section discusses the theme of efficacy and focuses on the extent to which the UKSA is able to produce statistics of relevance both to policy and changing user needs. Additionally, this Review covers how well the statistical system takes advantage of developments in the technology and the science of statistics. The question is raised as to whether the UKSA and its constituent parts have clear remits and mandates, with appropriate delivery models and measurable performance indicators. 

156. In summary, the recommendations are:

Recommendation 6:  The centre of government, led by Cabinet Office, HM Treasury and No. 10 must actively work to resolve the systemic, often cultural, barriers to data sharing between departments. All government departments, particularly those who own significant amounts of data, must prioritise data sharing for statistics and research purposes and, support the development of programmes such as the Integrated Data Service, to enable greater sharing of data across government for statistical and research purposes.  

Recommendation 7: The IDS is critical in facilitating the greater use of administrative data and bringing greater efficiencies to statistical analysis and decision making across government and academia. The Review therefore recommends that the ONS takes action to ensure that the purpose, scope and requirements of the IDS are clearly communicated and that the needs and concerns of departmental data owners are sufficiently understood. 

Recommendation 8: On the National Statistician role, Cabinet Office working with the UKSA Board should: 

• Commission a review of the role of the National Statistician well ahead of the next recruitment campaign, examining the many component parts of the National Statistician role in order to decide whether to propose changes to the role and what this may look like. This should also identify where changes to the Act may be required to facilitate the delegation or sharing of the National Statistician’s responsibilities; and
• Examine the talent pipeline and talent development structures it has in place across the GSS to ensure that those with potential to be future applicants for the National Statistician post and other senior roles in the statistical system are identified and nurtured. 

Recommendation 9: A suite of actions relating to communications should be adopted, including:

• Urgently improving the website so that it meets user requirements more effectively;
• Ensuring that there is a better understanding of the levels of uncertainty around specific official statistics, particularly economic, to reduce public (and government) surprise to revisions; and 
• Building partnerships with organisations that foster relevant communication expertise to improve engagement with the wider needs of users.

Recommendation 10:  The Authority Board should look to appoint a Non-Executive Director with relevant communications experience to advise and support the UKSA.

Recommendation 11: The Advisory Groups working with the National Statistician should become more formal: recruitment should be open and clearly advertised for applications. The style, design and attendance at the meetings should also be reviewed to ensure that they facilitate frank and constructive sharing of views and feedback. 

Recommendation 12: To demonstrate the commitment to user engagement and remind producers of its importance, the reference to ‘consult users before making changes that affect statistics or publications’ in the UKSA Code of Practice should be reinstated so that users are consulted before producers make substantial changes to statistical data collection or outputs. 

Recommendation 13: Internationally, the UKSA should:

• In consultation with His Majesty’s Government, prioritise the establishment and signing of a Memorandum of Understanding with Eurostat in line with the option provided in the Trade and Cooperation Agreement; and 
• Update its international strategy, Statistics for the Global Good, to provide more detail on how it will engage and lead within key global organisations, including the OECD and IMF. 

Recommendation 14: The UKSA should engage with the Cabinet Office to explore the consequences of mandatory completion of the Labour Force Survey.

Data sharing across government 

157. All national statistical offices in OECD countries are increasingly using, and realising the benefits of, administrative data supplemented by surveys in the production of their key outputs. There are multiple reasons for this: it is more cost effective; it reduces response burden, especially if the government already holds these data; widespread difficulties of maintaining survey response rates; technological advances including the ease of accessing data remotely and data merging developments; and the increasing demands for more timely data. There is also recognition that most difficult problems are multidimensional and need to be addressed by linking data across sectoral domains, though this is challenging. Some countries are further advanced than the UK in this respect - in particular those where there are national registers or identity numbers and where the public acceptability of the use of administrative data has been achieved by a focus on the value of data use.

158. To fully realise the potential of the data held by government, an effective system of responsible data sharing is critically important. This must be led at the highest levels within government. Evidence received by the Review highlighted repeatedly that a significant challenge to the UKSA’s efficacy is caused by the barriers to data sharing between government departments - this is most clearly seen in relation to the Integrated Data Service (IDS) programme which is essential to the development of better statistical systems in future. The costs of not addressing these challenges will substantially outweigh the costs of resolving the issues now.

159. Many people consulted by the Review highlighted the detrimental impact that these barriers have in undermining the ability of those working on data to deliver high quality statistics or research in a timely manner. The Digital Economy Act 2017 was introduced, in part, to facilitate better data sharing (though the research section notably excluded health data), in particular Chapter 7 of Part 5 which amended the Authority’s 2007 Act to give ONS greater access to data held within the public and private sectors to support the production of official statistics and statistical research. However, there is widespread disappointment that this has not resulted in the much needed step change of the quantity and frequency of data being shared. 

160. The general consensus from those that the Review heard from across the statistical system is that the challenge is more to do with culture than technical constraints. Though often cited as an excuse for not sharing, the legislative framework is in fact enabling. The challenging impact this has on data sharing for statistical purposes was highlighted by Professor Sir Ian Diamond when he appeared before the Independent UK Covid-19 Inquiry in October 2023. 

161. Whilst the Central Digital and Data Office in the Cabinet Office is responsible for data policy, standards and strategy, it is individual departments who are responsible for their own data. Departments can inadvertently see sharing data as a risk (often with little or no benefit to the individual department), this therefore incentivises departments to withhold data as a way of minimising risk for that department. Across government, clear incentives to share data (when appropriate) must be established, and action taken when data is erroneously withheld. Amongst senior leadership - both at Ministerial and Civil Service levels - there needs to be greater discussion and understanding of the significant costs of not sharing data across government, both to build a better understanding of the circumstances of the population and for the more efficient delivery of government services.  

162. In summary, the Review identified the following issues as barriers to effective and efficient data sharing across government: 

• Lack of clarity, or of consistency of interpretation, amongst departments as to what data sharing legislation permits, and whether their responsibility to protect privacy would be violated;
• Fragmented and technically incompatible systems;
• A lack of understanding as to the benefits or incentives for departments to share their data, with concerns amongst staff in departments that they are carrying the risks but see little benefit;
• Resource constraints, with departments reluctant to prioritise preparing data for research purposes over their core functions; 
• Cultural and organisational reasons, including a hesitancy to make administrative data available externally in case, for example, issues are identified regarding the quality of the data.

163. Unfortunately, these barriers to accessible data have existed in some form for many years. In his report for the Covid-19 Inquiry, Gavin Freeguard noted that over the past three decades there have been multiple reports, reviews or initiatives to tackle challenges to data accessibility^{[footnote 16]}. As the UK seeks to move away from surveys and rely increasingly on administrative data, resolving these tensions and blockages to data sharing is fundamental and the financial savings could be substantial. 

164. It is not possible for the UKSA to resolve this issue alone. Indeed the benefits of data sharing extend beyond statistical work and can support the more efficient delivery of public services. Thus the responsibility must lie with the Cabinet Office to bring departments together to remove the barriers to data sharing, to hold departments accountable when they are not sharing data as required by law, and to ensure that attention is given to the benefits of responsible, cross-government data sharing. In so far as there are financial constraints the involvement of HM Treasury at senior levels will also be required.

165. Thus what is needed is greater ownership of this problem at the centre, with a focus on delivery. In particular, the focus must be on producing tangible solutions to address the challenges, shifting cultural expectations and creating better incentives to promote sharing and address reluctance to share data. The centre of government, led by Cabinet Office, HM Treasury and No. 10 must actively work to resolve the systemic, often cultural, barriers to data sharing between departments. All government departments, particularly those who own significant amounts of data, must prioritise data sharing for statistics and research purposes and support the development of programmes such as the Integrated Data Service to enable greater sharing of data across government for statistical and research purposes. (Recommendation 6). 

Innovation and legacy technology

166. The ONS has a wide ranging and ambitious portfolio of innovation and transformational programmes with the aim of leveraging digital data and technology, in order to improve the production and utilisation of statistics across the expansive statistical system. 

167. The ONS’s Data Science Campus was created following a recommendation in the Bean Review to establish a new centre for the development and application of data-science techniques to the production of economic statistics. Situated in its own distinct space at the Newport office, the Campus has to date delivered a range of capability building work, individual projects and automated processes, for example helping to embed data science techniques such as reproducible analytical pipelines (coding best practice) and cloud computing throughout its statistical divisions across the GSS. They are currently researching the use of AI in statistical production and in responding to user queries, for example exploring Large Language Models (LLMs) to improve website search functionality via a tool called Statschat. 

168. The Review heard significant positive feedback concerning the contribution of the Campus during the Covid response, for example in developing the Coronavirus Infection Survey. Staff from the Data Science Campus also helped establish the No. 10 Data Science Unit, 10DS. However, the existence of small teams in the Campus, separate from potential users across the ONS and wider GSS, has meant that projects have not often impacted sufficiently on the regular work of official statisticians.

169. It is not unusual for innovative projects or schemes to be established outside of an organisation before being incorporated into the business-as-usual work. For example, in 2016 the Dutch statistics office (CBS) established the Centre for Big Data Statistics outside of its central office in order to identify new and existing big data sources and techniques. For ONS, now is the time to integrate the Data Science Campus more closely into the organisation in order to gain the benefits of greater collaboration. Synergies will no doubt be achieved by bringing data scientists closer to official statisticians and analysts across the ONS and GSS. There are several powerful examples of statistical offices incorporating research and development into their work. To name just one, Statistics Canada has a strong reputation for delivering quality work using multidisciplinary teams.

170. The Review heard from the leadership of the Data Science Campus who recognise this need for greater integration and is reassured by the plans they have which are centred around three core aims: to improve the quality, timeliness and granularity of information available to improve insights for users; to support the ONS in becoming more efficient; to build data and data science capability across the public sector. This work should be prioritised and supported by the UKSA to ensure delivery of these aims. The Campus should be ambitious in its work and seek not only to improve the ONS’s official statistics but those of the wider GSS too.  

171. More generally, the ONS has invested heavily in using new technology and methods in its Census and Data Collection Transformation Programmes, recommended in the Bean Review. This has helped to improve data collection and statistical processes, as well as introducing innovative Census Maps to improve data visualisation and access by place. The enhancements provided by these programmes not only helped underpin ONS’ swift and effective response to Covid-19 pandemic, but also subsequent surveys such as the Ukraine Humanitarian Insights Survey, facilitating the development and delivery of surveys in rapid time.

172. In order to consistently deliver high quality statistics amidst changing user requirements, technological developments and context requires the UKSA to be innovative in its approach. Statistics Netherlands (CBS) is an example of an organisation that has successfully developed an organisation-wide culture of innovation and openness to change. For example, it has established an internal grant for staff to bid for funding and/or release from regular work in order to experiment with innovation. This is a way in which the CBS takes some managed risks and improves staff motivation within a culture that is open to trying new things and finding better ways to deliver. Such a committee may help unlock potential innovations across the ONS. 

173. Inevitably, significant challenges in this area exist. In common with many other parts of government, the ONS is wrestling with legacy IT systems which act as a drag on improved productivity across the organisation. For example, the primary means by which users access data - the ONS website - requires a significant range of enhancements to ensure access and improved usability for users. Moving forward from legacy systems to new technology whilst not neglecting ongoing data collection and production is a challenge shared by many of the National Statistics Offices the Review engaged with. Whilst the respective technology and systems will vary across nations, the challenges and opportunities are similar. 

174. The UKSA should explore the appetite from their international peers to establish a UN Statistics Division ‘City Group’ to explore common challenges and opportunities to embrace new technology and reduce reliance on costly and burdensome legacy systems. It is critical too to involve users in the developments, as the point was made to the Review that their technological advances need to be synchronised with ONS’s since they too are dependent upon decades of coding.  Private companies have experience in how to fund and manage research and development alongside their regular delivery and ONS might benefit from building partnerships in this area.  

Data Ethics  

175. Ethical considerations are a core tenet of a statisticians’ work and increasingly important as new technologies, data sources and linking opportunities arise. The UKSA’s proactive work in this space to develop and share expertise across the statistical system both domestically and internationally is to be commended and supported. 

176. The Centre for Applied Data Ethics (CADE) was established in 2021 by the UKSA Data Ethics Team to lead the conversation and provide support to researchers and academics with the practical application of data ethics. For example, CADE has developed a data ethics self-assessment tool which helps analysts to apply the UKSA’s ethical framework to their research and mitigate ethical risks. It also works closely with the National Statistician’s Data Ethics Advisory Committee (NSDEC) to deliver transparent expert advice on the ethical use of data for analytical purposes and has published a wide range of guidance in collaboration with others across the statistical system. 

177. UKSA’s work as a leader in the data ethics conversation is highly regarded internationally.  Ironically it does not have the same profile amongst many of the users within this country. Given the important issues being addressed greater outreach and promotional work is called for, and the possibility of partnering with an academic department and a specialist institute might be explored.      

Integrated Data Service (IDS)

178. The IDS is aiming to generate a step change in the way data about our society and economy is made available and used, not only within government but through the statistical system. The programme is subject to a separate scrutiny process which the Review will not seek to duplicate, however the number of respondents who raised questions regarding the IDS with the Review team cannot be ignored given the transformative potential and importance of the programme to government.

179. The challenges of data sharing across government are long standing and significant. But as the UK moves increasingly to using administrative data sources and reduces its reliance on surveys, the IDS is critical in facilitating this transition and bringing greater efficiencies to statistical analysis and decision making across government and academia.

180. It is imperative that ONS works with government partners first to acknowledge the extent of any risks before setting out how they will be managed to facilitate the sharing of data with the IDS. It is critical that the UK joins other world leading National Statistics Offices who have already developed systems like the IDS. 

181. However, the Review heard a large number of concerns regarding the progress of the IDS with a variety of views as to what are the obstacles to greater progress. A general consensus exists that data sharing across government and the Devolved Administrations is critical to modern government, but a reluctance to provide data, including for the IDS, is a barrier to success and a central issue. However, views were also expressed that the programme itself could do more to explain the purpose, scope and requirements of the IDS, and that IDS staff could improve their understanding of and response to the needs and concerns of departmental data owners, as well as their engagement with academic users and providers of integrated data. 

182. Several users felt that they would appreciate greater clarity about the purpose of the IDS as well as how it will interact with the existing Secure Research Service which is used for significant academic research, much of it of direct policy relevance. The Review did not explore these issues in detail since there is a separate programme evaluation for the Integrated Data Service, but it urges the UKSA to be mindful of the implications of the transition for users and ensure that there is clarity to enable ongoing efficient access to data.

183. The IDS is critical in facilitating the greater use of administrative data and bringing greater efficiencies to statistical analysis and decision making across government and academia. The Review therefore recommends (Recommendation 7) that the ONS takes action to ensure that the purpose, scope and requirements of the IDS are clearly communicated and that the needs and concerns of departmental data owners are sufficiently understood.

Case Study: The UKSA’s Role during the Covid-19 Pandemic.

During the Covid-19 Pandemic the UKSA played a significant role in delivering data and analysis to support the pandemic response. The UKSA’s agility during the pandemic and work to deliver insights, including through the bespoke Coronavirus Infection Survey (CIS) was commended and admired by many of the International Statistics Offices the Review spoke to. This praise is given in full cognizance of the criticisms of the CIS, and there is no doubt that in less pressured and restricted circumstances significant improvements could have been made to the survey’s methodology, but the survey was commended and admired by many of the other National Statistics Offices who spoke to the Review, several of whom used ONS data during the pandemic because there were none available in their countries.

The ONS continued to deliver business as usual data as the pandemic changed daily life, adapting survey collection methods to ensure that economic and social measures including GDP and inflation data continued to be delivered.

ONS also adapted or introduced new surveys to meet data gaps and user needs. For example, the Opinions and Lifestyle Survey moved to collecting data more frequently examining the impacts of Covid-19 on daily life. Questions were adapted regularly in consultation with departments to seek insights relevant for policy making. Perhaps the most notable survey delivered by the ONS during the Pandemic was the CIS mentioned above. The ONS, working with partners, delivered the survey to provide information on Coronavirus infections and antibodies across the United Kingdom.

To ensure that user needs were met effectively the ONS embedded an individual within the Covid-19 Taskforce in Cabinet Office to better understand the data needs from policy makers and communicate them direct to ONS colleagues. This collaboration has been built upon as part of the ONS’s Policy Liaison Unit - a small ONS team that fulfils a similar role with policy departments promoting the use of statistics in policy making and understanding data needs.

The OSR  was also active during the pandemic, reviewing statistical publications, providing ad-hoc advice and monitoring the use of official statistics. They also, where required, took action to ensure that statistics were being used in line with the Code of Practice for Statistics in the public domain. For example, the then Chair of the UKSA wrote to the then Secretary of State for Health and Social Care to seek clarity about published testing targets.

National Statistician role

184. The National Statistician’s role is unique and complex, requiring - in statute - a strong statistical expertise together with the people skills and business acumen to lead a large organisation. Day to day the role requires managing the ONS; professional leadership of the GSS and the Government Analytical Function; representing the UK statistical system both domestically and internationally and acting as the chief advisor to the Prime Minister, Cabinet and Cabinet Secretary on statistical matters. 

185. Such is the breadth of the role, concern around the talent pipeline was raised repeatedly during the Review. It is important to note that this is not a reflection on the work or capabilities of the current National Statistician, nor his predecessors, but a concern as to who may fill the position in the future. These sentiments are well founded and the challenges of attracting and recruiting for this role are well documented. There are a number of reasons for this: the breadth of the role; the lack of a cadre of statisticians in senior civil service roles and the salary the role commands. 

186. The UK has been fortunate to have talented and passionate National Statisticians to date. The Review heard from many that the job description for the National Statistician is too broad. This is not the first time this has been observed, in 2019 PACAC recommended that the ‘UKSA should consider how to separate elements of the National Statistician’s role to make it possible to find suitable candidates’^{[footnote 17]}.  As one respondent to the Review noted, whilst post holders to date have done a commendable job delivering the responsibilities of the role “we are always seeking the hero individual”. 

187. In the absence of a significant cadre of senior statisticians across government, it is imperative that the role is attractive to external candidates also working in academia or the private sector. One way to do this is to review the salary available for the role, to ensure that it reflects the technical and specialist expertise required and attracts the best candidates. Consideration should therefore be given to the available salary ahead of any future recruitment campaign.

188. In light of the above, this Review recommends (Recommendation 7) that the Cabinet Office commission the UKSA Chair to review the National Statistician role and its component parts ahead of the next recruitment campaign. The aim must be to attract the best candidates. There should also be discussions with HM Treasury about the salary available for the role.  

189. The UKSA should also examine the talent pipeline and talent development structures it has in place across the GSS to ensure that those with potential to be future applicants for the National Statistician post are identified and nurtured. This needs to take place with some urgency as some of the skills to be developed might benefit from seconding individuals to policy roles or to other National Statistical Offices overseas.

Government Statistical Service

190. The GSS is one of the many professions working across government. It brings together civil servants across the UK who are committed to the collection, production and communication of official statistics. The GSS includes the Government Statistician Group (GSG) for those statisticians who have completed a ‘badging’ exercise, to ensure they meet professional standards. 

191. Since 2018 the GSS has sat within the Government Analysis Function (GAF), which houses seven analytical professions, from economists to social researchers to the digital and data professionals, into one functional group. While the distinction is not always immediately clear, the Analysis Function is focussed on government-wide standards and collective outcomes across the analytical professions, whereas the GSS pays more attention to the professional identity and practice of statisticians. The National Statistician leads both the GAF and the GSS, and a small secretariat for both is housed in ONS. There are strategies for both published on the Analysis Function website.  

192. The rationale for the GSS is clear. Evidence provided by members of the GSS to this Review highlighted the value of membership of the professional community and the importance of having a figurehead in the form of the National Statistician. It also allows the National Statistician to work nimbly across government to resolve cross-departmental issues as they arise (the Covid-19 response being the most obvious recent example) and for statisticians to work across departments effectively building skills and experience within a central framework. The GSS also has a Fast Stream Programme (for graduates to accelerate skills and experience within the Civil Service) and an apprenticeship offer, offering level 4 and 6 qualifications in data science and analysis. 

193. The GSS is integral to making the UK’s decentralised statistical system operate effectively and nurturing statistical talent. Heads of Statistical Profession, in addition to reporting to their Departmental management, have a dotted line and professional accountability to the National Statistician. Heads of Profession past and present welcomed this link and support from the National Statistician. They, and wider GSS members, also spoke effusively about the support they receive from OSR and they stressed the value of being able to have informal discussions with the regulator to sense-check any concerns they have about the impartiality and quality of their work. Given that many statisticians in departments are involved in gathering high profile and sensitive data this attention from the OSR is to be encouraged.

194. However, some members of the GSS - including departmental Heads of Profession for statistics - observed that there has been a sense that the profession had lost some of its status and identity in recent years following the introduction of the Analysis Function. This view is bolstered by the fact that the majority of Heads of Profession for Statistics are Deputy Directors (with grades varying between departments) who are identified and selected by their respective government department, in contrast with the more senior network of departmental Directors of Analysis who are selected by a recruitment panel that includes the National Statistician. 

195. This shift in focus reflects the wider government strategy following the introduction of the Functional Strategy under the Coalition Government. There is a clear benefit to bringing together the different professions delivering analysis - indeed many members of staff within the ONS are themselves members of different analytical professions, which is indicative of the operating environment in which statistics plays a vital role. Whilst the Review did not hear evidence to suggest that the purpose of the Analysis Function should be reduced, there was a strong sentiment that the status and identity of GSS should not be diluted, or devalued, by its existence.

196. More broadly, there does seem to have been a downgrading of statistician roles across the GSS, which is of concern. The UKSA should be more assertive in setting expectations across government on the expected seniority of the Head of Profession within departments: for those departments which produce a high number of statistics, and for the Chief Statisticians of the Devolved Administrations, the Review would expect these roles to be pitched at Director level. 

197. Much of the challenge around pay and grading stems from the wider Government People Strategy, which links pay to management rather than expertise. This has a detrimental impact on attracting and retaining specialists, such as statisticians (and others working in technical, scientific or analytical professions). The UKSA, including wider GSS, have no choice but work within the current system which limits the career offer to many talented specialists who have no desire to spend their professional life largely as managers. Complex pay remits, which are higher for Digital Data and Technology specialists, have exacerbated issues around pay, and impact on seniority and grading, causing further tensions within the wider Government Analysis Function. This Review would urge those working to reform the civil service to look into this particular issue to find a better approach to rewarding technical and scientific expertise, rather than the current focus on management skills, examining the opportunities provided by capability based pay. 

198. In terms of the management of GAF and GSS there is a perception that resources - including the time and energy of ONS’s leadership - which used to be directed to the GSS alone are now being shared across the wider function. Indeed, it is possible that this blurring of identity between the Analysis Function and the Statistics Profession is partially a consequence of both sharing the National Statistician as their figurehead.

199. It should not fall to the National Statistician alone to lead and support the GSS, and it seems clear that increased interest in the GSS from both the UKSA Board and Deputy National Statisticians would be welcome. To effectively do this, the UKSA Board must ensure that beyond high level updates they are sighted and engaged on key matters across the GSS regularly. The Review and Heads of Profession welcome the recent appointment of a dedicated Deputy National Statistician to engage with the GSS, and look forward to strong collaboration as has been the case in the past.  

200. Evidence from members of the GSS also suggested that the talent management currently in place to identify promising candidates across the GSS for the senior civil service was not sufficient to create a promising pipeline of talent. For example, the Review was struck that despite the strengths of the UK’s decentralised system the international links are mainly with the ONS. Although the Review was given examples of GSS members attending international meetings, which is to be encouraged since establishing strong cross-national links is of significant professional value to statisticians, more attention should be paid to international activities and ensuring they aid the professional development of statisticians throughout the statistical system. Domestically, GSS should make far greater use of secondments - for example, the current BBC secondment scheme - to build skills and ensure a more robust talent pipeline across government. 

201. Heads of Profession highlighted the work that they have been conducting with the ONS to consider how best statisticians in a range of departments can work effectively with the ONS to deliver the GSS strategy. This has been led by Heads of Profession with the support of the National Statistician and examines the themes of Governance, Standards, Capability, and Transformation. This work has the potential to address the challenges facing the GSS, reinvigorate the membership and strengthen its identity. The recommendations identified by the Heads of Profession should be supported by the UKSA. 

International Engagement 

202. As head of the UK’s National Statistics Office, the National Statistician represents the UK in international engagement and structures supported by the ONS’s International Relations Team and relevant technical experts. For example, the National Statistician leads delegations to the annual UN Statistics Commission and plenary meetings of the UN Economic Commission for Europe and the OECD. Such engagement is supported by the cross-GSS International Committee (GSSIC) which agrees the strategic direction for international engagement and functions as a communication tool across departments. Agendas and briefing for meetings attended by the National Statistician are shared with the GSSIC to enable departments to contribute and feed in.  

203. Since 2016 the ONS’s International Development Team has been working to support National Statistics Offices in the developing world. Guided by the international strategy Statistics for the Global Good 2020-2025, this work has helped partner countries to build capability, expertise and improvements. This work is largely funded through overseas development assistance and has supported a range of work including improving technical capability in Kenya ahead of their first digital census in 2019 and building data science capabilities in Rwanda. It is vital that this work is in full cooperation with the statisticians in the Foreign, Commonwealth and Development Office and that it is conducted in a way which is respectful of the priorities of the statisticians in the recipient countries, avoiding encroaching on data sovereignty.

Case Study: One Big Thing

One Big Thing is an annual Civil Service Wide training initiative focused on areas that have been identified as reform priorities. The focus of 2023’s programme is data, seeking to enhance data literacy and data science skills, with all Civil Servants encouraged to complete at least 7 hours’ worth of data training.

The ONS has played a large role in the delivery of the One Big Thing training including:

Practitioner Level Learning: In particular, delivering practitioner level learning with the Data Science Campus amongst others creating nine modules for Civil Servants covering topics including data ethics, data storytelling and data quality. At the time of writing almost 9000 Civil Servants have completed the practitioner level training since One Big Thing launched in September 2023.

Live Data Event: In October the ONS hosted a Live Data Event in Newport with speakers including the Cabinet Secretary and the Chief Operating Officer for the Civil Service. The event aimed to help civil servants learn about data ethics, quality and how to interpret and present data and was attended by over two thousand Civil Servants both online and in-person. Speaking at the event Cabinet Secretary Simon Case said “It’s on us as civil servants to bring data as early on in the decision-making process as possible to develop brilliant policy. If each of you can help five colleagues be more confident in using data, that will be an enormous contribution towards a better government. Your expertise can give them the confidence they want and need to be better and more effective in their work”.

One Big Thing Masterclass for Permanent Secretaries: In October 2023, the National Statistician was joined by economic journalist Tim Harford to deliver a masterclass for departmental Permanent Secretaries on communicating and using data to effectively to inform decisions.

Communication of Statistics

204. As illustrated throughout the Review, the importance of statistics in today’s society cannot be overstated. A crucial part of the UKSA’s role in safeguarding and promoting the publication of official statistics, relates to the effective communication of official statistics. Communications help convey the value of independent statistics and maintain public trust and engagement with them. As such, any shortcomings in effective communications risk these valuable foundations on which the UKSA sits. 

205.  The majority of ONS’s communications activity related to the c.1000 data and articles released annually. The style and frequency of publications varies across topics, according to the type of data being published and the range of user needs.

206. The Review heard from a large number of users regarding the ONS’s communication of statistics. Some publications were commended including the reporting of Covid-19 related statistics during the pandemic, the interactive Census maps published following the 2021 Census and the personal inflation calculator built for, and embedded on, the BBC website thus reaching a wide audience. Alongside the ONS’s publications, the names of lead statisticians are published allowing interested users to engage with them. The responsiveness and courteous nature of ONS statisticians in responding to users is appreciated.

207. Whilst these individual examples of good practice were highlighted, some users felt that the ONS’s overall approach to communicating statistics is not as effective as it could be and that it often lacks a coherent approach. In trying to reach a wide range of users and requirements, the ONS is communicating through many channels at high frequency, which poses a risk to the quality of materials and the engagement with them. For example, blogs published under the brand of ‘National Statistical’ on a separate standalone website, are at risk of ‘overwhelming’ a user interested in multiple topics and do not always deliver the quality of material that would be expected from an ONS publication informing public debate.

208. In September 2023, the ONS published the Blue Book, an annual update which details changes to Gross Domestic Product (GDP) to provide the latest and most accurate information based on newly available data from surveys and administrative data. The UK is one of a small number of countries that publishes GDP estimates on a monthly basis. Revisions are made as new data and information becomes available. An important bellwether for the economy, GDP statistics are amongst the core economic indicators published by the ONS. The Blue Book revisions published in September 2023 showed an upward revision of 1.1 percentage point which was met with surprise from some and commentary raising questions about the trustworthiness of the data, something that takes time and hard work to build. 

209. Whilst the ONS’s revision was not an outlier, with many EU nations subsequently revising upwards (including Spain and Italy) the surprise about these revisions from many, even those with significant expertise in economic data suggests that ONS’s communication was not as effective as it might have been. ONS did, at the time of release of the original GDP data, publish their concerns relating to the difficulty in collecting these data during a pandemic, but the message was not heard. ONS needs to consider how it can transmit a message about uncertainty more effectively. In this respect the Review welcomes the findings of the OSR Report^{[footnote 18]} on the GDP Revisions. Honesty regarding the limitations of data is critical to maintaining trust in the information, and statisticians must avoid over-claiming. Advice might be sought from the Science Media Centre in this regard. (Recommendation 9).

210. Given this relationship between maintaining public trust in official statistics and its effective communication underpins that, the Authority Board should look to appoint a Non-Executive Director with relevant communications experience to advise and support the UKSA (Recommendation 10).

Publishing and the ONS Website

211. The recent decision to move some market sensitive releases from a 9:30 am to 7:00 am publication time, a change delivered as a result of the pandemic and subsequently retained, has increased opportunities for the communication of ONS statistics. The earlier time facilitates coverage and interviews with ONS statisticians on morning radio and television that were not possible with a later publication time, and this greater visibility of official statisticians was mentioned positively by many of those contributing to the Review, who feel that this supports trust in the quality and impartiality of the data. However there has been a downside to the change in publication time since the pre-publication ‘lock-ins’ with accredited journalists, which allowed an orderly release of data and briefing to media recipients who would in turn communicate findings to their public audiences, can no longer take place. 

212. As a result of this change, media outlets highlighted that some users were unable to access publications at the moment they were released citing a problem which occurred due to so many organisations accessing the data at the same time. As a result the ONS has implemented short term ‘tactical fixes’ such as a Fair Use Policy to provide equal access to the website and publishes data on social media simultaneously to ensure equality of access should there be an issue with the website. The approach taken by other National Statistical Offices to ensure equal access to data should be drawn upon, including the use of Application Programming Interfaces (APIs) which enable two or more software programmes to communicate with each other.

213. The ONS website has long been critiqued for being difficult to navigate and search for required data. Many of those speaking to the Review shared these sentiments. In February 2016, the current version of the ONS website was launched following a redesign and delivered significant improvements. The Bean Review recommended that following the launch of the updated website the ‘ONS should continue to develop its new and greatly improved website in order to ensure that its full range of statistics can be easily accessed and viewed’. 

214. It is obvious that further work is required. A clear delivery plan, including user engagement and testing, should be created and sufficient funding allocated to and by the organisation to ensure that the ONS website is able to communicate statistics to users independently, clearly and equally (Recommendation 9).

User Engagement

215. Beyond central government stakeholders, there are a wide range of users of statistics, who not only have a wide range of different needs and working environments, they also have hugely variable skills and understanding of the data they wish to access. This makes it extremely difficult for statistical providers to service them all effectively. To assist with simplifying the different knowledge, requirements, and preferences of users, the ONS has identified five core user personas. These are identified as Expert Analysts, Information Foragers, Inquiring Citizens, Policy Influencers and Technical Users^{[footnote 19]}.

216. To meet the needs of the wide range of users, considerable effort is expended. This is guided by the User Engagement Strategy for Statistics. Examples of such engagement are the ONS Economic Forum events and Local Authority roundtables, together with the recent initiative of ONS local (with statisticians outposted to different parts of the UK to foster an understanding of the specific requirements in that area).

217. But the success of this engagement varies across theme or topic areas. Whilst some feedback described the recent User Engagement Strategy as ‘valuable’ the Review also heard from users about a tendency for the ONS to be overly defensive when receiving feedback and those working outside of government reflected that it can appear as if government users’ requirements are prioritised over others. Those who submitted evidence in relation to their experience on engaging with ONS over the 2021 Census questions and guidance on the questions relating to sex and the new question on gender identity stated that in this instance ONS’s approach to stakeholder engagement was unhelpful.  

218. To truly deliver Statistics for the Public Good, it is important that genuine user engagement is embedded across all theme areas, that it takes place throughout the life cycle of statistical outputs and is conducted constructively. Mutual respect and understanding from users, as well as producers, is needed to understand each  others’ position, particularly around the ONS’s need to prioritise given its significantly reduced funding envelope going forward (see efficiency section). 

219. Given the breadth of users drawing upon GSS statistics, UKSA must be cautious of trying to meet all user needs but should seek partnerships with the media, with academics and other experts to assist in the understanding of user needs and to develop a range of services and delivery mechanisms.  Professional organisations such as the RSS can help to coordinate and host users for fulfilling their roles as critical friends of the UK statistical system. 

220. Furthermore, the Review was surprised by the limited engagement from the business community during the course of the Review and is concerned that this might be an indication of a wider lack of interaction between the private sector and the statistical system. The Review urges the UKSA to seek to address this in enduring and sustainable ways.

221. Additional ways in which UKSA can improve its user engagement beyond governance include:

• An Annual Chair’s Address: The leadership of the UKSA plays a key role in communicating the work and priorities of the organisation. An annual public lecture from the Chair of the UKSA should be delivered to provide an update on the work of the statistical system and priorities for the year ahead for users to engage with. This would also raise the profile of the work of the Chair and Board, which would support user understanding of their work and give greater strength to the regulatory aspects of their work. The Annual Address should build on the OSR’s annual State of the Statistical System report and would also form part of the Triennial Statistics Assembly mentioned in Recommendation 1. 
• Open recruitment to Advisory Committees: To support the work of the ONS, the National Statistician has a set of advisory panels which provide him with advice on different topics such as Data Ethics and Consumer Price Statistics. The National Statistician’s Expert User Advisory Committee sits across these topics and includes representation from each of the Committees. The existence of these Committees is to be welcomed however they were little known amongst users contributing to the Review. Furthermore, it was proposed that their remit could usefully support the broader system of the GSS and OSR and not just the National Statistician. Given the valuable role these Committees play, it is important that suitable applicants and subject experts are aware when vacancies arise, and feel that their applications are taken seriously. Feedback from the Committees and recruitment processes should be as comprehensive as possible so as to promote user engagement but not at the expense of candour in meetings. The style, design and attendance at the meetings should also be reviewed to ensure that they facilitate such candour. (Recommendation 11).
• Amending the Code of Practice for Statistics: The Code of Practice for Statistics sets out the standards that producers of official statistics should follow. The Code in 2009 included the requirement to ‘consult users before making changes that affect statistics or publications’. Whilst the 2022 version of the Code includes references to user engagement, the requirement to consult users ‘before making changes that affect statistics’ no longer appears. Evidence received by users (particularly those working in local government) suggest that there can be major difficulties in their work if substantial changes to statistical releases are made with no prior warning. They pointed out that this is an issue for the GSS beyond the ONS. To demonstrate the commitment to user engagement and to remind producers of its importance, this clause should be reinstated in the Code so that users are consulted before producers make substantial changes to statistical outputs. Since the OSR has embarked on a review of the Code of Practice there is an opportunity to correct this omission (Recommendation 12).
• Given the success of the Code of Practice within government, the OSR should be encouraged in their commitment to roll out a ‘voluntary adoption’ scheme for all statisticians working in the UK, not just official statisticians. 

Regulating the use of statistics

222. The Review also examined the regulatory work of the UKSA and OSR. The Act sets out the UKSA’s role in safeguarding and promoting the production and publication of official statistics, a central part of which involves challenging their misuse. The public interventions made by the Chair of the UKSA and OSR were recognised by many who provided evidence to the Review as being central to ensuring the integrity of official statistics, and this was a role viewed as increasingly important in an age of misinformation. 

223. Many statistics producers, including Heads of Statistical Profession, commended the support and guidance they received from the OSR. In particular, they welcomed the OSR’s constructive and non-combative approach which allowed them to share and resolve difficulties heading off any potential problems. This is a deliberate approach, with members of the OSR team wishing to be seen as a supportive, critical friend to those within the GSS and as a ‘watchdog’ externally. Feedback suggests that cultivating this enabling approach through building trust between the regulator and producers is having a positive impact on compliance with the Code of Practice for Statistics.

224. Where official statistics are misused in public, the UKSA often seeks to correct misunderstandings and promote better use of statistics in future. Writing a public letter is one method that the UKSA uses to respond and promote behavioural change. Where appropriate, the UKSA engages with those misusing statistics privately, consistent with its ‘critical friend’ approach.

225. According to the OSR’s Intervention Policy, when deciding whether to publish an exchange they ‘consider the public interest in the issue and the value in making correspondence and statements public’. Whilst most correspondence comes from the Director General for Regulation, the Chair of the UKSA will typically respond where the issue relates to use by Ministers and elected representatives, whose prominent use of statistics is more likely to impact on public trust in official statistics if not aligned to the Code of Practice. In recent years the Chair has written to regarding the use of statistics both by the Prime Minister and Leader of the Opposition.

226. The public nature of the UKSA’s correspondence is expected to encourage behavioural change and the effectiveness of this was considered in evidence received by the Review. Options to improve the response to UKSA interventions were mooted for example, the introduction of fines. However this could divert resources away from organisations where the funds could be better used to produce quality statistics and could foment unhelpful arguments over the politics of the numbers.

227. The Review considers that the current approach is effective but could be strengthened further with increased, proportionate interaction with PACAC, or other Parliamentary Committees. Consideration should also be given to the tone and style of the letters sent. The UKSA’s Intervention Policy notes that it will consider ‘the impact of the breach’ when considering their response. Previous letters were considered by some to be too polite in tone and there was appetite amongst those the Review heard from for more robust responses to breaches of the Code of Practice; for instance, exploring the damaging impacts of the breach and requesting defined remedial action from the recipient such as the publication of a public correction. 

228. It should also be noted that success of the regulator should not be judged by the number of public letters sent. The approach of both private and public engagement by the Authority means that public letters are not always the best way to encourage behavioural change.

229. Consideration should be given as to how to involve PACAC more regularly particularly in relation to the work of the OSR in order to increase Parliamentary and public scrutiny of the issues being examined. This could enable the Committee to call to account those not meeting the standards of the Code of Practice. One possibility is for the OSR to submit a regular frank report outlining the interventions, explaining where success and improvements have been achieved and also highlighting more difficult issues such as repeat offenders. PACAC might wish to appoint a champion for statistics amongst its members who could be a key liaison for these issues.

230. Trust in statistics is undermined when they are misreported, particularly by people in the public eye, even when this misreporting is inadvertent. The Review wishes therefore to commend the strengthening of relationships between official statisticians and reputable fact checking organisations. Similarly, initiatives to help to raise the statistical competence of the media are to be encouraged.

231.   There is an opportunity for Ministers to take a lead in relation to the sound communication of statistics. The Ministerial Code currently states that ‘Ministers need to be mindful of the UK Statistics Authority’s Code of Practice which defines good practice in relation to official statistics, observance of which is a statutory requirement on all organisations that produce National Statistics in accordance with the provisions of the Statistics and Registration Service Act 2007’. 

232. It is the view of both the current Chair of the UKSA^{[footnote 20]} and his predecessor^{[footnote 21]} that this requirement could be stronger. PACAC also recommended in its 2021 report Government transparency and accountability during Covid 19: The data underpinning decisions that the Ministerial Code should be amended so that Ministers are required to abide by the UKSA Code of Practice in their presentation of data’^{[footnote 22]}. The Review agrees with this view and believes that bolstering the Code in this way will send a clear signal to the country that Ministers are holding themselves to the highest account.  

International

233. The UKSA is seen as a significant player in the international statistics community by senior staff in many other national statistical offices, who commended aspects of the UKSA’s work and leadership. Written submissions to the Review also recognised the strong role that the UKSA plays in international fora, for example at the United Nations Economic Commission for Europe. 

234. International comparability of data is important for three main reasons. First, because developing comparable data and standards provides National Statistics Offices with the opportunity to influence such standards and technical developments before changes are agreed. Second, in order to learn and share expertise from other peers. Third, ensuring that data are compiled according to international standards provides them with credibility and protects against inappropriate political interference with the data.

235. Following the UK’s departure from the European Union, the UK is no longer a member of Eurostat, the European Statistical Office. The departure from Eurostat reduces the influence which the UK can have on the development of methodology and standards for statistics across Europe, and also means that the UK has to replace the benefits of cross-national cooperation through working with other agencies and through bilateral cooperation. One important aspect of Eurostat’s work is the formal Eurostat Peer Review process, participated in by all National Statistical Offices (NSOs) of EU Member States and EFTA Countries. Peer reviews are conducted to assess members’ compliance with the European Statistics Code of Practice, but importantly they provide recommendations for improvements which benefit greatly from the experience of senior statisticians from other countries.

236. Whilst the Eurostat peer review process is no longer available to the UK (other than in the case of Gross National Income) there are many other high-performing NSOs that the UKSA regularly engages with whose expertise can be used to provide valuable feedback and constructive challenge. Notable peers and their strengths include: 

• Netherlands: who have established a strong culture of innovation
• Canada: who have high levels of public trust in statistics
• Ireland: who effectively engage with local needs for data
• Sweden: who maintain a reputation for high quality methodology and recently integrated this work with technological developments
• New Zealand: who produce inclusive statistics having built strong user engagement with Māori and Pasifika populations
• France: who have focussed on the professionalisation of statistics
• Australia: who promote the value of continuous improvement.
• USA: who build strong and productive relationships with high level economists and others with expertise of relevance to official statistics 

237. The ONS regularly engages with these and other peer nations. The UKSA should ensure this collaboration continues, and that it includes statisticians in other parts of the UK statistical system too, alongside exploring the possibilities of staff loans and interchanges. This would strengthen peer engagement and help ensure sufficient expertise is available for peer reviews. The UKSA should also build on the work that has been done by the National Statistician’s Committee for Advice on Standards for Economic Statistics established post-EU Exit.

238. An area identified by the Review for immediate engagement with peers relates to the falling response rates experienced by the Labour Force Survey (LFS). With the exception of the Census, the ONS does not mandate individuals to complete surveys unlike in countries such as Australia (though some of the business surveys are mandatory). The Review heard evidence that a change to statistical legislation to mandate responses to the LFS and other similar surveys would be welcomed both to improve response rates and statistical quality. The Review recommends that the UKSA should engage with the Cabinet Office to explore the consequences of mandatory completion of the LFS (Recommendation 14).

239.  The Director General for Regulation has also indicated that his team are already examining how it can work in the space left by Eurostat, through providing external scrutiny on statistical releases. With a team of fewer than 50 people the OSR does not currently have the capability to deliver external scrutiny comparable to that provided by Eurostat or other international peers. As OSR is beginning to step into this role it will require additional resources and must consider how to utilise the expertise of other National Statistical Offices. 

240. Though the UK is no longer a member of Eurostat, Article 730 of the Trade and Cooperation Agreement negotiated between the UK and EU provides that ‘EUROSTAT and the United Kingdom Statistics Authority may establish an arrangement that enables cooperation on relevant statistical matters’. To ensure that the quality of UK Statistics remains high and benefits from collaboration with other European Statistical Institutes, the UKSA should, in consultation with His Majesty’s Government, prioritise a Memorandum of Understanding with Eurostat in line with the option provided in the Trade and Cooperation Agreement. 

241. Finally, the UKSA’s international strategy for the statistical system entitled Statistics for the Global Good (2020-2025) has the ambitious mission to both ‘lead in the modernisation and harmonisation of international statistics and shape the development of internationally comparable standards to support international and UK policy objectives’ and ‘be at the forefront of the data revolution and use our knowledge and skills to help others respond to global challenges, and the UK to deliver its global priorities’. This is a commendable mission supported by clear steps to ensure it is delivered. Whilst the Strategy highlights that the statistical system must ‘adjust to the new circumstances, bolster governmental and people-to-people links across Europe’ it lacks detail on how this will be fulfilled. The UKSA should update its strategy on how it will engage and lead with key global organisations, including the OECD and IMF.

Case Study: OSR Review of HMRC

In September 2019, HMRC discovered a significant error in the published Corporation Tax receipts between April 2011 to July 2019, resulting in a revision downwards equating to £6.95bn over three years. In response, HMRC commissioned the OSR to carry out an independent review of the principles and processes underpinning the quality of HMRC’s official statistics. This was a different style of Review for the OSR who typically look at specific publications or themes of statistics. Led by Director General for Regulation Ed Humpherson, the Review aimed to assess the approach HMRC had taken to manage quality and risk in the production of statistics, while identifying improvements and providing recommendations to lower the risk or errors. 

Within HMRC, the Knowledge, Analysis and Intelligence (KAI) directorate is responsible for 87% of total official statistics production including those related to Benefits and Credits and Business and Personal Taxes. The review considered the roles and responsibilities of the managers, risk management and quality management and assurance processes. A wide range of stakeholders’ views were gathered, alongside a review of current guidance and best practices from other departments were examined.

In total, the review identified five key themes, comprising nine findings and recommendations; including increasing efficiency by reviewing and reducing the amount of publications HMRC produces, publishing information in a transparent way that is easy to find and ensuring knowledge sharing is prioritised at all levels. Ed Humpherson also acknowledged in correspondence with HMRC that current IT systems were not fit for purpose, arguing “improvements to quality management requires wider and sustained investment”. In response, Chief Executive and First Permanent Secretary Jim Harra said he was “grateful to the OSR for undertaking the review” while adding that “HMRC welcomes the findings and recommendations and we accept them all”.

Public Sector Equality Duty 

242. UKSA, ONS and OSR take their responsibilities under the public sector equality duty seriously. Internally an Inclusion & Diversity Steering Group (IDSG), made up of employees from diversity networks and other key stakeholders, evaluates new initiatives and measures progress against existing commitments and raises issues. It reports to the People Committee. Equality impact assessments are undertaken on key deliverables, such as the 2021 Census.

Accountability

243. This section explores the theme of accountability, specifically the adequacy of the current accountability arrangements between the UKSA and the Cabinet Office to determine if it is effective and appropriate. Furthermore, this Review explores whether ministers have enough assurance to meet their duties, by covering how accountability of the UKSA to Parliament is understood, specifically focusing on the relationship with PACAC. Additionally, the review sought assurance that the UKSA complies with relevant functional standards, including Data, Digital and Technology (DDaT) and analysis.

244. In summary, the recommendations are:

Recommendation 15: The UKSA should build on its work engaging with Parliamentarians and Select Committees of both Houses and in the Devolved Administrations further. It should continue to seek out opportunities to proactively add insights and value to the work and interests of the Select Committees. Equally, Select Committees should actively seek to use ONS data and analysis. 

Recommendation 16: The Cabinet Office should look to supplement the existing sponsor team to ensure that it is appropriately resourced to give the support and resilience required. 

Recommendation 17: The 2020 Memorandum of Understanding between the UKSA and Cabinet Office as Sponsor Body should be reviewed by both organisations to ensure that it reflects the requirements in the HM Treasury Framework Document Guidance for Arm’s Length Bodies, mindful of the UKSA’s statutory independence.

Relationship with UK Parliament

245. To safeguard the UKSA’s independence as a Non-Ministerial Department, the Act made the UKSA accountable to Parliament. Parliamentary Select Committees, whose Membership are made up of Members of Parliament or Peers (dependent on the House they sit in), are established to examine and scrutinise the work of Government Departments or specific issues. Recent scrutiny of the UKSA’s work has been conducted by PACAC and before that by its predecessor committee the Public Administration Select Committee (PASC). It is reassuring to note that in correspondence regarding the Review, the Chair of PACAC William Wragg MP stated that ‘Supporting the Authority’s independence has been, and will continue to be, a priority for Members of this Committee’. 

246. As part of PACAC’s scrutiny of the UKSA it regularly hears from the UKSA Chair, National Statistician and Director General for Regulation to examine their work and delivery against targets. Between June 2022 and May 2023 officials from the UKSA provided oral evidence to PACAC and other Select Committees nine times. In addition to regular joint sessions with the senior officials of the UKSA, PACAC should consider holding an individual session with the DG for Regulation on an annual basis to focus on regulatory matters. 

247. As highlighted in correspondence from the Chair of the Committee, PACAC’s scrutiny is also delivered through other means. For example, Committee Inquiries into different topical issues, corresponding with the UKSA on topical issues and holding pre-appointment hearings with the government’s proposed candidate for the role of Chair and post-appointment hearings for incoming National Statisticians.  However only four members of the Committee took part in the pre-appointment hearing for Sir Robert Chote in 2022. 

248. The UKSA regularly engages with other Select Committees in both Houses and those of the Devolved Administrations, publishing the written evidence it provides on the UKSA website. This proactive engagement with other Committees should be prioritised, especially by the ONS. As the producer of many of the statistics used by politicians they are able to offer unique insights into the data and how they can be best used. 

249. This Parliamentary engagement by the UKSA is to be encouraged, especially with the devolved legislatures (where its engagement is not as developed), to offer insight and expertise on matters of interest. Equally, Select Committees should be encouraged to actively seek to use ONS data and analysis in their work. Strong relationships with the House of Commons Library are valuable in this regard and the exploration of secondments for UKSA colleagues to the Library would have great mutual benefit (Recommendation 16).

250. In addition, politicians and select committees routinely digest and communicate statistics in the public arena. Since 2020 the ONS’s Data Science Campus has been delivering a Data Masterclass covering data and analytical literacy for senior public-sector leaders. This is a valuable and welcomed resource which contributors to the Review felt should be offered more widely across the senior leadership of the Civil Service. Mindful of a forthcoming general election, there will also be an opportunity post-election to brief new Members on the complexity of the statistical system and offer an invitation to the Data Masterclass. In the first instance, the UKSA could prioritise offering the Masterclass to those Select Committees it regularly engages with. 

Cabinet Office Sponsorship 

251. In addition to its accountability to Parliament, as part of its governance arrangements, the UKSA also engages with the Cabinet Office as its sponsor department in line with the Act which states that the Minister for Cabinet Office has residual responsibilities for the UKSA. These include: the appointment of non-executive members of the Board, accounting for statistical business in Parliament which can only be conducted by a Minister and, though it has never been used, the power as per Section 29 of the Act to issue a direction in the event of a serious failure by the Board.

252. Prior to the 2007 Act, HM Treasury was the sponsor department for the UKSA before it transitioned to the Cabinet Office. External users see this change as important in signalling greater independence for UKSA. 

253. The consensus from those working in both the UKSA and Cabinet Office is that the current relationship is effective and there are no calls for any change to be made to the sponsor department. The sponsor team currently sits within Corporate Services as part of the Governance Team and has two colleagues allocated to the UKSA who conduct this work as part of their wider roles (0.5 FTE of 1 x Grade 6 with oversight from a Deputy Director). 

254. The size of the sponsor team has fluctuated since 2008, and whilst the volume of work also changes throughout the year, the sponsor team in its current form is under-resourced. Without full-time resources working on matters relating to the UKSA, the team struggles to be across the wide range of matters related to the work of the UKSA and respond effectively to emerging issues. The Cabinet Office should look to supplement the existing sponsor team to give the support and resilience required (Recommendation 16).

255. In addition, a Senior Sponsor at Director Level is allocated to work with the UKSA. In November 2023 it was announced that the Senior Sponsor role would be split across the Heads of Analysis in the Cabinet Office and No10. The sharing of sponsorship responsibilities between two departments has generated some concern amongst the statistical community regarding the potential implications for statistical independence, particularly where views may differ.

256. The relationship between Cabinet Office and the UKSA is set out in a public Memorandum of Understanding signed by the Cabinet Office Minister, Chair of the UKSA and the National Statistician. This details the responsibilities of both organisations in clear terms. Since the publication of the Memorandum in April 2020, updated guidance on Framework Documents (which set out arrangements for sponsor departments to monitor and understand their ALB’s strategy, performance and delivery) for Arm’s Length Bodies has been published by HM Treasury. The Memorandum, which is overdue for its two year review, should be reviewed by both organisations to ensure that it reflects the requirements in the guidance for Arm’s Length Bodies, mindful of the UKSA’s statutory independence. The Review was assured that plans are in place to review the document in 2024 (Recommendation 17). 

Strategy and Business Plan 

257. In addition to the UKSA’s accountability to Parliament and the Cabinet Office as sponsor department, the organisation is accountable to the users of its statistics. A central way in which users can hold the organisation to account is via its five yearly, Board-approved, strategy document entitled ‘Statistics for the Public Good’. 

258. Published in 2020, the strategy has the mission to deliver ‘High quality data and analysis to inform the UK, improve lives and build the future’. It is a strategy for the UK statistical system, including the ONS, OSR and GSS. It sets out the UKSA’s ambition for the statistical system to build on the momentum achieved during the Covid-19 pandemic where data sharing, innovation and awareness of data were prominent. The Strategy is built around four key principles, Ambitious, Radical, Inclusive and Sustainable. These are worthy principles that any modern national statistics office should seek to embody to meet the challenges and, maximise the opportunities they face. 

259. The strategy is supplemented by additional publicly available documents including:

• ONS Business Plan and OSR Business plan, documents providing more detail on deliverables, timelines and resources.
• International Strategy. A five-year international strategy entitled ‘Statistics for the Global Good’ which details how the statistical system will engage and influence across the international context. 
• UKSA Annual Report and Accounts which are laid before Parliament each year, the most recent in July 2023. The OSR produces its own Annual Report which is submitted alongside. The Annual Reports reflect on the work conducted during the previous year, progress against strategic objectives and includes key information on the organisation’s finances, governance and adherence to functional standards.

260. The Review heard from users who felt that the UKSA strategy could benefit from more detail with one noting that they ‘believe that the strategy lacks sufficient detail on implementation’ and another observing that the current document did not allow for an empirical assessment of success or failure. Some also found it difficult to contribute to work planning as they are not well sighted on the process and how they could feed into it.

261. The ONS and OSR business plans that supplement the strategy with more detail on the organisation’s strategic objectives and milestones (against which progress is assessed in the annual report and accounts) were not identified by those engaging with the Review and are only mentioned in the final pages of the strategy. The UKSA should ensure that these documents are easily available to users and ensure that they contain sufficient detail for users to engage with and hold UKSA to account. 

262. In March 2023 the ONS hosted an event and panel session for statistics users to reflect on the work of the UKSA in implementing Statistics for the Public Good to date and look ahead to the remainder of the 5 year strategy. This style of engagement should continue to ensure that users can engage with and influence UKSA’s plans in addition to the Triennial Statistical Assembly proposed in Recommendation 1. 

Functional Standards

263. The Review has received evidence from the UKSA that it complies with Functional Standards and not identified any evidence to suggest that there is non-compliance.

Efficiency

264. The final section of the Review discusses the theme of Efficiency. As detailed in the terms of reference, the UKSA has already committed to achieving the efficiencies over the current spending review period and so consideration of further efficiencies is outside the scope of this review. However, the Review has examined evidence concerning possibilities to improve the productivity of the UKSA workforce, talent planning and retention and their adherence to existing financial guidelines. 

265. The Review has identified the following Recommendations:

Recommendation 18: The UKSA should develop a framework to follow when considering ad-hoc commissions for statistics in order to be open about the opportunity costs of such work. 

Recommendation 19: The UKSA should step up efforts to build partnerships outside of government, particularly with universities and think tanks, given the clear economic and social benefits to this collaboration.

Finances

266. Mindful of the Review Terms of Reference and ongoing engagement between the UKSA and HM Treasury, the Review has not sought to look at the detail of the organisation’s financial arrangements. This is due to the fact that in the Spending Review 2021 (SR21) settlement, UKSA agreed with HM Treasury to deliver cash releasing efficiencies worth 10 per cent of spending on ongoing baseline activities (or circa £21m). These efficiencies are being delivered primarily through restructuring support functions, streamlining business and social survey activities, more efficient use of the property estate and automation of repeated tasks and processes. Furthermore, HM Treasury provided UKSA additional programme funding at SR21 for investment in improving the quality and efficiency of statistical production through replacement of legacy IT systems and greater and more innovative use of large-scale administrative datasets.

267. The Chancellor subsequently announced a further Efficiency and Savings Review in November 2022. In response, UKSA made proposals to find additional efficiencies including through increased automation of statistical outputs via Reproducible Analytical Pipelines (RAPs), a revised commercial approach and an “online first” survey strategy.

268. The Review noted the unique role that the UKSA has in delivering the decennial census and the impact that has on its spending profile with significant funding increases and decreases on either side of the Census. UKSA received £453m in 2021-22 compared to £225m in 2022-23.

Prioritisation

269. Information from the UKSA noted that the Board oversees the delivery of organisational priorities set out in the business plan and receives regular updates from the Planning and Portfolio Management Team. This includes Monthly Integrated Performance Reports which set out delivery against the plan and a Quarterly Strategy Update. The prioritisation of ONS work is guided by a Prioritisation Framework alongside governance arrangements. As part of the wider work with Heads of Profession a Task and Finish group is seeking to better establish ways of working that guide the way that ONS and the GSS work together with collaboration and collective delivery as the default and optimise value for money from the statistical system as a whole.

270. Evidence to the Review observed that the ONS may sometimes be simply attempting to deliver too many outputs which poses risks to quality. The ONS’s most recent review of statistics and analysis using the Prioritisation Framework took place in Summer 2023. Some releases are being changed to provide data only (without complementary analysis) and others are becoming less frequent or stopping altogether. 

271. The Review understands that users have been informed, though not always consulted ahead of time, and that notices were included alongside the latest publication of affected statistical releases following the most recent prioritisation exercise. The update on the ONS website observes that changes will be made, however it does not provide any further detail on which publications will be affected. The UKSA should ensure that changes to ONS publications are accessible to all users on the website. Furthermore, ahead of any future prioritisation review, information should also be made available to users about how they can engage with the exercise. The recent consultation by the Northern Ireland Statistics and Research Agency (NISRA) on statistical prioritisation is a good example of this.      

Work for Other Government Departments

272. The ONS has capability and expertise that is often sought by other departments seeking analysis and insights. There are examples of ONS agreeing to deliver discretionary priority analysis and data collection for, and funded by, other Government departments.

273. This arrangement is not unique to the UK. Many other national statistics offices consulted by the Review also conduct separately funded work for different parts of their respective governments. Users of core statistical outputs put forward their view that such ad-hoc projects could direct resources away from the production of core, ‘business as usual’ outputs because ramping up and then down to deliver ad-hoc work is not easy. Where ad-hoc work may be required, opportunities for collaboration should be explored, including across the GSS, where Heads of Profession sometimes felt that their Department would have expertise to contribute or even lead on certain ad-hoc projects.  

274. The National Statistician noted that he is an advocate for meeting user needs and remarked that ‘Independence does not mean irrelevance’. This responsiveness to user needs from the ONS has been seen most clearly in recent years with its agile response to the Covid-19 pandemic. Whilst there is a Prioritisation Framework for overall work at UKSA, it was not clear to the Review that this was used when considering ad-hoc work requests. To mitigate potential risks identified in evidence to the Review, the UKSA should create such a framework to follow when considering ad-hoc work in order to be open about the opportunity costs of such work (Recommendation 18).  Such a framework should also draw upon the findings of the Statistical Assembly (detailed in Recommendation 1) and should address the following principles:

• That the UKSA should focus on delivering in areas where ONS capability and expertise bring unique value
• Work should not bring into question the UKSA’s independence
• The ONS’s teams that deliver ad-hoc work should be sufficiently resourced so as not to divert resource or capability away from core statistical outputs and priorities in a detrimental way
• The decisions should be sensitive to the position and expertise of statisticians in the broader GSS and should avoid duplication of work.

Working with partners outside of Government     

275. A significant amount of collaboration between the ONS and external organisations, most significantly universities, already exists. The longstanding relationship with the University of Southampton, for example, has delivered clear benefits to the UK statistical system. More recently the Economic Statistics Centre of Excellence, based at King’s College London, is an excellent example of the mutual benefits of fostering partnerships with academia.

276. There is scope to go further given the economic and social benefits of working with universities and think tanks, including two way interchanges or secondments, and this should be the focus not just of the ONS, but of the GSS and indeed the Government Analysis Function. This will help improve skills within government, extend the pipeline of talent available to the statistical system and bring broader cultural benefits associated with different ways of working and increasing diversification of the workforce (Recommendation 19).

ONS Estate

277. In addition to reviewing statistical priorities, the UKSA is identifying efficiencies in relation to its resources. Information from UKSA shows that their property estate includes offices in Newport, Wales and Titchfield, Hampshire and a paper records store in Christchurch. It also makes use of Government Hub Space (shared office locations for UK Civil Servants) across the UK (London, Darlington, Edinburgh and Manchester), and is making unoccupied space in its own offices available for use by others such as the NHS. According to information received from the UKSA, when considering sites across the UK factors including skills, diversity and location connectivity are all considered.

278. The Review recognises the benefits highlighted by UKSA that this approach to estates brings by increasing access to local employment markets outside of London which diversifies the staff pool and counters group think. It also increases opportunities for collaboration with the wider civil service for example the economic campus at Darlington. However, the UKSA must also ensure that it does not spread its workforce too thinly across the UK and thereby losing the benefits of collaboration and making it harder to manage the careers of specialists who are fragmented into differently-located small groups. In identifying hub locations it is important to deliver a workforce with the skills required for organisational priorities, and so partnership with local universities known for their statistics and data science research and training is an advantage.

Annex A - Structure of UK statistical system

Annex B - Stakeholder Engagement

This list of stakeholders that Professor Lievesley and the Review team have engaged with includes meetings with experts undertaken in a personal capacity rather than as representatives of their organisations. Alongside this, correspondence has also been received from various individuals.

As part of the field research Professor Lievesley and the Review team attended externally hosted conferences and events, including the 2023 World Statistics Congress in Ottawa, Canada, Public Service Data Live in London and events hosted by the Royal Statistical Society and Better Statistics, also in London. 

The Review team hosted round tables events on economic statistics and with local authorities. The attendees have been captured below. 

Within UKSA and the government more broadly Professor Lievesley attended multiple meetings, including a UKSA Board (in Darlington), a Regulation Committee, an ONS User Committee, a cross-government Director of Analysis meeting and a Heads of Statistics Profession (GSS) meeting. Internal roundtables were held with staff from ONS (in Newport and virtually), OSR, HM Treasury, Scottish Government (in Edinburgh) and Welsh Government (in Cardiff). 

Challenge panel:

• Professor Sir John Curtice
• Guy Goodwin 
• Professor Tim Leunig
• Professor Guy Nason
• Professor Rebecca Riley
• Professor Alice Sullivan
• Dr. Ben Warner 

Royal Statistical Society round table:

• Professor Paul Allin
• Simon Briscoe
• Siobhan Carey
• Dr Andy Garrett
• Peter Gordon
• Roger Halliday
• Mike Hughes
• Deana Leadbeter
• Rachel Leeser
• Jill Leyland
• Anna Powell-Smith
• Geoff Tily
• Dev Virdee

Local Authority round table:

• Vivienne Avery, Greater London Assembly 
• Nick Cassidy, Improvement Service Scotland 
• Jay Gamble, Liverpool City Region
• Jules Ient, Cambridgeshire and Peterborough Combined Authority
• Salman Klar, Richmond and Wandsworth
• Sichun Lam, West Midlands Combined Authority
• Thomas Peggs, Tees Valley Combined Authority 
• Christopher Pope, Greater Manchester Combined Authority 
• Juliet Whitworth, Local Government Association

Economics roundtable:

• Mike Brewer, Resolution Foundation 
• Professor Huw Dixon, University of Cardiff
• Andre Moreira, Bank of England
• Professor Rebecca Riley, King’s College London
• Professor Mairi Spowage, Strathclyde University

Interviews

UKSA Board:

• Sir Robert Chote (Chair)
• Professor Sir Ian Diamond (National Statistician)
• Ed Humpherson (Director General, Office for Statistics Regulation)
• Sian Jones (Deputy Chair)
• Dr Jacob Abboud
• Professor Sir John Aston
• Nora Nanayakkara
• Professor Dame Carol Propper
• Professor Sir David Spiegelhalter
• Penny Young

ONS:

• Rob Bumpstead
• Mary Gregory
• Sarah Henry 
• Sally Jones
• Mike Keoghan
• Liz McKeown
• Darren Morgan
• Sarah Moore 
• Alison Pritchard
• Osama Rahman
• Emma Rourke
• Ed Sheerman
• Tom Taylor

Government Statistical Service:

• Janet Egdell, National Records Scotland
• Steve Ellerd-Elliot, Department for Work and Pensions
• Scott Heald, Public Health Scotland
• Stephanie Howarth, Chief Statistician, Wales
• Ally McAlpine, Chief Statistician, Scotland  
• Neil McIvor, Department for Education 
• Paula McLeod, Foreign, Commonwealth and Development Office 
• Anita Morrison, Scottish Government 
• Tom Orford, HM Treasury
• Sian Rasdale, Foreign, Commonwealth and Development Office 
• Jennifer Rubin, Home Office
• Tom Smith, Department for Levelling Up, Housing and Communities
• Philip Wales, Chief Statistician, Northern Ireland
• Neil White, Scottish Government

Government stakeholders:

• Sue Bateman, Central Digital and Data Office
• Sam Beckett, HM Treasury
• Scott Bowman, Office of Budget Responsibility
• Sir Alex Chisholm, Cabinet Office
• Esta Clark, National Records Scotland 
• John Edwards, Information Commissioner
• Steve Farrington, Office of Budget Responsibility 
• David Foley, Cabinet Office 
• Roger Halliday, Research Data Scotland
• Tom Hemingway, HM Treasury
• Geoff Huggins, Scottish Government 
• Sir Richard Hughes, Office of Budget Responsibility 
• Dr Laura Gilbert, No 10
• Steffan Jones, Cabinet Office
• Sean Whellams, HM Revenue and Customs
• Professor Sir Chris Whitty, Department for Health and Social Care 
• Jonathon Wroth-Smith, National Records Scotland

National Statistical Institutes

• Anil Arora, Statistics Canada
• Jennifer Banim, Central Statistics Office, Ireland
• Joe de Beer, Statistics South Africa
• Barteld Braaksma, National Statistical Institute, The Netherlands
• Padraig Dalton, Central Statistics Office, Ireland
• Gary Dunnet, Statistics New Zealand
• David Gruen, Australia Bureau of Statistics
• Anders Holmberg, Australia Bureau of Statistics
• Lilli Japec, Statistics Sweden
• Francesca Kay, Central Statistics Office, Ireland
• Professor Shigeru Kawasaki, formerly Statistics Bureau of Japan
• Olivier Lefebvre, National Institute of Statistics and Economic Studies (INSEE), France
• Eric Rancourt, Statistics Canada
• Mikael Schöllin, Statistics Sweden
• Mark Sowden, Statistics New Zealand

Other:

• Professor Paul Allin, Imperial College London
• Professor Sir Charlie Bean, author Independent review of UK economic statistics (2016)
• David Bharier, British Chamber of Commerce
• Professor Sheila Bird, University of Cambridge 
• Maddy Bishop, Institute for Government 
• Simon Briscoe, Data Analysis Bureau
• Siobhan Carey, former Chief Statistician, Northern Ireland
• Oliver Chinganya, Africa Centre for Statistics 
• Len Cook CBE CRSNZ, former UK National Statistician
• Les Commons, Market News
• Professor Dame Diane Coyle, economist
• Robert Cuffe, BBC
• Dr Sarah Cumbers, Royal Statistical Society
• Dr Kevin Daly OBE, Goldman Sachs
• Tony Dent, Better Statistics
• Sir Andrew Dilnot, former Chair, UKSA
• Keith Dugmore, (retired) Demographic Decisions
• James Ebdon, Bank of England 
• Dr Ivan Fellegi, former Chief Statistician, Canada
• Dr Rita Gardner, Academy of Social Sciences
• Dr Andy Garrett, Royal Statistical Society
• Dr Ben Goldacre OBE, University of Oxford
• Emma Gordon, Economic and Social Research Council
• Tim Harford, BBC
• Dr Carlos Jarque, former President, National Institute of Statistics, Geography and Informatics (INEGI) Mexico
• Paul Johnson CBE, Institute of Fiscal Studies
• Simon Kirby, Bank of England 
• Sue Linacre, statistics consultant (formerly ABS and ONS)
• Phyllis Macfarlane, Better Statistics 
• Ciaran Martin CB, Blavatnik School of Government, University of Oxford
• Dame Jil Matheson, former UK National Statistician
• Will Moy, Campbell Collaboration 
• Lord O’Donnell (Gus), former Cabinet Secretary
• Dr Mark Orkin, former Statistician General, South Africa
• Professor Alison Park, Economic and Social Research Council
• Stephen Penneck, International Statistics Institute 
• Adam Phillips, Archive of Market and Social Research
• Huw Pill, Bank of England
• John Pullinger CB, former UK National Statistician
• Hetan Shah, British Academy
• Matthew Shearing, statistical consultant
• Sir Bernard Silverman FRS, statistician
• Will Snell, Fairness Foundation 
• Alex Steer, VML
• Geoff Tily, Trade Union Congress
• Gemma Tetlow, Institute for Government 
• Dennis Trewin AO, former Australian Statistician
• Walter Radermacher, former President of the Federal Statistical Office (Germany) and Director General of Eurostat
• Jordan Urban, Institute for Government 
• Ron Wasserstein, American Statistical Association 
• Rob Watson, BBC
• Professor Martin Weale CBE, National Statistician’s Committee for Advice on Standards for Economic Statistics
• Stian Westlake, Economic and Social Research Council
• Dr Michael Wolfson, Carleton University, Ottawa
• William Wragg, Chair, Public Administration and Constitutional Affairs Committee

Call for Evidence

The Review team received written evidence from the following:

• Richard Alldritt
• Professor Brian Bell, Migration Advisory Committee
• Tony Dent, Better Statistics 
• Steve Ellerd-Elliot, Department for Work and Pensions 
• Dr Jonathan Everett, Royal Statistical Society 
• Maya Forstater, Sex Matters 
• Stephanie Howarth, Welsh Government 
• Professor Yang Hu, Lancaster University 
• Thomas King
• Jill Burges Leyland
• Ally McAlpine, Scottish Government
• Fiona McAnena, Fair Play for Women 
• Kate McHugh, York and North Yorkshire Local Enterprise Partnership
• medConfidential
• MurrayBlackburnMackenzie
• Keith Miller
• Will Moy, Full Fact 
• Jane Naylor, Department for Business and Trade 
• Victoria Obudulu, UK Health Security Agency 
• Ordnance Survey
• Sir Bernard Silverman
• Professor Alice Sullivan, University College London 
• Sandra Tudor, Department for Levelling Up, Housing and Communities 
• Stian Westlake, Economic and Social Research Council 
• John Wildman
• William Wragg MP, Public Administration and Constitutional Affairs Committee

Annex C - Terms of Reference

Background

1. The Public Bodies Review Programme delivers on commitments made in the Declaration on Government Reform to increase the effectiveness and efficiency of public bodies. Reviews of public bodies assure the public and ministers that arm’s length bodies (ALBs) remain useful and necessary. Reviews provide independent challenges to ensure that ALBs are effectively delivering services that meet the needs of citizens; that they are doing so as efficiently as possible; and that they have clear governance and lines of accountability.

2. The UK Statistics Authority (UKSA) was established by the Statistics and Registration Services Act 2007 with a statutory objective to promote and safeguard the production and publication of official statistics that ‘serve the public good’. The main part of the UKSA is the Office for National Statistics (ONS) which is the UK’s largest producer of official statistics on a range of key economic, social and demographic topics. The National Statistician acts as the Permanent Secretary and Principal Accounting Officer.

3. The UKSA also contains a regulatory function called the Office of Statistics Regulation (OSR) which aims to increase public confidence in the trustworthiness, quality and value of statistics.

4. The last review of the UKSA/ONS was the 2016 ‘Independent review of UK economic statistics’ so it is right that an independent assessment is now made of how the organisation is performing, to provide assurance and offer fresh challenges.

Scope

5. The Review is organised according to the quadrants identified by the Public Bodies Review Programme - efficacy, governance, accountability and efficiency. The Review will be conducted according to the published Guidance on the undertaking of Reviews of Public Bodies which should be read in conjunction with these Terms of Reference.

6. UKSA has recently committed to significant efficiencies as part of the Spending Review process. This Review will not duplicate recent efficiency reviews or seek to identify significant savings on top of the existing efficiencies that UKSA is already implementing.

7. The following themes indicate the key issues the Review should cover. The Lead Reviewer will have discretion to consider other issues that arise during the course of the Review.

Theme 1: Efficacy

8. The Review will focus on the outcomes for users of the Authority’s work. It will explore whether the UKSA is producing statistics that respond to genuine user need, for example producing high quality, respected official statistics to support the government to manage the economy, drive growth, productivity and Government efficiency, and informing international trade policy, given its underpinning remit is to promote statistics for the public good.

9. The Review will cover the public bodies review requirements, including:

• Whether the UKSA and its constituent parts have clear remits and mandates and remain relevant; whether the delivery model is correct to deliver effective outcomes; and whether the UKSA has the correct systems and knowledge in place, including adherence to the Public Sector Equality Duty.
• Whether the UKSA has clear, measurable performance indicators.

10. In addition to the public bodies review requirements, the Review should consider:

• The extent to which UKSA produces high quality, respected official statistics to support the government to manage the economy, drive growth, productivity and Government efficiency, and inform international trade policy.
• The extent to which government needs for official statistics are being understood, anticipated, prioritised, selected and met by UKSA and how this compares with wider user prioritisation. Whether the form and functions of UKSA are optimised to ensure the right balance between quality and value for money.
• The way UKSA communicates statistics to users, and the balance of emphasis put on production of statistics versus providing analysis and interpretation.
• The extent to which ONS is taking full advantage of developments in data science capabilities and techniques to support improvements to quality and efficiency, across a range of its statistics that can make greater use of new forms of data.

Theme 2: Governance

11. The Review will consider the governance arrangements including the adequacy of the independence of UKSA and whether there is any conflict between the production and regulation functions it performs. The Review will also consider the UKSA approach to UK-wide data, board effectiveness reviews and board diversity.

12. The Review will cover the public bodies review requirements, including:

• Whether the UKSA board has a clearly articulated purpose, drives efficiency and effectiveness and supports the delivery of the government’s wider objectives.
• The processes by which UKSA board members are appointed, whether the UKSA board has the correct balance of skills and experience, and the procedures to ensure board members work to the highest standards.
• The arrangements in place for finance, risk management and internal control.
• Whether the UKSA board is open, transparent, accountable and responsive in its communications with the public and other key stakeholders.

13. In addition to the public bodies review requirements, the Review should consider:

• Whether the UKSA is operating at an appropriate ‘length of arm’ to ensure the right balance between alignment with government priorities and the need for technical expertise and impartiality.
• The roles of the ONS, the OSR, how they work with the UKSA Board, whether there is any conflict between the production and regulation functions, and how all parties maintain the necessary degree of separation required. This section should consider whether the regulatory regime requires reform to deliver public confidence in statistics and whether the current legislative structure for the UKSA is fit for purpose.
• How the UKSA works in relation to supporting policy development across government. This section should include how the UKSA manages its role with ministers and ministerial department officials, and the governance of decisions about what emerging topics to prioritise.
• How well the UKSA performs its wider role with respect to the whole UK official statistics system, including relationships with the devolved administrations and the effectiveness of efforts by UKSA to improve the coherence of UK-wide data.
• Whether the arrangements put in place after leaving the EU are appropriate for enabling the UKSA to assess if UK statistics are internationally comparable, with a particular focus on fiscal and economic statistics produced by ONS.
• Whether the UKSA has arrangements in place for identifying changes in statistical reporting after leaving the EU. This could include whether any statistical measures inherited from Eurostat can be improved or removed to better suit the circumstances of the UK, consistent with the highest statistical standards.
• The system of pre-release access to statistics and the impact of changes to the regime introduced by the National Statistician in 2017, including whether it had the desired impact on public and market perceptions.

Theme 3: Accountability

14. The Review will focus on the adequacy of accountability arrangements with the Cabinet Office and whether ministers have enough assurance to meet their duties as set out in Managing Public Money.

15. The Review will cover the public bodies review requirements, including:

• Whether there is an effective and appropriate relationship between the UKSA and the Cabinet Office as the sponsor department, bearing in mind the status of the UKSA as a non ministerial government department, the requirements in Managing Public Money, and the role of the Minister for the Cabinet Office.
• How the accountability of UKSA to Parliament is understood and whether it is effective. This section should include how UKSA works with PACAC and whether the committee is satisfied with the relationship.
• Compliance against the relevant functional standards, for example, digital, data and technology and the analysis function.

Theme 4: Efficiency

16. The government has asked all ALBs to consider additional efficiencies contributing to an overall reduction in spending, in the context of existing settlements and the 2022 Autumn Statement.

17. The UKSA has already committed to achieving efficiencies over the current spending review period, including efficiencies worth 10% of its baseline budget in 2024/25. As part of the Efficiency and Savings review that was announced at Autumn Statement 2022, the UKSA are looking to deliver further efficiencies in 2023/24 and 2024/25. Therefore, this Review will not seek to identify further efficiencies on top of the existing plans for the spending review period but will assess if the UKSA’s plan to deliver the current commitments is feasible and any opportunities for longer term efficiencies. This section of the Review will also cover the following public body review requirements:

• Adherence to existing financial guidelines including Managing Public Money, the Financial Reporting Manual and Cabinet Office Spend Controls.
• Possibilities to both improve the productivity of the UKSA’s workforce.
• How future-proofed the UKSA is in its plans to develop or attract a cadre of senior civil servants, ensuring an appropriate balance between turnover in senior appointments to ensure groupthink does not develop, and the need to retain deep technical expertise at senior levels.

Annex D

The Review team supporting Professor Lievesley:

• Thalia Baldwin (Senior Responsible Officer)
• Neil Warsop
• Joseph Moore
• Alex O’Donoghue

Glossary

Acronym	Meaning
10DS	No10 Data Science
AI	Artificial Intelligence
ALB	Arm’s length body
API	Application Programming Interface
BBC	British Broadcasting Corporation
CADE	Centre for Applied Data Ethics
CBS	Centraal Bureau voor de Statistiek (Statistics Netherlands)
CDDO	Central Digital and Data Office
CIS	Coronavirus Infection Survey
DA	Devolved Administrations
DDaT	Data, Digital and Technology
EFTA	European Free Trade Association
EU	European Union
FTE	Full-time equivalent
GDP	Gross Domestic Product
GSG	Government Statistician Group
GSS	Governmental Statistical Service
GSSIC	GSS International Committee
HMRC	HM Revenue and Customs
HMT	HM Treasury
IAC	Inter Administration Committee
IDS	Integrated Data Service
IMF	International Monetary Fund
LFS	Labour Force Survey
LLM	Large Language Models
NSDEC	National Statistician’s Data Ethics Advisory Committee
NSEG	National Statistics Executive Group
NSO	National Statistical Offices
OBR	Office for Budget Responsibility
OECD	The Organization for Economic Cooperation and Development
ONS	Office for National Statistics
OSR	Office for Statistics Regulation
PACAC	Public Administration and Constitutional Affairs Select Committee
PASC	Public Administration Select Committee
RSS	Royal Statistical Society
SAIL	Secure Anonymised Information Linkage
UK	United Kingdom
UKRI	UK Research and Innovation
UKSA	UK Statistics Authority
UN	United Nations

1. As at 31 March 2023. 2023-24-strategic business plan ↩

2. Fund principles. ↩

3. Resolution adopted by the General Assembly on 29 January 2014 (PDF, 37KB) ↩

4. Browse our statistical bulletins by A to Z ↩

5. Independent Review of UK Economic Statistics Professor Sir Charles Bean (PDF, 5.1MB) ↩

6. Code of Practice for Statistics ↩

7. Declaration on Professional Ethics ↩

8. The Tyranny of Numbers by David Boyle ↩

9. This role is referred to as Head of Assessment in the Statistics and Registration Service Act 2007. ↩

10. Public Bodies’ Review guidance ↩

11. Resolution adopted by the General Assembly on 29 January 2014 (PDF, 37KB) ↩

12. Pre-release access to Deaths registered weekly in England and Wales, provisional ↩

13. National Performance Framework ↩

14. SAIL Databank ↩

15. response report outputs consultation ↩

16. Expert report for the UK Covid-19 Public Enquiry (PDF, 10.4MB) ↩

17. Governance of the statistical system ↩

18. Revisions of estimates of UK Gross Domestic Product (GDP) ↩

19. User personas ↩

20. Pre-appointment Hearing: Chair of the UK Statistics Authority, HC 1162 (PDF, 175KB) ↩

21. Chair of the UK Statistics Authority, Sir David Norgrove (PDF, 219KB) ↩

22. Government transparency and accountability during Covid 19: The data underpinning decisions (PDF, 811KB) ↩