Governance of official statistics: redefining the dual role of the UK Statistics Authority; and re-evaluating the Statistics and Registration Service Act 2007 Contents

Conclusions and recommendations

Use of statistics

1.We agree with the evidence we received that those producing official statistics do not understand all of today’s users and potential users of statistics and how statistics are used. It is surprising that UKSA and the GSS more generally seem not to have carried out research into users such as that produced by the Statistics Commission in 2007 or actively followed up the main conclusions of that report. Producers are, therefore, not able to close statistical gaps or appropriately refine the presentation of existing statistics so they are more useful. Similarly, with only a modest sense of how the public uses data and no evidence of the unmet needs, UKSA is not delivering public good as required under the legislation. We recommend that UKSA should lead cross-government research to build an evidence base of how statistics are used in practice, taking into account the full breadth of stakeholders (not just users) and to establish where data gaps persist. We also recommend that UKSA should conduct sector by sector reviews, to understand what stakeholders need or want, and to make statistics more relevant. (Paragraph 22)

2.UKSA needs to take a more strategic approach to engaging with users. The government statistical system’s engagement with users is an issue that still has not been effectively addressed, despite being highlighted many times and even in the Office for Statistics Regulation’s (OSR) own assessment reports. Some mechanisms exist to engage users and there are pockets of good practice, but the evidence suggests there is a real need to improve overall use and usefulness of official statistics in order to deliver the public good as defined by the Act. UKSA should develop a strategy for engaging all users of official statistics that works for the Office for National Statistics (ONS) and for other producers across government over whom it does not have direct control. It must, at a minimum, report annually on progress against that strategy. Departments that do not make sufficient and rapid progress on engagement with users of statistics must expect to be held accountable in Parliament. UKSA should develop a strategy for engaging all users of official statistics that works for the Office for National Statistics (ONS) and for other producers across government over whom it does not have direct control. It must, at a minimum, report annually on progress against that strategy. Departments that do not make sufficient and rapid progress on engagement with users of statistics must expect to be held accountable in Parliament. (Paragraph 31)

3.The National Statistician should make every effort to encourage individuals at all levels in ONS and the Government Statistical Service (GSS) to actively engaging with users so that the user perspective is central. Above all, UKSA must set the best possible example of engagement with users of statistics, as part of its overall governance of the statistics system, in order to increase public confidence. Government statisticians must also work with the media to ensure that the correct interpretation of the statistics and trends is found in their outputs. (Paragraph 32)

4.Official statistics must be presented in a way that is easy for non-experts to use. We heard evidence of improvements to the ONS website in recent years but it is clear that there is much more to be done to make ONS data accessible. We recommend that UKSA conducts a review of how official statistics are published and presented and develops a plan to address any weaknesses found. We also recommend that UKSA conducts a programme of user research to see how easy it really is to find the basic core statistical series. (Paragraph 43)

5.UKSA has not yet done enough to bring important data together in one place, giving users the big picture across a domain, as used to happen in, for example, the now-defunct annual Social Trends and monthly Economic Trends and Financial Statistics publications. We recommend that UKSA resumes publication of compendia, albeit in a modern format consistent with accessing data via the internet, that allow related data to be found and are not limited to ONS statistics. (Paragraph 44)

6.There is also evidence that gaps in data of many shapes and forms have arisen for different reasons, with the result that users are not getting data that is relevant to them. We recommend that UKSA takes note of the Institute for Government report on data gaps, of other third-party reports highlighting gaps in data and of the user consultation recommended above. It should identify the most significant data gaps and use its influence across the public sector, and by engaging with Parliament and international organisations, to set out a plan to close them. (Paragraph 45)

7.GSS and ONS statisticians should also do more to guide users how to use their statistics, explain how they are typically used, outlining their strengths and weaknesses, providing commentary and advice. (Paragraph 46)

Governance of official statistics

8.Opinion is divided on the effectiveness of the current UK statistical system. The system is considered by some to be better than what went before; and survey data suggests it has strengthened public trust. The UK system is also well regarded internationally. We have heard evidence that the Code of Practice has improved the quality of statistics but there has been criticism that its detailed application has distracted UKSA from big strategic issues. (Paragraph 66)

9.UKSA has made some tough public interventions that were unpopular with government but has been too slow to act on serious statistical and methodological issues such as the classification of items on the national balance sheet and has seemed to have acted only when under pressure from others. We praise ONS for ending the pre-release of ONS statistics to Ministers and the Bank of England, but the practice of pre-release of data across the GSS remains. While we recognise the need for a safe space for UKSA to discuss complex and important issues with Government, UKSA has not been transparent about the meetings it has with Government and significant stakeholders and the purpose of those meetings, for example, in the way policy towards RPI was developed. (Paragraph 67)

10.UKSA has been too reluctant to make recommendations if it does not think they will be acted on. It has not used its powers to lay reports before Parliament where it disagrees with a department. It prefers to take a more collaborative approach. UKSA’s most significant failure has been in its handling of RPI, which we discuss further in Chapter 5. (Paragraph 68)

11.We recommend that UKSA should demonstrate more proactive, quicker responses to concerns about the accuracy and misuse of statistics and should more clearly demonstrate its independence from key stakeholders, when it has significant disagreements with producers of statistics. UKSA should engage with Parliament, including using its powers to lay reports before Parliament to highlight concerns about statistical practice. UKSA should make recommendations to Government, regardless of whether it considers they will be taken up. UKSA should be transparent about meetings with external bodies, such as with the Treasury and the Bank of England, and about the purpose of those meetings. The Chair of UKSA should make more strident efforts to encourage Ministers to end the practice of pre-release access to data across the public sector by referring departments that continue with the practice to Parliament. (Paragraph 69)

12.The legislation presents significant challenges to UKSA in exercising the dual functions of production and regulation of statistics and does not give clarity to those roles. The Committee believes that combining regulation and production in a single organisation was misconceived since it placed a higher value on avoiding public disagreements than improving statistics for the public good. UKSA, set up as a non-ministerial department, does not have the independence, for example, of an independent regulator, such as a parliamentary body. There is a risk of direct interference from government departments and other stakeholders or indirect interference as they seek outcomes that favour themselves rather than the public good as defined by the Act. (Paragraph 73)

13.The Committee believes that the dual role has created a conflict and recommends that the legislation should be changed to create a separate and independent Office for Statistics Regulation: guardian of government-produced numbers, reporting to Parliament, regulating a central statistical institute and the disaggregated producers of statistics across government. This would also provide the opportunity to tidy up other parts of the legislation and align UK law better with best international practice. (Paragraph 74)

14.We recognise that, in the current climate, legislation is unlikely to be revisited within the next three years and so have made recommendations throughout this report to promote improvements without the need for legislation. We believe a lack of legislative change should not impede necessary change in UKSA and therefore we recommend that UKSA take steps, starting immediately, to improve and demonstrate its independence and to minimise the vulnerabilities of having production and regulation within one body. We call on UKSA to report to this Committee each year on progress with the implementation of the recommendations presented in this report. If early improvement is not forthcoming, we will press for time to be found to replace or improve the legislation. (Paragraph 75)

15.It is welcome that UKSA has sought to distinguish OSR from the rest of UKSA and has expanded its capability. However, with only 29 staff, it is under-resourced when set against the Office for National Statistics with over 3000 staff, and the thousands of GSS statisticians working across the system that OSR is required to regulate. We do not believe that OSR can effectively regulate the UK’s statistical system with the level of resources it has. This is evidenced in the lack of action in key areas like RPI and student loans, and the reliance that OSR has on departments presenting their mistakes and failings to OSR, rather than OSR actively seeking problems. The Committee concludes that OSR is not being sufficiently ambitious in its role and should be more proactive in safeguarding statistics for the public good. We believe that UKSA must do more to promote a proactive, ambitious and better resourced OSR, with a higher profile and ability to act as a champion of good quality and useful, independently produced, statistics. We therefore recommend that OSR sets out what more it could achieve if it doubled or trebled in size. We recommend that the non-executive members of the UKSA Board, who comprise the Regulation Committee, should have a special responsibility to examine the budget of OSR and the balance of resourcing between OSR and ONS. (Paragraph 85)

16.Despite the creation of the Office for Statistics Regulation, the Committee believes that the UKSA Board has still not sufficiently separated its dual functions. Staff responsible for production and regulation of statistics remain co-located in the same office space. The National Statistician is Chief Executive, Accounting Officer and Permanent Secretary of UKSA and nominal Head of the GSS. We believe this makes it unclear to external observers whether OSR staff are independent and separate from these reporting lines. OSR can only claim independence if it demonstrates more clearly that it is operating independently of ONS. This needs active support from the non-executives on the Board. The Committee therefore reiterates previous recommendations that UKSA should take more concrete and visible steps to separate its roles of producer and regulator. The only way to secure public confidence is to ensure that OSR has a separate role and is independent. We recommend that OSR relocates its offices to separate premises from ONS. (Paragraph 86)

17.OSR has produced two annual reports since it was established in November 2016. The Committee recommends that OSR produces its own annual report every year. (Paragraph 87)

18.We believe that the accountability structures for OSR need to be more transparent. We therefore recommend that UKSA takes steps to ensure its non-executive members, sitting on the Regulation Committee, operate more transparently by, for example, announcing when meetings will be held, and publishing agendas and minutes as soon as possible afterwards. (Paragraph 88)

Governance of UKSA

19.The Committee agrees with witnesses who highlighted the lack of appropriate oversight by the UKSA Board in dealing with RPI. This was an example of the UKSA Board being too passive and not taking sufficient action to protect its reputation as an independent regulator. The fact that there was a need for the Treasury to commission its own external review of economic statistics on ONS highlighted limited foresight and engagement from the UKSA Board. (Paragraph 94)

20.It is clear that the Board should do more to demonstrate how it is holding the statistical system to account, addressing the big issues and doing its best to ensure that official statistics serve the public good, as defined in the Act. The Board’s lack of action on RPI, student loans and economic statistics, raises concerns about how issues are being managed in other areas. Even when mistakes were highlighted by external observers, the actions of the UKSA Board have often been slow or non-existent. We recommend that UKSA reports, for example through Board minutes, on how non-executive members provide accountability and challenge to the executive staff of UKSA and statistics producers in other bodies. We are told that meetings of non-executives take place but there are no publicly available agendas or minutes for those meetings. We recommend that UKSA reports publicly, using agendas and minutes, all committee meetings of non-executives. We also recommend that when new non-executives are recruited, it is with the expectation of working more hours than they currently do, so that they can give more time and attention to their role. (Paragraph 95)

21.Scrutiny of UKSA by Parliament is important and the results of that scrutiny are something that the UKSA Board should take seriously and act upon. Our evidence underlines that UKSA has not been sufficiently quick, thorough or transparent about its progress in responding to PASC recommendations. Nor has UKSA responded as readily as hoped, to requests from PACAC. (Paragraph 100)

22.The Committee will call UKSA for an annual hearing to be held after the publication of the annual report of UKSA. In addition to the UKSA annual report, the Committee requests the following documents to consider at that session: an annual report of OSR agreed by the non-executive members (through the Regulation Committee); a report identifying the most significant issues facing the statistical community, including a plan to address them; and an update on UKSA’s progress against recommendations from this Committee. We also ask the UKSA Chair to write to the Committee on a quarterly basis to update it on progress against Committee recommendations and results achieved, and on any other major statistical issues. (Paragraph 101)

23.We recommend that the Director General for Regulation keep the Committee informed of OSR actions by copying correspondence relating to interventions in the statistical system. In addition, we recommend that OSR write to Departmental Select Committees at least annually to update them on the data and statistical issues that OSR has identified in their remit. (Paragraph 102)

24.The Bean Review helped UKSA to secure additional resources and we welcome initiatives such as the Data Science Campus and Economic Statistics Centre of Excellence. However, we do not think UKSA should have to be subjected to an external review for it to secure what it needs. If the UKSA Chair, National Statistician and Head of Assessment were fulfilling their roles, they would be less dependent on external reviews. This is why we have called for an increase in OSR funding and a more dynamic set of actions by the UKSA Chair. However, we question whether UKSA has done enough to communicate its response to external reviews in the medium to longer term. It still remains unclear whether all the recommendations from external reviews have been accepted and what changes have been implemented. UKSA should report annually on its progress in implementing the recommendations of external reviews, including Parliamentary Select Committee inquiries, after its initial timely response. If it decides not to implement a recommendation, it should publicly explain the reasons for that decision. (Paragraph 107)

25.Evidence to this inquiry suggests that UKSA is not as transparent about its decision making as it should be. For example, publication of Board minutes has not been regular and timely and has not provided sufficient detail. We recommend that UKSA should publish its Board meeting agendas and papers, and its minutes as soon as they are approved. Minutes should be more informative about the nature of the questions under discussion. (Paragraph 113)

26.Evidence to this inquiry was unequivocal: the governance of the statistical system is unclear, even to those who work closely with it. This reflects a fundamental shortcoming of the Statistics and Registration Service Act 2007: that the term “Board” can mean different things, depending on whether it relates to its production, or assessment functions. UKSA is not to blame for confusion arising from the Act, but it should address it. UKSA should produce a framework document which sets out clear roles and responsibilities of the different parts of the statistical system, including the different elements of the UKSA Board, ONS, OSR, GSS, and departmental Heads of Profession to whom UKSA intends each of these parties to be accountable. A clear description of the framework should bring clarity where there is ambiguity in the legislation and should strengthen public confidence in the accountability of these bodies. (Paragraph 117)

RPI–a case study

27.The evidence about the mishandling of RPI, the influence that HM Treasury and the Bank of England have exercised, and the reluctance of UKSA to confront them, raises the most serious questions about the effectiveness of the governance of statistics, which is the primary responsibility of UKSA as a regulator. UKSA has allowed what was originally a simple mistake in price collection of inflation data to snowball into a major unresolved issue for a decade. It is clear that UKSA does not exercise consistent governance to ensure that the public can have confidence that it is safeguarding the measurement of inflation - either in the production of the numbers or the assessment and regulation of them, despite the production of RPI being primarily a duty prescribed by the Act. (Paragraph 132)

28.The Committee acknowledges that the legislation gives UKSA a dual role which is confusing but agrees with the conclusion of the House of Lords Economic Affairs Committee that “In publishing an index which it admits is flawed but refuses to fix, the Authority could be accused of failing in its statutory duties”. We are concerned that UKSA’s response to the report, promised but still not delivered, has been held up by its deference to the Treasury. This demonstrates a continued lack of independence on the part of UKSA from the influence of the Treasury. (Paragraph 133)

29.The handling of RPI resonates with other findings in this report, that the governance of statistics provided under the present statutory framework is inadequate. UKSA commissioned reports but did not follow through with recommendations and failed to resolve the various conflicts, demonstrating the inherent shortcomings of the statutory arrangements that have established a body with the twin roles of producer and regulator. We believe that in this case, UKSA has not fulfilled its responsibility to ensure the quality of statistics, but has become drawn into concerns about the consequences, which should be left to the Treasury and others. (Paragraph 134)

30.We recommend that UKSA publishes its recommendation for addressing the shortcomings in RPI immediately. This will enable Parliament to hold the Government to account for the mishandling of RPI and provide Parliament with an opportunity potentially to support UKSA’s recommendation. (Paragraph 135)

Governance of the statistical system

31.John Pullinger has demonstrated strong leadership of the Office for National Statistics over the last five years. The next National Statistician is a vital appointment to lead ONS and the GSS in developing a culture of listening to users and innovating in order to secure a statistical system that meets the current and future needs of the country. Ambitious leadership is needed to unleash the power of data to transform government and bring widespread benefits to society. User focus and innovation need to permeate all levels of the statistical system. We agree with the Royal Statistical Society that there is a systemic capability challenge to combine the technical skills required to produce numbers with skills in listening, communicating, influencing and innovating. (Paragraph 140)

32.The Committee is concerned that UKSA has not been successful in identifying a new National Statistician. This job requires a rare combination of skills, including a high level of understanding of statistics along with leadership and management ability. We recommend that UKSA should consider how to separate elements of the National Statistician’s role to make it possible to find suitable candidates. We recommend that UKSA produces a note for PACAC setting out a strategic approach to succession planning and how it is developing a sufficient talent pool to fill senior roles, including that of National Statistician, in the future. (Paragraph 141)

33.The Committee agrees with John Pullinger that having professional statisticians in large numbers across government, following a common code of practice, is a strength of the UK system. However, this disaggregated system needs to be able to maintain its independence from inevitable pressure from Ministers and senior officials about the way any data is presented. This has been made more difficult by the decline in seniority of statistical Heads of Profession in departments and the lack of influence that UKSA has over the GSS. (Paragraph 151)

34.The UK Statistics Authority should strengthen the governance of the wider statistical system. It should:

a)bolster the statistical Heads of Profession in departments by recommending they are given a more appropriate level of seniority in their departments;

b)report on the effectiveness of senior ONS roles with system-wide responsibilities (e.g. Head of Communications) in supporting the independence of GSS personnel across government;

c)publish a programme of work to tackle the big statistical issues across the GSS;

d)consult with users to invite external challenge and contributions to that programme; and

e)appoint a Board member to be responsible for monitoring and encouraging the independence of GSS personnel in departments.

35.We recognise that UKSA has made good investments in technology, including changes to its website, and the platform for the next census, but there is much more to be done. Innovation is key. The Data Science Campus has been recognised for making good progress, but UKSA should ensure that the outcomes of the Campus’s work are more transformative in ONS and across the whole of government. We have heard that some progress is being made in making use of administrative data, for example by use of tax data in analysis of changes to the economy, but this has been slower than hoped. (Paragraph 161)

36.UKSA should be requesting data sets so that it can take a lead in using the potential of data to transform government, at the same time demonstrating how the public will be protected from misuse of data. Statisticians should also anticipate the future needs of decision-makers, so that UKSA can shape the statistical system to meet those needs. (Paragraph 162)

37.We recommend that UKSA takes a stronger leading role across technology, data science, data ethics and influencing improved sharing of data which recognises UKSA’s role in the governance of the whole system. We would like to see UKSA highlight to Parliament where public bodies are not sharing data in the spirit of the Digital Economy Act. UKSA must make efforts to ensure that innovation in technology and data science, for example through the Data Science Campus, impacts the whole statistical system. UKSA must sufficiently resource continued activities to bring about further innovation with data. We call on UKSA to report to this Committee on how much it spends on the Data Science Campus and what it could achieve with varying levels of additional funding. (Paragraph 163)

38.UKSA has the opportunity, in developing its strategy for the next five years, to address the main points that have been raised in this report. The Committee believes that the UKSA strategy for the next five years must focus on UKSA’s role in the governance of the statistical system so that it seeks the public good and supports good statistical practice. OSR must be resourced to provide consistent regulation across the entire GSS and must have clearer separation from ONS; the UKSA Board must make this happen. The strategy must be more outward looking, providing the mechanics of how UKSA will listen to and be responsive to all types of users and engage with others across the public sector working with data. It must also be more forward looking, with a comprehensive programme of work to address known strategic issues and a process for identifying and anticipating future requirements, to shape a world-leading statistical system that meets the future needs of the UK. UKSA must also identify what resources it requires in order to carry out its role and functions to ensure the effective governance of UK statistics. (Paragraph 168)





Published: 18 July 2019