Appointment of the Chair of UK Statistics Authority

Written evidence submitted by Royal Statistical Society (UKSA 01)

Summary

1. The Royal Statistical Society (RSS) considers that priorities for the new UK Statistics Authority Chair are:

· Promoting the need for coordination in the decentralised and devolved UK statistical system; this includes keeping under review current cuts in spending and "sounding the alarm" if any cuts appear to damage the overall integrity of the system

· Ensuring that government statisticians pay more attention to all actual and potential users, not just those in central government departments

· Supporting the National Statistician in improving the status of official statisticians and in enhancing their professional development

· Dealing satisfactorily with the UKSA’s dual role in both producing and scrutinising official statistics

and above all

· Taking all possible steps to improve public confidence in official statistics.

2. We note though that the UKSA cannot achieve these aims, particularly the last one, on its own. It needs the cooperation of departments and UK governments and the support of PASC and Parliaments to ensure that this happens.

3. The main changes we see needed to the official statistical system are

· The cultural change of paying more attention to all actual and potential users

· Recognition throughout UK governments of the disadvantages, as well as the advantages, of a decentralised and a devolved statistical system and active consideration of how to offset these disadvantages including paying more attention to the UKSA and the National Statistician.

4. The RSS considers that the reduction in nominal time commitment for the new Chair is acceptable but is concerned over the implicit message concerning the status and importance of the post that the reduction in salary sends.

About the Royal Statistical Society

5. The Royal Statistical Society (RSS) is the UK's only professional and learned society devoted to the interests of statistics and statisticians. It is also one of the most influential and prestigious statistical societies in the world with an international membership. It is active in a wide range of areas both directly and indirectly pertaining to the study and application of statistics. The role of statistics in monitoring the economic and social conditions of the nation and their use in public policy has been a concern of the Society since its earliest days. It engaged actively in the process of reform undertaken by the previous government playing a particularly active role in the consultation, discussions and legislative process that led to the 2007 Statistics and Registration Service Act.

Background

6. Valid and accurate statistics are critical in understanding society and the economy. To be effective, and to meet the objective in the 2007 Statistics and Registration Service Act of "serving the public good", official statistics must be relevant, respond to current and future user needs, be produced cost effectively to high standards with high integrity, and must command public confidence. Prior to the 2007 Act UK official statistics, while generally being produced to high technical and professional standards, only met the rest of these criteria partially; public confidence in them was (and remains) very low.

7. We believe that the United Kingdom Statistics Authority (UKSA) under Sir Michael Scholar has made an excellent start in its role. The Code of Practice was introduced, the Assessment Programme is well under way, timely and well-informed reports have been produced. The UKSA practice in the Assessment reports of usually giving the statistics concerned conditional "national statistics" designation (with certain requirements which have to be fulfilled before the designation is confirmed) has proved an effective method of improving statistical handling by official statisticians. Sir Michael and the Authority have made a number of well-judged and balanced public interventions, some of which have attracted significant media interest. Together the UKSA actions have resulted in noticeable improvements to the statistical process and, importantly, to the handling of statistics by government departments. In particular the guidance on statistics handling, including confirmation that the statistics Code of Practice applied to all civil servants, issued in early 2009 by Sir Gus O’Donnell, following Sir Michael’s intervention in the "knife crime" affair (coupled with PASC’s interest in the issue), was a significant step forwards.

8. The goal is a statistical system which is relevant, responsive to current and future user needs and produced cost effectively to high standards, with high integrity and commanding public confidence. However, attaining such a system is a long-term goal; it cannot be fully reached in one term since it often involves changes in culture and procedures. More importantly, since the UK statistical system is both decentralised and devolved, the UKSA and the National Statistician cannot, by themselves, bring about all the changes needed. They require cooperation from departments and from devolved administrations and we believe that this is not yet sufficient (see paras 10, 12 and 19 below). They also need the influence of Parliament, and in particular of PASC, in helping to ensure that needed changes happen.

9. Official statistics have a profound influence on decision making across society – whether by central government, local government, government agencies, business, academic researchers or individuals. Whether it is informing democratic processes such as local elections, providing information on the state of the NHS, influencing resource

allocation in the state and voluntary sector, guiding business decisions, shaping academic research which in turn shapes policy, annual uplifting of pension payments, influencing individuals’ choice of school for their children or many, many more examples - statistics matter. Thus official statistics must not just be the by-product of government; they must be carefully designed to reflect the needs of a wide range of users, produced well in a coordinated way, fully and carefully explained, and easily accessible when needed. These imperatives are as yet only partially acknowledged by government departments and other producer bodies; raising recognition of these facts is an essential task that the UKSA has started on but where it has a huge mass of departmental inertia or outright resistance still to overcome.

Priority issues facing the new Chair of UKSA

10. Uncoordinated cuts to official statistics. We fully accept, as does the UKSA, that official statistics have to bear their share of the cuts in government spending. But cuts should not be done without relevant departments consulting or dialoguing with users and considering actual and potential needs other than their own. Otherwise there is a real danger of an uncoordinated and ineffective statistical system. However while some departments have discussed proposed changes with users the UKSA has already had to protest formally [1] about two announced cuts. The first was the decision by the NHS information service to discontinue its share of the funding of the Lifestyle Survey (subject to ratification by the Department of Heath) – a decision apparently taken without any consultation of users. The second is the proposed discontinuation of the Citizenship survey by the Department of Communities and Local Government despite a consultation which showed strong need for it, including by other government departments and in support of current government policies (such as measuring national wellbeing and the ‘Big Society’ initiative), without any substantive discussion as to how needed information could be obtained through other means. With detailed decisions on cuts by government departments still in their early stages we fear that these two instances might be only the tip of the iceberg.

11. Developing a more coordinated statistical service generally. The problem mentioned in the preceding paragraph is an acute example of a more general problem – that of balancing departmental autonomy with the need to produce a coordinated statistical system. While departmental autonomy needs to be respected there should be an obligation on departments and devolved administrations to consult and pay attention to the UKSA and/or the National Statistician as appropriate. The benefits of a decentralised and devolved system can only be realised if there is a strong centre to ensure coordination. We greatly regret that the Cabinet Office has not seen fit to endorse this view.

12. Issues can also arise regarding coordination between the devolved legislatures or between devolved legislatures and the UK as a whole. While we recognise the need to ensure that statistics support individual countries’ needs, efforts must also be maintained to produce harmonised statistics for the UK as a whole.

13. Supporting the National Statistician in improving the status of official statisticians and in enhancing their professional development. Thirty years ago senior statisticians in major government departments were at least at Director level. None of them are now, while in some smaller departments the Head of Profession is not even in the Senior Civil Service. At a time when statistical understanding is ever more important to the political judgements that have to be made this loss of capability is alarming. This requires both greater recognition of the value of statistical skills and enabling statisticians to develop the broader skill sets needed for higher level analytical posts. It is also necessary that the National Statistician has sufficient influence over the training and professional development of government statisticians.

14. Greater attention to the needs of all users; enhanced commentary and presentation. The UKSA recently published a summary of the findings from its first 100 Assessment reports covering 450 sets of statistics [2] . In many ways the statistics assessed already met the requirements of the code of practice. However the UKSA has found three areas where statistics frequently fall short: communication including commentary and presentation; improving engagement with all users (not just those in central government whose needs were in practice heavily prioritised); and improving documentation. Encouraging official statisticians to improve these areas of their work and supporting the National Statistician in these endeavours will remain a key priority of the new Chair. The RSS is ready to play a full part in this work to improve the understanding and use of statistics across UK society. The ten year RSS Getstats campaign launched in 2010 provides a real opportunity for the world of official statistics to engage with wider audiences.

15. Dealing with the UKSA’s dual role. While we do not think it ideal that the UKSA is both a "producer" of official statistics, via its responsibility for the ONS, and responsible for scrutiny, in practice the arrangements put in place by Sir Michael go a long way to overcoming this problem. The new chair will need to review the system to see if further enhancements are needed.

16. Improving public confidence in official statistics. While we have put this factor last in the list in practice it is perhaps the most important long-term priority and was a clear aim of the 2007 Act. It is a very long-term goal which will not depend solely on the work of the UKSA but also on other factors including the attitude of governments more generally. We believe that the work being done by the UKSA, provided its efforts continue, will ultimately facilitate an improvement in confidence although it will not on its own be sufficient. The causes of low public confidence in official statistics are multiple but perceptions of government interference and control are clearly one of the key elements. (In the latest survey into confidence in official statistics [3] carried out in 2009, only 32% of respondents agreed that "official figures are generally accurate" while 40% disagreed, the remainder not expressing an opinion. Only 16% agreed that "official figures are produced without political interference".) As paragraphs 10 to 12 and 19 below demonstrate, while deliberate and malevolent instances of political interference are very rare, in practice ministers and Whitehall still have too much "soft" influence over the official statistical process. The new Chair will need both to seek to reduce, or at least balance, this influence and to continue to grow public recognition of statistical independence. We also regret that the Coalition government (and the devolved administrations where it is within their power) has not seen fit to endorse the UKSA’s recommendations on further restricting pre-release access to statistics.

17. An essential step in improving confidence, particularly among users, is to ensure that official statistics are "carefully designed to reflect the needs of a wide range of users, produced well in a coordinated way, fully and carefully explained, and easily accessible when needed – as paragraph 9 explains.

Changes which may be required to the official statistics system

18. Paragraph 14 above indicates some changes in practice which official statisticians generally need to adopt. We would like to highlight the need to listen to, and engage with, all actual and potential users of statistics. The decade prior to the 2007 Act saw an increase in the effective influence of Whitehall departments over official statistics, partly as a result of changes within departments which have seen statisticians increasingly "bedded out" in policy sections. Meanwhile the work of the Office for National Statistics has in practice become influenced by the "service agreements" it has with organisations such as the Treasury and the Bank of England. As a result even professional users of statistics outside central government often found it hard to have their voice heard and their needs paid attention to. In recent years the tide has turned but much progress remains to be done not just as regards actual professional statistics users but also with potential users. We accept that engaging with users, who are usually themselves busy people with little time, is not always easy but efforts must continue if official statistics are "to serve the public good". And there must be more recognition by official statisticians of how they are in practice still overly subject to central government influence.

19. Changes in organisation. The decision to remove publication of crime statistics from the Home Office highlights the perennial debate over the advantages of a centralised statistical system versus the decentralised and devolved system that the UK has. There is no perfect system since the advantages in perceived independence and coordination of a centralised system means that government functions generally, including policy making, have less input from statisticians. There is also the danger of an "ivory tower" mentality developing in a fully centralised service. Further, devolution enshrined the rights of devolved administrations over relevant statistics. For the moment, therefore, the RSS continues to support the principle of a decentralised system but we recognise that the "costs" of this system in terms of the lack of coordination and the perceived lack of independence are becoming more apparent. It is imperative that there is greater recognition of its disadvantages and that additional consideration is given to what checks and balances could be put in place to counter them (including ensuring that harmonised statistical outputs can be presented for the whole of the UK). This highlights in particular the need mentioned in paragraphs 11 and 12 for departments and governments to pay more attention to the UK Statistics Authority.

Whether a two-day week is sufficient time commitment

20. We feel that the proportional reduction in time from a notional three to a notional two day week for the Chair of the UKSA is not a problem. The role is non-executive and the initial notional three-day week reflected the need to set up the system. It remains the case that the effective time required is likely to be greater than 2 days a week just as Sir Michael in practice devoted more than 3 days a week.

21. We are more concerned with the reduction in salary. At the time of writing we do not know whether there have been sufficient applications from credible candidates to justify the reduction. We are particularly concerned about the implicit message that the reduction sends regarding the importance of the post, which is regrettable. We note that the recent Hutton "Fair Pay Review" provided comprehensive arguments rebutting the notion that pay should not exceed, pro rata, that of the prime minister. The pay of the UKSA Chair should be comparable with other public sector appointments of similar importance and responsibility.

April 2011