Select Committee on Treasury Appendices to the Minutes of Evidence


Memorandum from the Statistics Users' Council


  The Statistics Users' Council was formed in 1970 by Sir Claus Moser to act as a forum for the exchange of views between Government statisticians and users. The main activity has been the organisation of an annual conference at The Royal Society with a different topic each year—Education, Environment, Service Sector, etc. During the Rayner years these conferences were supplemented by seminars drawing attention to the harmful effects of the cutbacks in statistics, particularly business statistics. As we pointed out in 1991, UK companies no longer had the data to allow them to measure the loss of markets due to competition from imports. In order to maintain a meaningful dialogue, the Council set up User Groups for specific sectors—12 in total—Business, Health, Education, etc with an overall membership of 1,700 individuals. The Statistics Users' movement in the UK is unique. There is nothing comparable in any other country. It developed precisely because we did not have an Official Statistics Act with formal procedures for consultation.

  The Council unreservedly welcomes the introduction of National Statistics. This is a campaign we have been waging for over 20 years, starting with a paper on the need for an Official Statistics Act at our Annual Conference in 1978. Our only concern is that we may have misunderstood the Government's intentions. The White Paper and Framework Document eloquently express the role of National Statistics in providing a window on the work of Government, enabling the public both to monitor the performance of Government and to participate in the public debate on major policy issues. This accords well with our own preferred definition of National Statistics as "an essential part of the infrastructure for evidence based decision making in a market orientated democracy". We are moving to a participative democracy in which, rather like the Athenian model, all citizens can be involved, albeit connected by the Internet and the media rather than gathering in the city square.

  The aims and objectives in the Framework Document stress the need to consult with and develop statistics of value to users, but the operational guidance in the Framework Document is dominated by procedural matters and the perceived need to restore public confidence in official statistics. Guidance on the practical implementation of the National Statistics/User interface is hard to find. There is the pregnant phrase "the coverage of National Statistics will evolve over time", a reasonable statement provided the mechanisms are in place for the identification and evaluation of additions to the current statistical series listed as National, but again there is scope for misunderstanding.

  The euphoria we may feel about these developments needs to be tempered by the thought that we have done no more than join the community of developed democratic countries. The UK is one of the last, if not the last, of our compatriots to introduce a National Statistics Act or its equivalent. The Netherlands celebrated the centenary of their Act last year, and the United Nations Declaration on the Principles of Official Statistics initiated to help the emerging countries of Eastern Europe is almost 10 years old.

  The key discussion points in moving from the principles to the practice of national statistics include:

    1.  The scope of national statistics.

    2.  How do we identify and evaluate user views?

    3.  How do we inform and educate users (raise the profile of statistics)?

    4.  Budgets and governance.

  Those points that could lead to action are listed under the Agenda for National Statistics heading.


  The Framework Document does not define national statistics. It only describes those official statistics that fall under the national banner and instructs the parties to develop consultation procedures to take account of the needs of users, so there is the opportunity for change and addition. The National Statistics Code of Practice, adds another dimension:

    "National Statistics will provide for the integration of statistics from diverse sources through common UK standards across all sources and practices"

  but we are still not clear what is on offer. Is it:

    —  the existing published output?

    —  the detailed databases from which the published output is derived?

    —  improvements to the existing datasets to meet the needs of users including the filling of gaps, improvements in timeliness, frequency and detail, better correlation between series, record linkage?

    —  improved access through the publication of compiled CDRoms such as the one with all the relevant cancer statistics and the launch of CDRoms and Internet databases with sophisticated search and analysis programmes?

    —  the development of entirely new "official" series to service new policy initiatives and monitor performance against targets?

    —  the integration of statistics from other official bodies such as the regulators (OFWAT, etc), the police, immigration authorities, etc;

    —  the integration of data from other sources such as the BMA, CBI, Bank of England, etc?

  If we are only talking about the existing published output and improvements in access to the existing databases, what is all the fuss about? Len Cook talks about the national information base, an impressive and exciting concept, so hopefully we will move towards the position where we start with the decisions that need to be made on issues of public policy, wealth creation and monitoring government performance, and develop the appropriate statistical series to support those decisions either from within the GSS or from outside as appropriate. One not so casual observation—where does the initiative come from? Does National Statistics wait to be asked or does it take the lead?

  The scope of National Statistics is more than just the statistics that are available. It extends to how they are used. Which leads to another not so casual observation. Does National Statistics sit on its data or actively promote both availability and application?


  Statistics do not quite fall into the mousetrap syndrome. Better statistics can be built but the world will still not beat a pathway to their door. The market needs to be told what is available. The launch of National Statistics has come at an opportune time. Information technology, especially the CD and the Internet, offer unparalleled opportunities for low cost dissemination. These changes simplify the publication procedures. Aggregation and selection are no longer required. The user can access the database and create their own tables. The key is the development of sophisticated search engines so that the sales message is simplified: "We could have exactly the information you need at little or no cost—try us".

  The Marketing Department at the ONS has had a chequered history, and there is still the feeling that marketing is not fully understood. It was disturbing to note that the KPMG Efficiency Review for the ONS considered the marketing budget was one of the areas where cuts could be made. Contact between the GSS and the User Groups has been excellent. Government statisticians serve on the committees and provide speakers for the meetings. Those sectors of the market outside the Groups, however, are not well served. For example, a first class booklet "Profit from Facts" has been produced but not fully exploited. Part of the problem is that the DTI, having lost the Business Statistics Office, seems to have lost interest in detailed statistics, in spite of having set up the Business Links precisely to encourage business to become more competitive through careful market assessment.

  On the broader front there is a well developed press release system, but this tends only to relate to a particular series being published at that time. There is a yawning gap for a rapid response unit that can pull together the relevant statistics when "issues for public debate" arise. Either the moment passes and no statistics are quoted in the media, or a whole raft of conflicting statistics are launched on the public. "It is essential to consolidate and make sure that agreed figures only are used. The utmost confusion is caused when people argue on different statistical data. I wish all statistics to be concentrated in my own branch as Prime Minister". Winston Churchill's prescription for setting up the CSO is the right model to emulate. If public debate is to be advanced, it needs to be supplied with authoritative statistics.

  The advent of National Statistics has created a buoyant mood among statisticians. We seem to be returning to the nineteenth century position when statistics were the ammunition train of the social reformers, when Prince Albert formed the Statistics Dining Club and four members of Gladstone's Cabinet were on the Council of the RSS, but for the wider world the low profile of statistics has been maintained. The launch of National Statistics was a non event, just a few column inches in the broadsheets. References to the odd missing statistic, but no in depth analysis of the implications of the change from Official to National. The SUC has suggested the organisation of a National Statistics Week to bring home to the public at large the value of statistics in helping to develop rational policies over virtually every aspect of their lives. The summer of 2002 could be an appropriate time, as discussions are under way for the UK to host the bi-annual conference of the International Association for Official Statistics.


  We could scarcely be operating in a more ill defined market. We do not know how many users there are, let alone their "size and shape". When national statistics moves outside government customers it is operating blind. Historically, the ONS and other government departments have been production orientated. If any of the published series have been of value that is the users good luck. If not, bad luck! That is not to say that many series are not of the greatest value to users, but that it is really a matter of chance.

  The Framework Document is full of references to consulting users and meeting the needs of users, but not forthcoming on how this should be done. That is left to the National Statistician and the Statistics Commission. The task is not straightforward, as the over-riding characteristic of users is their variety. The non government market can be segmented in many different ways:

    —  By sector—the public, business, financial markets, academia, local authorities, development agencies, institutes, charities, pressure/lobby groups. Each of these sectors can be sub-divided. Business, for example, splits between manufacturing industry and the service sector, and each of those into many more specific markets. The important point is that each of these sectors has its own requirements for detail, frequency, timeliness and accuracy.

    —  By level of knowledge—a major distinction can be drawn between the expert user who understands the data being used and the passive user who simply reacts to the statistics quoted in a media story. There is, of course, no clear dividing line between the two, but a continuum. Frequency of use is another variable. There is a wide variation in understanding between the regular and the occasional user.

    —  By type of decision—strategic decisions taken by the Board of a company tend to require macro economic statistics, whereas market share decisions taken by product managers need highly disaggregated data. If you are selling malt whisky you need market statistics for malt whisky, not data for the whole Spirits sector.

  The attributes of statistics—detail, frequency, timeliness and accuracy—vary widely between the sectors. Minutes are critical in financial markets, detail and frequency essential for market share decisions by business, accuracy essential for the Bank of England Monetary Policy Committee!

  A major distinction should be drawn between the public and other sectors. The public, unless we take the media as a proxy, tends to be a passive user, which throws the responsibility for determining what statistics should be collected, developed, improved, back onto national statistics. It is a derived demand from the policy and other decisions under consideration. Ivan Fellegi in Canada has led the way. In his paper to the 1997 SUC Conference he demonstrated how statistics could become involved as an integral part of decision making by reference to the Canadian Policy Group's Report of 1996, which identified the main pressure points that were likely to arise in Canada over the next 10 years as a result of economic, demographic and social trends. The Group identified the main forces as globalisation, the information technology revolution, environmental pressures, an ageing population—a good base for any similar analysis in the UK, to which we might add, as a result of the Chancellor's recent threat or promise following the fuel blockade of a debate on taxation and benefits to the public deriving from government expenditure—more or less government. Ivan Fellegi's paper went on to specify the policy challenges in more detail under headings including economic growth, health, crime, income, distribution, productivity, etc.

  The challenge for National Statistics is how to emulate this approach. The ESRC seems a prime candidate for involvement on the wider front, but in sectors such as health the BMA and the various medical professional associations such as the Royal College of Surgeons, would appear to be front runners for integration, viz the health scandals of recent years—the Bristol Infirmary baby deaths, Harold Shipman, and in recent weeks the report that at the St George's cardiac unit the death rate was five times the national average. Crime and punishment is an unresolved quandary. We incarcerate more and more prisoners, and yet the crime rate rises. Safety is always in the news. Discussions of draconian measures for speeding and drink driving are followed by reports that tiredness and inattention are the cause of the majority of motorway accidents. These examples are only the tip of the iceberg. On this sector approach the national statistics theme groups would appear to be well placed to identify the issues that need investigation. We need to bring together the existing and develop new statistics that will inform the public debate in these key fields. Just what are the forces at work, measuring them at two levels:

    —  what is actually happening?

    —  leading to an understanding of economic and social phenomena.

  The expert users, by comparison, have much more mundane requirements:

    —  Inclusion of floor space in construction statistics.

    —  More timely detailed production statistics: The fact that UK manufacturing industry was disadvantaged by comparison with our major international competitors because their data is monthly and ours is annual was commented on by the last Treasury Select Committee review of the ONS, but no action has taken place and none is in sight. We still have annual PRODCOM published almost a year later.

    —  Record linkage in databases to facilitate analysis, eg location, size and ownership of companies so that output and trade figures can be analysed more extensively. The identification of imports and exports by multi-nationals would be an important addition to our knowledge of international trade patterns.

  This part of the market is highly segmented with each sector having very specific requirements which can only be properly identified and quantified by a well structured questionnaire with attitude rankings. The present consultation exercise is a good start and should help to create an awareness of the questions to be asked in the full survey. As it stands, however, the consultation exercise can only generate anecdotal evidence. The sample to be interviewed should be drawn from:

    1.  The purchaser/user of GSS publications.

    2.  The membership of the Statistics Users' Groups and similar bodies.

  These two can be supplemented by lists held by the ONS and other government departments.

  The questionnaire needs to be fully structured and to probe for each publication into the precise tables which are used and the purpose for which they are used. Attitudes to the main determinants—detail, timeliness, etc—need to be collected using rating scales, not open ended questions. The results then need to be analysed by each type of user. All this is just standard market research practice, but using these survey methods is considerably more expensive than the present consultation process. If we are serious, however, about identifying and evaluating user attitudes and requirements in the form in which the responses can be converted into product improvement—changes in national statistics in line with user (market) needs—then such surveys are essential.

  The use of the Internet to invite comments has enormous potential but, again, for the moment would only seem capable of generating random observations.


  The Framework Document is cautious—"the aims and objectives will be taken forward within the resources made available", and the Minister, at the press launch, when asked the question "all this looks rather costly" replied: "we are not anticipating major changes in the budget. In fact, National Statistics is a process of actually assuring higher quality and greater trust, but does not involve a massive budgetary cost". Are we seeing a particularly brazen example of political equivocation, or when the time comes will the Treasury accept that the transition from Official to National will inevitably lead to increased expenditure? There is certainly plenty of scope for increases. The ONS budget is only around £150 million and it is one of the most under resourced "National Statistics Institutes" in the EU. Only Greece, Spain and Portugal have a lower staff cost per head of the population than the UK, while countries like Germany, the Netherlands, Sweden and Denmark are one-third higher. On a straight staff count, France has just over twice (8,141) and Germany over two and a half times (9,752) as many employees as the UK (3,760). These comparisons are for 1998-99 and are derived from an annual Eurostat report on staffing and cost levels.

  The UK budget for National Statistics could be doubled and still not be expensive or represent more than the smallest of Treasury small change.

  The CSO, when first set up, was directly responsible to the Prime Minister, a position it held until the Pickford report in 1989, when it was transferred to the Chancellor of the Exchequer. There was a certain logic in this move while the CSO was predominantly macro economic, but the enlargement to the ONS, and now to National Statistics, makes the link inappropriate. There are precedents. Statistics Sweden has jut been moved from the Finance to the Justice Ministry, subtitled the Ministry for Democracy. There are similarities between the impartial administration of the law and the impartial control of the basic statistics that define and measure the government policies that influence our everyday lives, so perhaps the idea is not so far fetched. Consideration should at least be given to a return home to the Cabinet Office.

  The relationship between the National Statistician and Statistics Commission is not entirely clear. Both are given responsibility for determining user views. The relationship between the National Statistician and User Groups is straightforward. User Groups are the customers and the "theme" working groups the product managers. This is the detailed level at which product improvement and development takes place.

  The relationship between the Commission and Users, however, is not straightforward. In setting up National Statistics, the Government favoured the small supervisory body rather than the Grand Council representing all user sectors. The Statistics Commission represents the interests of users rather than being representative of users, so instead of a duologue between users and the National Statistician we have a triologue, hopefully more in tune with the harmony of a trinity than the antagonism of the eternal triangle. The Commission is charged in the White Paper with reporting within six months on its plans for consulting users and we look forward with great interest to the description of the "machinery for covering the interests of User Groups". As we noted earlier, the face to face debate on the needs takes place between the theme groups and the users under the aegis of the National Statistician. The role that users would appreciate from the Commission is the endorsement of such jointly derived plans and their support for any budget increases required to implement them, together with the facility to appeal to the Commission should those direct discussions not generate the results sought by the users. To undertake this role the Commission will need access to expertise and the User Groups are an obvious source. Consideration should be given to setting up a formal advisory structure with representatives from each of the User Groups and other user constituencies. The problem with the Grand Council approach is that it is too big to reach decisions, a situation normally resolved in practice by creating a small executive. The UK Statistics Commission has the reverse problem. It needs to develop a root structure to tap down into the user market and it would seem sensible to build on the already strong and well developed user community. The prescription given in our Green Paper evidence is for the creation of a Statistics Users' Forum—through an extension of the present structure of the SUC and its Groups. The SUC was set up because there was no formal link between Government statisticians and the Users. Now that the Commission has been given that legal requirement to interface between the National Statistician and the Users, the SUC is technically redundant.


  The following points are a user based assessment of the developments deriving from the transition of official statistics for government to National Statistics for both the public and the government. Some are obviously the responsibility of the National Statistician or the Commission, but we are operating in very grey areas and there is considerable scope for discussion between the two parties about who does what.

1.  Hold a National Statistics Week

  A target date to meet (Summer 2002) is a good discipline. We have an exciting message to get across to the public and this will provide exactly the focus needed to bring all the elements together.

2.  Exploit existing statistics as fully as possible

  There is a wealth of valuable under utilised data that is not being fully used. We need to raise awareness and improve access using the latest information technology.

3.  Build on an historical position

  UK official statistics under Sir Claus Moser were well on the way to a de facto national statistical service with the creation of the BSO and OPCS, the publication of Social Trends and plans for a mid-term census in 1975. The quarterly Business Monitor series sold half a million copies a year (the PRODCOM equivalent is around 30,000). National conferences to draw attention to the value of statistics in management decision making were organised with The Times newspaper, and a train mounted exhibition of Government statistics toured the country for 11 weeks, playing to 7,000 delegates. A Guide to Official Statistics was produced, which won a Library Association award. In several areas we are still not back to the position we left in the late 1970s.

4.  Absorb international ethos

  The last 20 years have seen significant developments in other countries. The Canadians, under Ivan Fellegi, are not the only country to have thought out and implemented the role of statistics for users. The Dutch, the Danes and the Finns have all developed exciting initiatives and Eurostat has become a virtual treasure house of comparative reports and theme studies. The George Als report is a classic. Attempts have been made to provide cost benefit reviews, and research units set up to study problem areas. The latest was a unit set up a few weeks ago to study the trade cycle. The annual conferences of the International Association for Official statistics have provided an excellent focus for many of these developments. Looking at them over time, there is an interesting shift of emphasis. In the early 90s the pressure was on marketing and selling statistics. Now the overwhelming mood is for statistics as a public good.

5.  Establish framework for interfacing with users

  Both the National Statistician and the Commission are charged with consulting users. Exactly how do they divide the responsibilities up between themselves? The link between the User Groups and the Theme Groups is relatively straightforward, but the mechanism for consultation between the Commission and Users urgently needs to be established. As was mentioned under Governance above, the effectiveness of the Commission as the User's champion requires a formal framework that provides the Commission with continuous access to the expertise required not only to properly assess the report from the National Statistician, but also to provide its own initiatives.

6.  Devise a research programme to identify user requirements and attitudes

  Consultation has limited value if product development and attitudes are to be properly evaluated. A full scale survey is needed to take account of the high degree of market segmentation and to generate actionable results, ie specific changes to specific series.

7.  Carry out product audit

  What use is made of each publication? Here we should include government departments among the users. Are there any other examples like the General Household Survey without a departmental protector?

8.  Develop and publish criteria for National Statistics

  Just what is it that defines "National" and what are the mechanisms for bringing statistics within the national umbrella? Several possibilities are discussed below.

9.  Set up a research facility

    —  To identify both the ongoing and emerging policy issues. Explore links with other organisations such as ESRC Throw the debate open to the public through the Internet.

    —  To provide a rapid response to issues that suddenly arise—BSE, fuel prices, rail safety, immigration—bringing together at least the available official statistics and ideally adding data from other reputable sources.

    —  To examine business decisions and internal information flows in order to determine how best to tap those flows and develop the statistics management needs to provide the market perspective to their internal company data. The ideal would be a System of Business Accounts based on the annual company account.

    —  To keep a check on the misuse of statistics and write appropriate letters to the press or take other suitable action. The ideal would be a Statistics Standards Authority along the lines of the Advertising Standards Authority, but that is probably impractical, so the best we can look for is react quickly to the worst abuses.

10.  Establish relationship with official organisations

  The National Audit Office, the various regulatory bodies such as OFWAT and many other organisations at regional as well as national level produce and use statistics. How should they fit into National Statistics?

11.  Integrate departmental surveys

  A large number of surveys are carried out every year in a wide variety of sectors. They need co-ordinating and publicising.

12.  Review private sector surveys

  To determine the reason behind them and the extent to which they can be incorporated either under National Statistics or some new hybrid. There are, for example three House Price Indices—by Nationwide, the Halifax, and the Land Registry. The problem is that they are not always consistent and the DETR is developing its own improved version. Salary surveys are much more widespread and often of dubious quality, yet a study of comparative incomes is of great importance to a large number of people. It is a topic that needs more critical attention. Timeliness is a key element in statistics and the private sector has developed a wide variety of trend surveys because the official statistics are not sufficiently timely. Is there scope for National Statistics versions designed specially to meet this demand for timely short term indicators?

  Attitude surveys are also widespread and of variable quality. Are there any partnership possibilities with the NCSR?

13.  Develop marketing plan

  Not only to inform users of availability, but also to raise the profile of statistics as an essential element in public debate and for the improvement of industrial competitiveness and productivity.

14.  Develop form filling burden/benefit balance sheet

  Only 2 per cent of forms that companies complete are for statistics and these, in fact, are among the few areas in which there is a direct payback to the company in useful data. Even the Armstrong-Rees report found very little objection to form filling, and the Whitting report on trade statistics recorded the message that a large proportion of respondents were prepared to pay the government to keep collecting the detailed trade figures.

15.  Develop programmes for culture change

  Meeting the needs of users outside government requires a very different approach to the traditional commitment to meeting only government needs, both in the nature of the statistics—the need for disaggregation rather than aggregation—and in the approach to product development. Historically most changes have occurred as a result of international requirements (especially for Eurostat) or a specific government requirement such as new forms of the RPI. The market outside government is both varied and largely inarticulate, requiring an active product development policy based on market research as part of the full marketing plan.

16.  Devise priority criteria

  With national statistics as a virtually free commodity the problem is potentially unlimited demand. There is no market mechanism to regulate supply and demand through price, so it would be helpful for users to know the ground rules for determining the priorities between competing claims.

17.  Publish list of Government targets

  Together with statistics available to monitor progress and plans for developing the relevant statistics where there are gaps.

18.  Review Government legislation

  Lawyers and lobbyists should not be the only people going through new legislation with the proverbial fine tooth comb. There is also the need to determine the potential statistical requirement for monitoring the outcomes of that legislation and including those costs in the Act.

19.  Decide the right home for National Statistics

  Should it be moved from the Treasury. If so, where? In Sweden, the Statistics Office has just been relocated from the Finance Ministry to the Ministry for Democracy (see discussion under Governance above).


  The Government has taken a bold initiative in setting up a National Statistics service that facilitates democratic debate and provides the public with the basis not only for entering into the discussion of major policy issues, but also passing judgement on the performance of Government itself. The only real cloud on the horizon is budgets. The change from "Official" to "National" Statistics, if it is to have any real meaning, will require extra expenditure. We have crossed the Rubicon. Whether we advance slowly or rapidly will depend on the funds made available.

  As a not entirely light hearted footnote, we will know that National Statistics has come of age as a truly independent force when, in response to a statistically backed argument, the Prime Minister feels as frustrated as Disraeli, and repeats or rephrases that famous quotation.

30 October 2000

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2001
Prepared 18 January 2001