Select Committee on Treasury Minutes of Evidence

Examination of Witnesses (Questions 280-299)


14 JUNE 2006

  Q280  John Thurso: In our first evidence session all three witnesses—who I would broadly describe as the user groups—raised the question of the fragmented nature of UK statistics, particularly with regard to devolution. One of them stated that the 2001 population census—the most fundamental of statistics—was so fragmented that very few outputs are available for the UK as a whole. When I pressed on this, they said that it is not so much that devolution created the problem but that it has brought it into a fairly sharp focus. I want to make it clear that I am not in any of this challenging the devolution settlement itself; but it does seem that there is a real problem in producing joined-up UK statistics. Is this a problem the Government is aware of, and do you have any remedies that you might propose?

  John Healey: We also have no intention of reopening the devolution settlement.

  Q281  John Thurso: No, we can leave that aside.

  John Healey: Let me be equally clear and emphatic about that. You are right, Mr Thurso, it is not a new problem. In Scotland, they have been producing certain sets of statistics separately and independently for 100 years or so. It is not a new issue, therefore, but perhaps it is being emphasised by the statistics being part of the devolution settlement. Yes, we are aware of it. We have sought to deal with some of the statistical difficulties that that throws up in two ways. First of all, the Framework for National Statistics, you may remember, contains a chapter on each devolved administration and their statistical approach. Also, we have settled as part of the Memorandum of Understanding between the UK Government and the devolved administrations in 2001 a concordat on statistics. Essentially, this is a commitment by the administrations to work together to produce agreed, coherent, reliable, and as comprehensive as possible sets of statistics. However, there will inevitably be variations, as there have been for some time. The Scottish law, health and education systems are different. Some of their outputs are different. Some of the information, monitoring and statistical requirements will therefore be different, and that is a consequence of the devolved function and the differential decisions that are made. We have been discussing quite closely with the devolved administrations, in particular the leading professional statisticians, how we ensure that we can maintain, as far as we can, an ability to see compatible datasets and comparability, where we can achieve that. To the degree that the devolved administrations are willing to look at renewing and perhaps reforming and strengthening the concordat, we would be very keen to do that.

  Q282  John Thurso: That was going to be my next question, namely that all of the concordats were written at a time when devolution was a new horizon and a new adventure, and to ask you if the Government would be prepared to look at rewriting the concordat, to make it more appropriate. I think you have answered that. What responsibilities do you think should be given to the National Statistician in overseeing the coherence and consistency of the UK-wide statistics—if any?

  John Healey: The majority of economic statistics are produced centrally by the Office for National Statistics at the moment. The National Statistician has oversight and will continue, as Chief Statistician, to do so. Many of the stats of the Office for National Statistics also have an underlying element that the devolved administrations feed into. I would hope and expect that that would continue. The ONS regularly receives around 250-300 datasets from the devolved administrations, which go into the compilation of the statistics the ONS is responsible for. There is also that area of statistics that, as the UK Government, we are obliged by international obligations to supply. Once again, the agreement we have in place with the devolved administrations allows us to fulfil those international obligations. In all those areas I would expect to see a continuing, important, professional lead role for the Chief Statistician. The area where clearly there are the differences and the difficulties are those devolved policy areas where the activities may be different according to the devolved administration and, therefore, the statistics and data available or collected. Here, the role that the Chief Statistician for England will have, dealing with her counterparts—as she does at the ONS at the moment, to try to make sure that we secure as great a degree of comparability, where that seems necessary—would also continue.

  Q283  John Thurso: At the moment, the Statistics Commission have a UK-wide remit. Do you propose that the new board would maintain that UK-wide remit, in terms of a kind of overall quality audit?

  John Healey: Nothing in the new arrangements will  interfere with the devolution settlement. Fundamentally, therefore, devolved administrations —because it is a devolved function—have the responsibility and the scope to develop and run their statistical system as they choose. Clearly there are the shared imperatives, which we have just touched on, about comparability and consistency. When we introduced the code of practice as part of the 2000 reforms, broadly the devolved administrations accepted, signed up, and generally followed that. I would hope that we could get to a similar situation, but ultimately it will be for the devolved administrations to decide the degree of buy-in and involvement they have in the new system that we will legislate for.

  Q284  John Thurso: It seems to me, both through this dialogue and through previous evidence from other witnesses, that a new concordat may well be highly desirable, and that such a concordat will clearly be quite important. Would it be possible for that to be consulted upon before it is finalised? Is that a possibility in the system?

  John Healey: I see no reason at all why it should not be. I could see the significant advantages if it was properly and publicly consulted upon. Certainly within government we would need to consult very carefully across government. Particularly as it would essentially be in the statistical area—it would be a sort of formal agreement between the UK Government and the devolved administrations—it would be difficult to see that process completed without quite serious consultation.

  Q285  Mr Todd: One of the sensitive areas which we have frequently had touched on is the access of ministers to statistics before they are released. The Statistics Commission favoured moving to a position in which there is no pre-release disclosure to ministers, and indeed that equivalent access is given, if there is any pre-release, to Opposition spokespeople. What is your response to that?

  John Healey: I probably have three responses to that. First of all, you will have seen in the consultation document, Mr Todd, that we have included the question of pre-release to see what views we might get as part of the consultation and we will look at those carefully.

  Q286  Mr Todd: We have not heard anyone say, "We're happy as it is", but you may within your consultation. I do not know.

  John Healey: We may. We may get people saying, "This is a good system. It should be left untouched". We may get some people saying that, on principle, this is a system which is not acceptable and—

  Q287  Mr Todd: We have chosen our witnesses poorly if there are people in that first group!

  John Healey: . . . or we might get people saying, "There is a sensible case for pre-release, but it could be reformed in a way that would make it tighter and perhaps inspire more confidence". If there are suggestions about reforming the pre-release system, therefore, we will clearly look at those. There are two further things that I would want to make clear. First of all, I would certainly accept that the pre-release arrangements contribute to the perception of interference in statistics. In fact, the cases of abuse of the system and actual interference in the production, and indeed even in the release of statistics, are very few and far between. Nevertheless, there is a perception there. Part of the drive to legislate now to entrench the independence is to try to deal with some of the problems that are still there in perception. Finally there is the question—and this is a question that people must make their own judgments about—of is there or is there not a case for any form of pre-release arrangements? In my view, the principle of pre-release is justified largely because—particularly in today's world, with the sorts of imperatives and pressures—ministers are required, expected, as part of our duty to be accountable for the decisions and what is going on in government, to understand and respond immediately to challenges that might come from the production of statistics. In those circumstances, I think it is right and sensible that there is some degree of pre-release. Second, I think that the principle of pre-release is quite widely accepted internationally. The details may be different but the principle of pre-release is accepted. It is accepted in Australia, New Zealand, the USA, Ireland and France. So I think that the principle of pre-release is sensible and defensible. The practice of it is important; the details of it we will be prepared to look at; and we are looking forward to the views we get through the consultation on that.

  Q288  Mr Todd: There is a little bit more than a perception of abuse, because when we questioned the Statistics Commission on their investigations into claimed abuse they made a rather ambiguous remark, in which the word "ambiguous" appears quite a number of times. "We have found at times . . . some difficulty in getting the responses from departments, and I think if you do not get the information you require then it is somewhat difficult to be certain, especially in this highly ambiguous situation of whether something is a breach of an ambiguous code . . . I think there are many cases where there have been representations of the statistics saying one thing—and often before the statistics have come out in the public domain . . . ." That was a remark made to us, which did not give that ring of confidence which you have expressed, which is that there may be a perception but perhaps not a problem. They feel that the regulatory process they carry out is incomplete and perhaps rather hobbled by the code that they are using.

  John Healey: The Statistics Commission will have to explain the evidence they give before the Committee. If I heard you right, Mr Todd, they said that there was often a problem. Actually, the number of breaches which the Statistics Commission is responsible for investigating is small in number. I gave an indication of the 2004-05 figures, and it looks as if the 2005-06 figures are fewer than that. If the Statistics Commission are not getting the information from departments that they require, I would expect that to be the sort of thing that they would draw attention to and report as part of their general duty, as they are currently constituted to draw to ministers' and others' attention where they see problems with the operation and integrity of the system at present.

  Q289  Mr Todd: You feel that they have not done that? Obviously they have made a remark which is scarcely a ringing endorsement of their own function in this role.

  John Healey: My point is more that the number and the nature of breaches of the code on pre-release are rather few and far between. Of the six they investigated in 2004-05 they found three were accidental. Three were breaches, but they were content, as I understand it, about the action that was taken to deal with them.

  Q290  Mr Todd: Can I turn to a different topic? We have asked questions of a number of witnesses on the access to data which is the basis on which assembly of statistics is achieved. We have had comments about the difficulties, sometimes varying difficulties, of obtaining core data from departments for independent statisticians and third-party business users then to develop their own series of information sources for their own purposes. Would not this legislation be a good opportunity to set that straight and make much clearer how core data can be obtained for use by third parties?

  John Healey: The legislation evidently is an opportunity to deal with some of the questions that are around data access, in particular access to administrative data. The issues around access to administrative data are quite complex. However, we have clearly signalled our interest in hearing views during the consultation process of the extent to which the current arrangements—which will obviously need as a minimum to be entrenched in legislation—could be developed further. Also, how at the same time some of the appropriate safeguards on confidentiality, particularly of micro data that can identify individuals, can be safeguarded—because that is obviously the flip side to that.

  Q291  Mr Todd: So there is an open mind on this topic. Because the way in which this debate has been conducted in large part has been almost "Westminster village" stuff: of how to stop politicians being horrible with statistics, when there is a broader community who rely on access to data and use it for their own purposes, who would love to be included in the broader remit of this bill.

  John Healey: There is a broader community with interest in access to administrative data. There is also an interest within the statistical professionals, particularly in the ONS, in access to some data that other government departments have at the moment that they do not. There are potentially benefits. There are potentially benefits on the business side for reduction in some of the survey requirements and the burdens of responding there. There are also benefits in terms potentially of statistical quality, which would allow the ONS and others to develop much more finely grained data and analysis, which would contribute in many way. However, there are these knotty and residual or core questions, which revolve around how you make sure your safeguards are properly in place to ensure that. We have said that we will consider, and we will, the views that come back to us through the consultation process.

  Q292  Mr Todd: Finally, stuck within this process is the Register Office function, which the Government is suggesting transferring to no particular destination, but away from where it resides now. Do you have a feeling of where it might be relocated? We have had a signal from statisticians that their only discomfort in this is in ensuring that the core data which the Register Office function does collect remains easily accessible to those who need to use it. I do not think that we have had anyone saying, "No, no, we wish to keep the function of the registration of marriages and deaths as an important activity of the statistical function of government".

  John Healey: No, we do not have a fixed view on this at the moment. It seems to us—and we have indicated this in the consultation document—that where formally there are statistical responsibilities currently identified as the responsibility of the Registrar General—who happens to be one and the same with the National Statistician—it is sensible for those such as the Census to be properly with the Office for National Statistics in the future. However, because there are functions of the Registrar General overseas areas that are not necessarily statistical, which require the involvement of ministers, then it seems sensible to us to be expecting to separate that function out and locate it elsewhere than within the current ONS. There are certain policy functions, particularly around the registration roles of the GRO, which necessarily involve ministers. There are also legislative functions around that which directly involve ministers which in our view do not make it appropriate, in the new statistical world, for it to sit with an independent statistical office.

  Q293  Ms Keeble: I want to ask a bit more about the development of new statistics, in particular to help to support evidence-based policymaking. So, if you like, the small-scale statistics—and you have mentioned the need for more fine-grained statistics—on which the neighbourhood renewal programme depends, or research for some stats to measure productivity in the public sector. How would you see those being developed under the new system that is being proposed? Would you see ministers commissioning them and, if so, would it then be from the independent office or would it be from their own departments?

  John Healey: In general terms, if there was a strong policy or government need for statistics, then clearly government and the lead department in all likelihood would want to commission those. The judgment would probably have to be in two stages. First of all, are they best commissioned from the Office for National Statistics, as the central statistical service, for instance in the way that the development of the neighbourhood statistics and the Neighbourhood Statistics' website have been done? Or are they best commissioned and developed within the department itself, given that our Government Statistical Service does have a presence, professional expertise, and heads of profession in each department? The second issue facing ministers in circumstances like that, it seems to me, would be, "Are these important enough—these statistics that we want to develop—for them to be classified and given the seal of approval as national statistics?", in which case there would be a question of proposing those to the board; probably having them provisionally designated as national statistics, but assessed and approved by the board as national statistics because they meet the code that the board has drawn up.

  Q294  Ms Keeble: I think that this issue about independence, integrity, credibility and so on, is a really difficult one. Does that not mean that you will have two tiers of statistics? One which is produced by the ONS and which is seen to have integrity, independence, and this, that and the other; and the other, which might be greatly used for functional purposes but would be seen to be dodgy or tainted or something, because they do not quite make the grade. Will it not mean that there will always be a dispute about political statistics and non-political statistics—which is actually what this whole debate is about, is it not? People dispute the statistics that deal with the politically most contentious issues.

  John Healey: First of all, I accept that we have a situation at the moment where, if you look at the omnibus survey commissioned by the ONS and the Statistics Commission, 62% of the public think that ONS statistics are more trustworthy than those produced within government departments. That is precisely the reason why we are looking to bring in a new system for national statistics, in which they are not just declared to follow a code of practice but are independently assessed as such, and independently adjudged and approved as such. A system which will reach into the departments where the majority of the key statistics—in other words, the national statistics, as I have explained to the Committee—will still be produced. What I hope, therefore—and this is the intention of the new system—is that independent assessment and health check, if one likes, is equally applicable to what the ONS produces as it is to the national statistics that are produced within the departments. The only way of approaching that differently would be to say, "All National Statistics have to be produced by the Office for National Statistics". You would have to take out the production of a large range of statistics from government departments, if you pursued the concern that you have to a logical conclusion.

  Q295  Ms Keeble: Can I come back on this point once more, because it is a really key one. Supposing you have an issue about measuring poverty, for example. This is a really key issue because, on this, all our rates, council tax subsidies and suchlike are calculated, are they not? Large amounts of government money go on that. If there is a really big dispute about what constitutes poverty—for example, how you factor in ethnicity or measure sparsity, what happens to rural poverty or, for instance, having been a councillor in Peckham, there might be an acute concentration of poverty which, in terms of London-wide statistics, completely disappears, with the Office for National Statistics saying, "That's not poverty" or "We wouldn't accept that that's a way to measure poverty", and a great body of opinion, which would probably be politicians, saying, "Oh, yes, it is"—how would you deal with that?

  John Healey: But is not that precisely what this system—and perhaps I have not made myself clear—will actually deal with? What matters is not who is producing the statistics; what matters is precisely how they are produced. In other words, is the information, the data, gathered, collated and produced in a way that meets the professional standards, which would be set independently in this code of practice and assessed and adjudged independently of government? That is the quality check, the confidence check, that this new system promises. It is there, available to what ONS produces, as indeed to datasets and stats that come out of departments.

  Q296  Ms Keeble: All the procedures in the world cannot take away the fact that at some point people have to make decisions and assessments. To take a really practical example, when Ofsted was set up it did league tables which measured school performance; it took kids' assessments. Your party did it. When we were not MPs and running these councils, the schools all cried foul and said, "Those aren't valid, because they don't take value added into account". Some years down the line, you then get value added included. It is disputes like that, which are about ways of measuring things. However much you try to sanitise it, there are political judgments in there. Would you accept that, however hard you try, independence will not be able to resolve some of those really difficult political issues?

  John Healey: I think that it will help considerably. I said a while back that if I am the minister responsible for a policy area and I am looking for data to be generated in that area that tells me how it is going and which I can use to tell Parliament and others how we are doing, clearly I have the option of doing that within the department. If there is a dispute or a challenge to the way that that data is produced, how much confidence people can have in it, whether it is professionally robust and reliable, I am in a relatively weak position if my argument is essentially one of contesting a judgment made by a professional statistical service like this. From my point of view as a minister, there is quite an incentive for me to say, "This is important. It will help me and it will help confidence in what we are doing if it is brought into the national statistical framework and assessed as that".

  Q297  Chairman: I am still not quite clear, Minister, if your real answer is that all these statistics will hopefully now be produced to the same high standard through the code of practice, as to why you will still end up with two classes of statistics. If you rightly are enhancing the perception of national statistics, is not the danger that you will still have this kind of category II of government statistics? Lord Moser—and it is unfair to refer his evidence to you because you were not here—simply did not understand why there could not be a single definition of statistics.

  John Healey: I did try to explain before that the nature and range of statistics produced by government that are official statistics are virtually unbounded. The question is can you devise a system, and would you want to devise a system, that somehow gave a quality assurance check on all those? Our judgment is no. Then the question comes: what are the statistics that you give priority to? It seems a sensible starting point, particularly compared to other countries—although they calculate them differently and we have significantly more national statistics that would fall within this framework than perhaps others—to say, "This data is important and it is particularly important for holding the Government to account, or because it tells us the most important things that we need to know about, where the economy and society are changing".

  Q298  Chairman: I understand that but, in creating these A-list statistics—the alpha statistics—and enhancing the way they are branded, you are almost admitting that there is a whole chunk of official statistics that cannot be quality-assured to the same standard. It seems to me that you are actually weakening the perception of those statistics that will not be National Statistics.

  John Healey: They could be quality assured if there was a decision that they should be brought into that system. What I am arguing, however, is that there is a huge range of data that is produced—in some part as one-offs, sometimes in the natural course of just running departments—that probably does not warrant that sort of status and scrutiny.

  Q299  Peter Viggers: You mentioned that 62% of the population had more confidence in national statistics.

  John Healey: The ONS-produced statistics.

previous page contents next page

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

© Parliamentary copyright 2006
Prepared 26 July 2006