---

Lord Moser: My Lords, I am very grateful to the noble Lord, Lord Marlesford, for introducing this subject, which is close to my heart after a lifetime spent in statistics.

I need to distinguish the three elements of the set up: ONS, the Office for National Statistics, and its head called the National Statistician; all the departments—each ministry—have their own statistical office; and, independently of those two, the Statistics Commission, which the Government set up in 2000.

It is true that there have been some errors, especially in the construction of macro-economic statistics. If I had lots of time I would try to explain the incredible complexity of GNP. It is rather a miracle that all over the world statisticians in government can piece together from literally tens of thousands of bits of information on income, expenditure, production, and so on, the national accounts. It is a remarkable achievement, but it is so complex. Bearing in mind that every little bit probably has an error attached, it is even more remarkable that the sum total—the ultimate GNP—is as accurate as it is. Incidentally, it is widespread opinion round the world that our statistical system is one of the best in terms of accuracy, but occasional errors will occur, and the noble Lord has mentioned one or two.

It is crucial that we distinguish between errors and revisions. It is inevitable that if one tries to produce economic or any other statistics very quickly, there will be errors that will be corrected when later information is available. These are called revisions. Every country does them. It is part of the ordinary process of producing good statistics that when later information is available, one corrects the initial estimate.

In this country, the record is extraordinarily good for recent years. Between the first and third estimates, the difference is usually of the order of 0.1 per cent, which is fairly trivial. But there will be revisions. One of the examples given by the noble Lord was a revision caused by new information becoming available. There is a trade-off. We could lessen the revisions if we took more time, but the Government's pressure is always on timeliness. So data are published very quickly, then later revised.

One reason for revision is later information. The other reason is that methodology is constantly being improved. That is the role of the statistical office. For all of my 10 years in charge, one was always trying to improve methodology. Perhaps I may say that road maintenance, which has now hit the press, is the latest example. It is
 
23 Feb 2005 : Column 1275
 
absolutely correct for the National Statistician, with his office, to seek constant improvements—one such was sought in 2002 on this subject.

Whether the statisticians are right in making that change is a technical matter. On the whole, they are governed by IMF, World Bank, OECD and United Nations rules. They may be wrong in that technical change, but it has nothing whatever to do with, as, sadly, the shadow Chancellor said, fiddling the figures. That is a ministerial and political point. One can argue about the methodology, but one cannot call it fiddling the figures. In fact, Ministers did not even know that that change was being discussed by the technicians.

Everyone, including politicians, has to get used to the fact that there are errors and revisions. When I was in charge, my main problem was Ministers and the opposition parties over-interpreting the accuracy of figures, which is a greater problem in causing public mistrust than faulty figures. That is my first point on the present concern about accuracy and revision.

In my view—I am no longer in any way involved—the ONS now has a very fine record of macro-economic statistics. It also has a good record on social statistics and on widening user links, because the Government are not the only customer for official statistics. Equally important are local government, the trade unions, the business sector, the academic world and the general citizen. The ONS record in widening its customer base is probably one of the two or three best in the world. It is also very good at research and innovation. I could give a number of examples, but I will not take that time. All told, the ONS, including its leader, Len Cook, the retiring National Statistician, deserves its high reputation for statistical accuracy and leadership.

Public trust remains a major problem. It is probably greater in this country than in any other. It relates partly to the statistics about which the noble Lord, Lord Marlesford, spoke, but it is also a wider issue. The trouble is that many of the Government's statistics are not published by the ONS. As I said, each department has its own statistical office. For example, National Heath Service waiting lists are the responsibility not of the ONS, but of the Department of Health; migration and crime statistics are the responsibility of the Home Office. We have a semi-decentralised system where some of the most sensitive statistics are published by departments, independent of the ONS.

The National Statistician, which was the post that I held, is widely regarded as responsible for those statistics. He tries to influence departments, but that influence is very limited. He has strong influence over appointments to statistical offices, but not over what departments do. In my view, the greatest problem regarding public trust links not to the ONS but to the individual departments. There are errors in those departments and they do not always obey the codes of conduct that the National Statistician has turned out.

In my day, the National Statistician, and that was me, was responsible directly to the Prime Minister via the Cabinet Office. In a strange way, that gave me, and
 
23 Feb 2005 : Column 1276
 
my office, a greater reputation for independence than if I had been under the Treasury. It is a matter certainly for discussion whether the statistical office in future should remain so closely linked to the Treasury with its very clear policy interests, as opposed to under the Prime Minister or, in some sense, separate.

I conclude with a few final remarks. First, the Statistics Commission was set up primarily to monitor and scrutinise all government statistics. It does a first-class job. Its publications—now some 20-odd reports—are of the highest order. It is independent of government and, through its independence, achieves the kind of aims that the opposition party is now looking for. The Statistics Commission should be strengthened and should feel more free than at present to criticise Ministers and the media when they misuse statistics.

Secondly, I would keep the National Statistician and the Office for National Statistics as they are. It would be a disaster from a statistical point of view to separate analysis from collection, as has been proposed. The ONS should build on its strengths with full government support.

Finally, there is a great deal to be said for the proposal from the Statistics Commission that, partly in order to keep the departmental statisticians under greater control, there should be statistical legislation to bind it all together under the authority of the Statistics Commission as a statutory body, which I hope the Government will consider.

Lord Brooke of Sutton Mandeville: My Lords, it is a daunting prospect to follow the noble Lord, Lord Moser, for reasons of my own very slight personal experience in this area, but it is nevertheless a great privilege.

All our speeches in debates such as today's open with a conventional paean of praise and appreciation for whichever noble Lord has opened the debate. But my thanks to my noble friend Lord Marlesford are the more heartfelt because, by choosing this debate, he has opened my eyes to how much has been going on in the field of official statistics over the past decade.

As to the opening of the debate by the noble Lord, Lord Marlesford, I still retain pleasure at the barrister who, after affording consultation in his Chambers to a solicitor and his client would conclude the occasion by saying, "Well, there we are. I don't know where we are, but there we are".

My credentials as a statistician to create a statistical paradox are less than minimal, though in the second year of an MBA at the Harvard Business School 45 years ago I did do the optional course of advanced applied statistics under the legendary coupling—legendary at least within Harvard—of Schlaiffer and Raiffa. Schlaiffer was a full professor with a background in ancient history and I think that Raiffa was an assistant professor. I can still recall moments when they would turn towards the class and say that they wanted to have a private bilateral debate for five minutes but since they would take short cuts we were under no obligation to try and follow what they were
 
23 Feb 2005 : Column 1277
 
doing on the blackboard. I also learnt more than a little at the feet of Sir John Kingman when he was chairman of the Science and Engineering Research Council and I was his inadequate Minister.

I alluded to the developments in this field over the past decade, going back to the 1996 Act under the previous government. They will be familiar to those taking part in this debate, which obviates the need to repeat them, but I want to dwell on some aspects of the use or abuse made of statistics by politicians. I am conscious that I am at the margins of the subject of this debate in doing that, but I notice that there have been other references in earlier speeches. The integrity and adequacy of statistics is an important foundation to the confidence the public place in them, and if politicians undermine that confidence, they are not rendering any service to the common weal.

It is an odd anorak addiction to admit to, but my party—and no doubt other parties too—maintains an e-mail running commentary on the statistical background which our opponents are applying either in support of their postures or on a rebuttal basis to observations that we have made on their records. I fear that the instant rebuttal principle which we have absorbed from the United States is the enemy of intellectual honesty—I picked up on what the noble Lord, Lord Moser, said—yet intellectual honesty is critical to the good health and reputation of statistics.

I am particularly concerned when Ministers take public exception to statistical analyses produced by their own official departmental body. There was one such by the Health Secretary in commentary on an article in the Guardian in October 2003 about ONS statistics from 1997 showing deterioration in NHS productivity over the previous six years. I have not the slightest objection to Cabinet Ministers disagreeing with their own official statistics, but the common weal is better served if collective responsibility is observed in public and internal argument conducted in private. As to accusations across the political divide of lying, even if described as "fibs" buttressed by statistical arguments, I sigh for the demise of one of my favourite 19th century earnest endeavours, the Society for the Suppression of Mendacity.

One of the immense virtues of the Minister who is replying to this debate is his own intellectual honesty. He is robust in defence of the Government's achievements and policies when he believes alternative views are wrong, but he is magnificently generous when, as with David Willetts's unveiling of conceptual errors in official pension statistics, he believes that credit is due elsewhere. He is a model to the generality of the administration in which he serves.

Some of us in your Lordships' House will be looking forward to Sir Christopher Foster's new book on public policy to be published next month, not least because he has served over three decades as an adviser to governments of different cues and in different departments. None of us can be sure until publication day what he will say, but I have some cause for thinking that he will deplore the decline in the amount of statistical background to policy choices which
 
23 Feb 2005 : Column 1278
 
government official papers of whatever colour now display in favour of the alternative of setting out the outline of the Government's ideas and then laying greater emphasis on consultation.

If I am right, we shall have to await the historians' verdict on this Government as to whether this is the general cultural change or a predilection of this administration. But I am personally greatly attracted by my party's desire to see the independence and expansion of official statistics tied to the same kind of gold standard as the National Audit Office provides. The Government who conferred independence on the Bank of England might sensibly be attracted to the same principle. In the mean time, it is in tune with my right honourable friend Michael Howard's emphasis on accountability.

The Royal Statistical Society has given my right honourable friend's intentions a guarded welcome, while sensibly implying, as it did to the similar general pledge by Labour in 1997, that the proof of the pudding would be in the eating. Meanwhile, the Statistics Commission, created as we know by the present Government, has shared the view of the Treasury Select Committee, the Royal Statistical Society and the report of the independent government communication review group—the Phyllis review—that although the non-statutory framework for official statistics introduced in 2000 has been beneficial, legislation requiring government departments and agencies to follow a new statutory code of practice, enforced by a statutory commission, is necessary.

I conclude with a final bee which has buzzed in my bonnet throughout this Government's life—their insistence on referring to running costs as investment. In the last Parliament, in response to a press release, I took my life in both hands and taxed the present chairman of the Labour Party, when he was at the Cabinet Office, with Labour's divergence from Sir John Hicks's magisterial work on national income accounts. He effectively adopted the attitude of Humpty-Dumpty that he could call anything whatever he liked.

I listened carefully to the observation of the noble Lord, Lord Moser, about the roads expenditure example. I cannot help feeling, however, that the description of revenue expenditure as investment runs the risk of undermining public confidence in the process.

---