Statistics and Registration Service Bill

My Lords, I am sure that we all know why the Bill is important, not only for Government but for us in society, and not only for finding out what is going on in government but what is going on in society generally. It is in everyone’s interests that government statistics are of high quality and, equally, that society can have confidence in them. Here is the dilemma: by and large, our official figures are of very high quality but, again by and large, they do not get the public trust they deserve, and this is where the Bill comes in. I approach the Bill and the reforms as a non-political, professional statistician, knowing from experience that we are dealing with a highly difficult and sophisticated activity. Designing, collecting, analysing and interpreting data requires the highest technical skills, and that is what government statisticians do, always with clear integrity and, I hope, free from political interference. Spin is foreign to their approach. This is the world in which I have spent most of my professional life, including more than a decade of being in charge of official statistics, combined with international responsibilities. I cannot resist mentioning one personal aspect of my career which is probably unique—it began behind barbed wire. I was one of the Jewish refugees who were interned in 1940 as part of a regrettable Government panic. There, in the Huyton internment camp, a fellow internee set up a statistical office and asked me to help him. That began my interest in statistics. Perhaps I should be grateful to the Home Secretary of the day for imprisoning me. Anyway, little did I think then that statistics would be my career and that one day I would be talking about in your Lordships’ distinguished House. So here we are. The legislation started in a formal initiative from the Chancellor of the Exchequer, very helpfully developed over the months by the Financial Secretary. As a professional, I warmly welcome it, because it intends to give official statistics the independence from Ministers that they deserve and thus to enhance public trust. But things are easier said than done, and the Bill, in a few respects—not, I think, too controversial—falls short of the initial aims. I will talk about two or three of these, and I know that my views—my warm welcome for the initiative and my qualms about details—are aligned with reactions from the Royal Statistical Society, the independent Statistics Commission and the chief statistician of Canada, Ivan Fellegi, probably the leading official statistician in the world—he certainly leads the world’s best statistical system. So I feel on strong professional ground. I must say a word about our system because it is unusual, and it has bearings on the reforms. In most countries, all official statistics are dealt with in a single office. That is where they are designed, collected, analysed and issued. This makes life infinitely easier for integrating statistics and clearly shows them to be at arm’s length from politicians. Public trust is more easily achieved and in most countries is not a problem. By contrast, we have always had a decentralised system. At the centre there is the Office for National Statistics and each Ministry has its own statistical activities. The point is to achieve greater policy relevance for the statisticians and keep them out of the back room, so to speak. This makes sense as long as the whole is run as a single integrated statistical system, led by the National Statistician from the centre. The Minister gave a very clear outline of the aims and basic contents of the legislation. The main intent is to create a structure that will enhance public trust. Trust, however, is a complex matter. For one thing, trust in statistics is part and parcel of trust in government themselves, and, indeed, in politicians in general. That aside, trust in statistics relates much less to their actual quality than to the way in which they are used. The suspicion, whether justified or not, is that political spin is at work. As a result, trust has undeniably dropped significantly in recent years and is now more of a problem here than in almost any other country. In an independent study, the Statistics Commission found that only 17 per cent of people think that government statistics are free of ministerial intervention—hence the importance of the Bill. At the heart of the Bill, as we have heard, is the new Statistics Board, which will replace Ministers as the top layer of governance for the ONS. However, it will, we hope, have a far wider remit. It will be independent of ministerial control, but will still be in the Treasury, which I greatly regret. I would have much preferred the residual ministerial functions to move to the Cabinet Office, where they were in my day. That would work much better. In passing, the board also replaces the Statistics Commission, which has done an increasingly powerful and influential job, although, of course, without the legal teeth now envisaged for the new board. I should also stress again—this point has already been made—that the board will report directly to Parliament, giving Parliament a new and crucial role in official statistics. Two strategic features of the board as envisaged are unsatisfactory. The first is its basic function, as originally conceived, for supervising the whole system, the ONS and, we hope, all official statistics. We have been assured by the Minister that the supervisory role is seen as distinct from the executive functions in producing the figures, which is obviously the role of the National Statistician and her professional colleagues. I say ““obviously”” because, despite the reassurances, this distinction between overseeing on the one hand and producing on the other is critical but is not clear in the Bill. In fact, what has emerged seems to give the board both functions, with repeated references to its role in producing statistics. No wonder the Royal Statistical Society refers to this as a muddle. This is not a major point, but I hope that the Government will consider the tidying-up needed to separate very clearly the board’s supervisory role from the National Statistician’s professional delivery functions. The other feature of the Bill that needs more clarity relates to the Treasury parentage of these reforms. As a result of that parentage, there is, for my money, too much emphasis on the board’s role in the ONS as opposed to non-ONS policy activities, where 80 per cent of the figures come from. The system is centralised but must be viewed as a single integrated whole, and the board and the National Statistician, as head of the Government’s statistical service, must be seen clearly to be responsible for the whole system and not just for the ONS. It would make no sense whatever if the Chancellor’s welcome new vision for independence from Ministers were to benefit only the ONS, while leaving departmental statisticians to remain as exposed to ministerial involvement as ever. It would leave the situation unchanged where, in my view, change is most needed. The ONS is widely and rightly regarded as one of the best statistical offices in the world, whereas departmental statistics are, to put it mildly, patchy. But it is those statistics that are most sensitive to public distrust; your Lordships could think of the figures on crime, migration, health, pensions and education, et cetera. The Bill needs strengthening where life beyond the ONS is at issue. I turn briefly to an historic anomaly which distorts the Bill; namely, the distinction between so-called national and non-national statistics. This was invented in 2000, which I suppose was meant to confer extra status on some statistics that then had to obey the strictest standards. In fact, it is a distinction without a difference, which in reality is meaningless and actually harmful in relation to a public trust. At present, only a tiny fraction of all statistical series are categorised as non-national, but the whole thing is arbitrary. For example, quarterly figures for hospital waiting lists are quite properly national statistics. But why on earth are monthly figures non-national statistics and therefore outside the most severe scrutiny? Other examples of non-national statistics are UK energy projections, monthly prison population counts and estimates of immigration population, et cetera. Not only does that make no professional sense, it is also contrary to the very purpose of the Bill because it leaves Ministers with the power to decide what are national as opposed to non-national statistics; in other words, which statistics can and cannot escape the strictest treatment within the crucial code. The right course is now to abandon the term ““national statistics”” or, at the very least, it should be made clear in the Bill that it is the task of the new board to decide what are and what are not. Certainly, Ministers should not hang on to that role. Finally, I turn to the most sensitive issue, which all speakers already have mentioned; that is, the way in which official statistics are released. The public are most suspicious of Government interference when political comment is mixed with neutral statistical comment and, above all, the fact that Ministers, advisers and officials get sight of figures well before publication. On the comment mix, we have heard from the Financial Secretary of his very helpful decision to set up a publication hub, which I will not go on about. However, on pre-release access, our situation has gone from bad to worse and we are one of the most lax systems in the world. In a number of major countries, no pre-release is allowed. In many others, it is very limited with only a handful of people getting access to very few series, mainly to those which are market sensitive. Moreover, even when permitted the length of time is very short, often just an hour or two. The President of the United States gets key figures a half-hour before they are published. It is beyond my thoughts to think what he does with those figures in that half-hour, but that is not my main point. Our situation is now very lax. Even market-sensitive data are seen by 10 to 20 people, most of whom have no need for advance sight. Some statistics are pre-released even more widely, often 40 hours or even longer, in advance. I am aware that some tightening up is in the pipeline, but if we mean what we say about improving public confidence, radical change has to be faced. Release has to be controlled by the statistical authorities, not by Ministers, and should be very strictly limited and disciplined. In my view, and indeed in the view of the Royal Statistical Society, it would be best to have no pre-release. However, that may be beyond hope. In any case, the decision to release should be left to the new board. I have mentioned a number of key strategic qualms about the reforms, but I end by warmly welcoming the initiative in the hope that a bit of government redrafting will take place before the Bill comes back to us at the Committee stage. I am aware that I have not commented on many topics of enormous importance: the crucial regional and local dimensions in the reformed structure; the importance of making the whole thing totally comparable with international practice and rules; the need to put greater emphasis on the voice of the user; and, of course, the important role of the administrative sources, surveys, and much else. The challenges ahead are formidable, not least because of the major funding cuts facing the ONS which I believe risk undermining quality. With all this in project, let me remind your Lordships again that new responsibilities are coming to Parliament as part of the reforms. They will require arrangements beyond the present committee structure in each House and, surely, involving both Houses. Given the right structure, Parliament will be ready to play a highly constructive role in the proposed reforms.