| Previous Section | Index | Home Page |
“lies, damned lies and statistics”.
His experience was based on the then Government statistician, whom he described as an imbecile who spent his time constructing libels with which to damage the Government. The problem was that the Liberals appointed the chief statistician. Party political badinage on the subject is therefore long established.
My third basic point is that although I welcome the move to independence, I worry—this is not a party point—about the tone of many of the professional comments from those who have been invited to comment on the Bill, including the Royal Statistical Society; the chairman of the Statistics Commission; Lord Moser, who is the towering figure in British statistics of the last generation; and the Statistics User Forum. All those bodies have broadly welcomed the Bill, but have been very critical of substantial parts of it, some of them damningly so. Bill McLennan, who was head of the Australian system, which has some merits as a model, said that the Bill would set back statistical independence for several generations. That is the most extreme expression of criticism, but it is clear that the professionals are not happy with the Bill as it stands.
John Healey: Does the hon. Gentleman agree with Bill McLennan’s assertion, and can he see anything in the Bill that would substantiate such an absurd assertion?
Dr. Cable: I recognise that it is the most extreme criticism, but Bill McLennan has no political axe to grind. I was reassured by the Financial Secretary’s saying that he was willing to listen and to debate the issues, and that the Government are open to amendments. If that is the spirit in which the legislation is considered, here and in the other place, I am sure that we will emerge with improved legislation that will not justify Mr. McLennan’s pessimism.
On my first point, Government statistics are important for different reasons. The first reason is the nature of the debate that we conduct in the political world and the need to ensure that it is consistent and sensible. We have had several examples in recent years of how a lack of confidence in statistics can make that debate very difficult. The most obvious case, which the hon. Member for City of York (Hugh Bayley) mentioned, was concern about the quality of employment figures. The hon. Member for Worthing, West (Peter Bottomley) intervened with a justification, but there is little doubt that there was considerable loss of confidence in employment statistics at that time, and it has continued.
We see another example every Wednesday lunchtime, when Members on one side of the House say that crime is increasing, based on figures in either the household survey or police recorded statistics—whichever shows an increase—while Members on the other side of the House claim that crime is going down, citing figures in the other survey that suggest the opposite to the findings of the first one. One of the most fundamental arguments—whether there are improvements in reducing crime rates—depends on two sets of statistics, which are mutually inconsistent and do not enjoy full confidence.
A more technical point, which is important to the debate and is of particular concern to the Economic Secretary, relates to the whole issue of child poverty, about which there will be much discussion over the next few years. One of the difficulties is that because we are talking about not an absolute but a relative measure, the extent to which progress is made is critically dependent on the methodology statisticians employ to measure it. We need to have absolute confidence in their integrity.
More important even than that is the quality and output of Government spending, with which we shall be preoccupied over the next two years. Over the past few years, there has rightly been growing concern about whether Government spending is productive, which links to the wider debate about resource accounting—something that is not only difficult but could also be subject to abuse.
One example of the difficulties in that area relates to the efficiency of Government spending on education. The ONS, using national statistics, estimates that over the past decade real spending inputs in education increased by about 2 per cent. a year and that outputs increased by 1 per cent. a year. In other words, we are making progress although the system is not enormously productive.
It was suggested that the statistics were not as meaningful as they could be, so they were modified to take account of quality changes, movements in tests and value added, which tended to show that outputs were falling. Department for Education and Skills statisticians then tried another measure, which was to take account of quality changes resulting from the number of pupils passing GCSE at grade C, and concluded that educational output was increasing substantially. Everybody has a different view, but it is interesting that the Department published the last survey, not the other two. That is the sort of problem that arises when there are differences in official and national statistics, and Departments and their Ministers have discretion about which statistics they choose to use.
One sense in which statistics are important relates to political debate; another relates to people’s livelihoods. As has been said, economic statistics have enormous implications for individuals and business. I shall give a few examples—some important and some less so. The argument about inflation statistics was cited early in the debate. We all accept that the move from the retail prices index to the consumer prices index was not political, but was technically driven. It has had enormous implications, however, and there is much frustration that the CPI does not properly capture property price movements, and that it diverges enormously according to which income group is under discussion. If there is to be any change in the measurement of the CPI, or if it is to be augmented by some other measure, which will of course have an impact on pensioners and benefit recipients, we must have absolute confidence that the people responsible for the process can be wholly trusted.
There are public finance examples, such as the degree of debt and Government deficits. At present, they are largely at the level of only public debate, but as the Financial Secretary has pointed out, such things have implications for bond markets. If we were ever to get as close as applying for admission to the economic and monetary union, such factors would determine whether we could be admitted and, if we were part of that system, whether we would be fined for complying with them. Our debates in this place are replicated in arguments in France about France Telecom debt and in Italy about the valuation of Italian gold. Our
arguments are not unique. Confidence in the statisticians is vital in that regard.
In discussing the importance of confidence, I shall use a minor and mundane example concerning how I got involved in this debate many years ago and why people working at the coal face of statistics have found the system frustrating. I was asked to work with the World Bank to produce a new approach to trade statistics—the Minister mentioned that we have been doing this for four centuries. We embarked on an exercise—this will really set the hearts fluttering—of reconciling three-digit industry statistics with five-digit trade data.
That enormously exciting subject had considerable practical implications, one of which was that it provided a measure of industrial competitiveness by sector. In addition, it allowed us to compare how open different countries were to manufacturing input and competition—that was why the World Bank encouraged us to do the work—and thus helped to frame the terms of reference of international trade negotiations.
The figures were produced and published by the Department of Trade and Industry and were used greatly, but in the early 1980s, they started to produce rather embarrassing results showing that the rate of market penetration in many industrial sectors was increasing alarmingly, and one day they were suddenly pulled. An enormous investment of the UK Government’s and others’ time was thus wasted, and useful and rich, albeit not terribly high-profile, data were simply abandoned. We thought that that was done for political reasons, although no one ever explained the decision. The Bill is designed to prevent that kind of possible political intervention at a rather low level from interfering with the useful production of data.
On the consensus on independence, I think that we all agree that the statistical service and the chief statistician have to be protected. The underlying reason for doing so—the lack of trust—has been mentioned. The hon. Member for Wolverhampton, South-West (Rob Marris) challenged the hon. Member for Chipping Barnet (Mrs. Villiers) because she had only one set of data. I have two, but they are cross-sections, rather than a time series, so I am not sure whether they entirely prove the point. It is worth citing the Office for National Statistics, which has suggested that only 14 per cent. of the public believe that the Government handle data honestly. A MORI survey provided confirmation of that by indicating that only a third of the public think that official statistics are accurate.
In a sense, the comments of professionals are more important than our views, so I have compiled a few quotes on the Bill from some of the leading people in the field. Before the Bill was published, Lord Moser, who, as I have said, is probably regarded as the leading authority on British statistics, cited the central problem:
“When it comes to reducing trust in official statistics, ministers are the main culprits”.
He welcomed the published Bill, but said that he was surprised “to put it mildly” by the way in which the governance was structured and the Government’s approach to pre-release.
The chief executive of the Statistics Commission commented that the successor body of the ONS would
not have “enough authority” to “resist pressure” from Ministers and their advisers. The Statistics Commission has outlined four major areas in which it believes that Bill should be substantially changed. Tim Holt, the president of the Royal Statistical Society, says that the Government have not bitten the bullet on pre-release. I know that the Minister has partially bitten the bullet, and we shall return to that in a moment.
The statistics users forum suggests that there are five major areas of concern about the Bill, and I shall run through what they are, because they have been discussed with varying degrees of thoroughness. The first is what is said to be the limited scope of the statistical board. The Minister has rightly said that its legislative responsibility reaches right across Government. The key phrase that keeps coming back in most of the comments of the professionals is that the statistical board, and the code of practice that it will use, will have responsibility, but not authority. There is an imbalance between the two. There are two specific problems with that, and we discussed one of them at some length on the Floor of the House—it was the idea that Ministers should decide on the distinction between national and official statistics, although that might, in some cases, be a politically sensitive decision. The other point, which has not yet been stressed adequately, is that even if the statistics board judges that, effectively, the code of practice is not being complied with, it has the power to name and shame, but not, as I understand the Bill, the power to make its decisions binding. I hope that the issue of the degree of authority, as opposed to responsibility, will be clarified in Committee.
The second point is on governance, and although it has been dealt with effectively by the Opposition spokesman, the hon. Member for Chipping Barnet, I shall touch on it briefly. The issue is partly one of clarity: the statistics board appears to have both executive and oversight roles, but tucked away in the Bill is the suggestion that a new executive grouping, which will produce the statistics, will be created within the board. It is not entirely clear whether the roles of production and oversight are separated. The example has been cited of the role of the statistics board chairman. He might, under some circumstances, both pass a judgment on the methods used to calculate a particularly sensitive set of figures and, in his other capacity, justify the conclusion. It will need to be absolutely clear that such a conflict of interest will not arise.
The third set of problems relates to pre-release, and the Minister advanced the discussion on that subject this afternoon. I may have misheard him, but I think that he said that the pre-release period will now be 40 hours, although that is not in the published Bill; is that correct?
Dr. Cable:
The period will be 40 hours, so as to be aligned with the provisions for market-sensitive data. That is an improvement, and we acknowledge that the Minister is trying to address some of the criticisms, but it is worth comparing that 40 hours with the situation in other developed countries. In the United States, the
time is half an hour; in France it is an hour; and in Australia it is three hours. The industrial country with the longest period is Canada, where it is 17 hours. Even given the Government’s concession, we are far from having a response time that other countries would regard as normal. The Minister made a good, eloquent case for pre-release, and anybody acquainted with government will accept the need for such a provision, but I would argue that the best practice, as demonstrated by most Anglo-Saxon countries, particularly the United States, is to allow a much shorter period.
My final set of points again reflects the comments of the professionals. They are on the way in which the statistics board will be subject to the residual powers of Ministers, and the board’s reporting requirements to Parliament. Those are separate issues, and we will perhaps need to touch on them separately. The residual powers will rest with the Treasury, but I agree with the hon. Member for Chipping Barnet that that is not right. I do not have anything against the Chancellor or the Treasury under this Government, but there are good reasons for not giving the powers to the Treasury. It is a major consumer of statistics, and so has a strong interest in how they are used, and it is also the funding Ministry, so its decisions will determine the effectiveness of the statistics board as a producer of statistics. It seems a better idea for decisions to be made by the Cabinet Office, but we will, no doubt, argue that point in greater detail.
It is certainly true, as the Minister said, that the Bill will create a much greater degree of parliamentary accountability, and there is scope for a good deal of debate about where that parliamentary accountability should lie. As things stand, I imagine that the board would be subject to scrutiny primarily by the Treasury Committee. Although that is an admirable body—I was a member of it for a short time—it deals purely with economic matters, and many of the issues that the statistics board and the chief statistician will be concerned with are non-economic issues. There is a strong argument for creating a system of parliamentary accountability that goes way beyond the Treasury Committee and makes use of the resources of both Houses of Parliament.
With those qualifications, I welcome the move to independence. I particularly welcome the indication that the Minister gave in his opening statement that he will be flexible and will listen. I hope that he will accept that, as the professionals suggest, there is need for substantial amendment.
5.45 pm
Fiona Mactaggart (Slough) (Lab): I welcome the opportunity to contribute to our debate which, everyone recognises, hinges on the important issue of trust in official information. We are self-obsessed in believing that it is the actions of politicians that undermine trust in official information, because the biggest factor in undermining that trust is whether the information is accurate. I welcome clause 7, which makes clear the Bill’s objective of
“promoting and safeguarding...the quality of official statistics”.
It refers not only to impartiality, which we have discussed at great length, but to
“accuracy...relevance, and coherence with other official statistics”.
I want to illustrate that important issue with a local example. I counsel parliamentary colleagues against falling into the trap that the media have dug for us by implying that the only reason why politicians wish to become involved with such things is to spin and fiddle them. I do not believe that that charge can be levelled against most politicians, whatever their party. We all do ourselves a disservice by conspiring in that debate, as it is a big mistake to believe that that is why Governments want to do things with statistics.
Mr. Newmark: Is the hon. Lady honestly saying that Health Ministers do not try to spin figures to suggest that things are better in the health service? Is she saying that the Home Office is not trying to spin criminal statistics to suggest that things are better than they are? People look at the reality on the ground, because their own personal experience does not match politicians’ announcements. That is the problem, and that is the cause of the lack of trust.
Fiona Mactaggart: The hon. Gentleman is right to say that a lack of trust arises when people’s experience differs from the figures. That is my point. He says that the difference arises only because politicians make up facts, but he is wrong. In fact, in many cases statistics are not accurate, and the most undermining factor for any statistical series arises when it is simply wrong. We need to ensure that the Bill addresses that issue properly.
Julia Goldsworthy (Falmouth and Camborne) (LD): Does the hon. Lady agree that it is not only the content of statistics but the timing of their release that undermines trust? For instance, a spate of reliable but damaging statistics can be undermined by the fact that they are all released at the same time—just before Parliament goes into recess for Christmas? It is not just the content of the statistics but their timing that undermines the trust of the wider public.
Fiona Mactaggart: The hon. Lady is doing something that I counselled hon. Members not to do, by assuming that politicians act in that way to bury bad news. I am sure that sometimes they do, but it is not the case on most occasions. Politicians on both sides of the House who conspire in assuming that it is the case most of the time undermine trust in politics. I believe, perhaps naively, that most people in politics, whatever their party, genuinely work for what they believe is the public good, and do not wish to connive in concealing the truth from the public. We are making a mistake when we imply that misrepresentation is widespread—the normal pattern of behaviour. I do not believe that it is.
What most undermines trust in official statistics is questionable accuracy. The Government will always get the blame. Let us take as an example a statistical series that most people do not believe is connived at or spun by Ministers. I refer to the census and the mid-year census statistics, which are more overseen by Parliament than any other statistical series that I can think of.
In August the leader of Slough borough council said:
“By now most people realise that the Government continues to underestimate the population of Slough”.
In his new year message he wrote:
| Next Section | Index | Home Page |