| Previous Section | Index | Home Page |
“Many Londoners believe Muslims get more than a fair deal”
—does she understand from it that 77 per cent. of those surveyed did not agree, and that 23 per cent. were described as “many”? That is the kind of misleading start to information that stops people reading the figures, and allows Ministers to get away with what they should not get away with. It is an argument for figures being announced first and then Ministers trying to put a spin on them, rather than the other way around.
Kali Mountford: The hon. Gentleman has made an important point, with which I was about to deal. It involves complexity, and questions that are often asked. It is wrong to talk of statistics as though they were produced in a value-free way. The question that is posed in the first place is often the real problem—the crux of the matter. The question that is posed, and from which the statistician then analyses the results, is the question, rather than the information that comes out of that.
I have been looking at a set of figures on how we can trust statistics, and it was difficult to understand. I will not bore Members by explaining why it was difficult, as it would take an hour for me to analyse that set of statistics about statistics, but it is broken down in
several ways that are relevant to the question that the hon. Gentleman posed about what people might or might not believe is preferential treatment for a particular ethnic or religious group. If questions are asked in a particular way—a way that has a sliding scale—complexity is introduced. In the current case, that complexity is about whether one comes to an issue with a preconception about statistics. There is then also a question about trust and then, added into that, questions about one’s age, political opinions, background, where one lives and one’s other interests. Therefore, all the issues surrounding a question become increasingly complex.
Unsurprisingly, at the end of that process, whether a Minister has had sight of such information in advance almost becomes immaterial, because the information coming out of it is almost impenetrable to the average person. If we look at a graph that is as simplified as possible, it can still be very difficult accurately to discern what we are meant to understand from a set of figures.
Let us look again at the question about whether people think that a disproportionate amount of resources is spent on a particular ethnic or religious group in some part of London, and whether they think that that is fair. It is sometimes difficult to understand whether the question was fairly put, whether the nature of the resources spent is properly understood by the participants in the survey, and whether the Minister who asked the questions in the first place knew exactly what questions were being posed. Whether the Minister knew the answers before they were published is another question entirely. In any event, under this Bill, none of those questions would produce figures that would be national statistics for the board.
Sir Robert Smith: The hon. Lady says that, in respect of a complex set of data, it does not matter that the Minister has 40 hours to go through that, but in fact that gives the Minister a huge advantage because they can then pick the most eye-catching part of, or way of presenting, that data, and the easiest way of steering the media in one direction—and the media have deadlines to meet. It can then take ages for everyone else to get into the detail of the data, with the Minister having had the unfair advantage of a 40-hour head start.
Kali Mountford:
I am sure that the hon. Gentleman will intervene on me if what I say is incorrect, but he seems to be suggesting that Ministers should not have early sight of data. This was not going to be the subject of my speech, but let me say that I do not know of any European country where Ministers do not have early sight of data. There are good reasons why Ministers would want to see at least some data before publication. Having commissioned data, it would be bizarre if they were unable to have sight of that before anyone else—if only to be able to think, “My goodness, the policy we were proposing has gone badly wrong.” If things had been got badly wrong, we would like to think that the Minister concerned would at least want to be able to apologise to the House about that—or if things were a terrific success, to be able to claim credit,
or to say something else if the result were somewhere in between those extremes. It would usually be right for Ministers to be able to make some sense of any figures they might have had a sight of before presenting the figures to the House. I do not find that at all bizarre. That has been the practice in our political system not only over decades, but over the many years since statistics were first collected.
The most-used phrase about statistics was coined long ago, back in the age of Disraeli, so it is clear that the relationship between politicians and statistics goes back a long time. It would not be fair to suggest that it started in 1997, or even in 1979. We should try to get away from the question of whether Ministers can somehow massage figures in the 40 hours or five days or however long they might have. I do not think that that is the case.
I want statistics to inform public policy. That is the key. That is what statistics should be used for. I see that the hon. Member for West Aberdeenshire and Kincardine (Sir Robert Smith) is frowning, but that is what they should be used for. Some years ago, I was a councillor serving on Sheffield city council and I remember feeling completely nonplussed when I saw better financial settlements being secured for areas in the south-east that seemed to me to be extremely wealthy compared with the ward that I represented, which was the sixth most deprived ward in England. I could not understand how any economic indicator could result in some such wards getting more money than the one that I represented in Sheffield. That did not make any sense at all.
That brings me back to earlier remarks on how we determine that kind of policy. We have made changes, but it might be asked whether they were fair or based on accurate data, and whether the collation of data was properly conducted. That is what we need to get to the heart of—the proper process. Because there is a perception of interference, politicians want to defend their own constituency interests, which is right and understandable. The accusation is therefore made that, “You’re only making this change for political reasons, because you’re northerners representing northern Labour seats and you want to take the money away from the south-east because that’s where you think us Tories—or members and supporters of whichever other party—are. You’re doing this for political reasons.” Therefore, we must have true and proper data, so that not only is the right thing done but it can be demonstrated that the right decision has been made for the right reasons.
Yet what have we heard about in the debate? My hon. Friend the Member for Slough (Fiona Mactaggart) made some telling points, even about as simple a question as how we can know who lives where they say they live and how many people live in a particular town. It might be said, “Well, we’ll count the number of people who live there”, and that sounds easy to do, but it turns out to be very complicated. We cannot possibly think that Ministers try to interfere in how many people live in a particular house on a particular day, but it turns out that even that is a complex question. Therefore, although I, along with all Members present, welcome the Bill and want it to progress, I also want the Government to go further—if not in this Bill, then later—because more can be done to make statistics more accurate and
dependable. That should not be done just to remove the perception of political interference; there is more to this point than that.
We need to look not only at the guidance for use of statistics or the acceptance of statistics, but also at the national guidance for statisticians and the integrity of statisticians. I cast no aspersions on statisticians in saying that. When I was a civil service research officer, we worked in teams, and I believe that that practice has been developed in the time since I left the civil service. I asked some questions earlier about interdisciplinary teamwork, and it is important that teams work together—that they collect data together in an interdisciplinary way, and properly research what is happening so that we understand what people are doing in their lives and make policy decisions based on what people really do, rather than on a snapshot view on one day when we do a simple head count. That is not enough to know how people are experiencing their lives.
Earlier in the debate, a Member talked about understanding deprivation, and he seemed to have already made a judgment about what he thought might be an underlying cause—the breakdown of family life. That judgment seemed to be based on the kind of feeling that we all sometimes have: we feel that crime or family life might be better or worse. Sometimes, however, we have actual data, and sometimes those data are counter-intuitive, so we say the data must be wrong, because they go against our gut-instinct. However, it is not wrong just because it goes against our gut instinct. Such data cannot be a snapshot on one day; they have to extend across a period of time. A proper study of family life would have to extend over at least a decade; indeed, we would have to follow families over decades to see what actually happens to them, and have proper cohort groups. We would need to work across Government Departments, and we would probably have to involve other, non-governmental agencies. The statistician’s work would have to inform at the outset how such a study should be set up, and other researchers, such as economists and social scientists, would have to be involved.
We should not examine just a single moment in time at which we thought we knew what was happening. We should not undertake a simple count, saying, “Last year, x number of marriages broke down, y number of people got married and z number got divorced”, without looking at whether the marriage rate or the divorce rate had gone up over the previous decade. We must examine the whole context to see whether children were thriving in spite of, or because of, divorce, or whether family counselling had an impact on family life.
Given that perception is often regarded as reality, the independence of the board and all the other elements in the Bill will go some way toward building trust and might have some impact on public confidence. However, if the statistician’s work is to be meaningful and really to inform public debate on policy, I hope that we will at some stage go further and examine how the statistician can inform policy in a more meaningful way inside Departments. I also hope that the role of the statistics board in examining the standard of the statistician’s work will assist Departments in informing public policy.
7.22 pm
Stewart Hosie (Dundee, East) (SNP): The hon. Member for Colne Valley (Kali Mountford) seems to have a touching faith in the integrity of statistics and of Ministers, who never spin at all. The difficulty that I have arises from a letter that the then Tory Secretary of State, Ian Lang, wrote to the then Prime Minister in 1992, when the general expenditure revenue Scotland statistics, which I shall refer to later, were being developed. The then Secretary of State said that he judged that it was just what was needed at present in the Tory campaign to maintain the initiative and undermine the other parties, and that it could score against all of them. GERS was designed to be a political tool, not to offer an enlightened or accurate description of the Scottish economy. I suppose that the great shame—this is the last political point that I shall make in this speech—is that the Labour party continues to use GERS to this day for exactly the same purpose as the Tories did in 1992.
However, I give a broad welcome to the Bill and to some of the Financial Secretary’s opening comments. I particularly welcome what was said about pre-release access, which is to be put on a statutory basis, but I shall listen carefully to the argument that the board may wish to set the rules regarding such access. I remain concerned that the 40-hour period is far too long. The idea that a Minister and his team cannot in 40 hours find one little nugget to spin or one bit of obfuscation to hide a bad bit of news is clearly wrong.
I also welcome the announced creation of a publication hub to separate the communication of statistics from political comment on them. If the Government hold to that, it will make an enormous difference in terms of the public perception of statistics and the way in which they are launched. Too often reports are leaked in advance, or the political comment on them immediately thereafter is given a particular spin. As was said earlier, by the time that the journalists and the rest of us have got into the guts of the statistics, the story has gone and a perception has been left.
However, in general terms and for the sake of good policy making, it is important that statistics are produced and delivered in an independent and impartial manner; fundamentally, that is what statistics are about. In the case of Scotland in particular, it is time for the figures that tell us so much about the success or otherwise of policy to move outwith the control of Ministers. Of course, we ultimately wish to see an independent Scottish statistics board, but we very much welcome the move that the Government have made through the Bill. Our primary concern is the future structure of statistics: we want adequate, independent and entirely impartial statistical information, which presents a true picture of Scotland at all times, to be available. All such statistics are essential in developing and measuring policy, in assessing need, in measuring the economy and in almost every other Government and local government decision-making process.
Furthermore it is vital that the public, media, voluntary organisations and society as a whole have access to independent impartial statistics in which they can have complete trust. Independence is essential to ensure that there can be no implication of manipulation, selectivity or politicisation in the
production, publication and dissemination of statistics at any point in the future—which brings me nicely back to GERS, because that was its sole purpose.
We have heard from many people that public trust in statistics is low—they had their own different reasons for saying so—and that some publications, including GERS, have been heavily politicised. I have no doubt that the independent board proposed in the Bill would have given short shrift to the Tory attempt, and the subsequent Labour attempts, to use partial and incomplete figures to make a narrow political case. I want to see the removal of any suspicion from our statistics, and to go further by guaranteeing that they are based on real evidence. We must therefore make certain that the Bill as drafted really is sufficient to deliver confidence in the statistics published.
I shall use the GERS paper as an example of the difficulties that we face and of the problems that must be overcome. GERS is supposed to be a snapshot of the Scottish economy two years in arrears. It is supposed to look at total income and total expenditure and work out the balance sheet for Scotland—and I shall quote from it to show just how inaccurate the process is. The net borrowing figure—the deficit included—in GERS is based on expenditure and receipts calculations that by necessity include some estimation. GERS says:
“It should therefore be used with some caution”.
It fails to include a single penny piece of revenue yield from the Scottish sector of the North sea. It is a wholly misleading figure.
On income tax, GERS uses the survey of personal income, yet for the 2004-05 statistics it used the 2003-04 survey of personal income data, because the 2004-05 data were not available. Nor is the calculation based on a known and fixed taxpayer population, so it is wrong in two regards.
On VAT, Scotland’s share of UK VAT revenue was estimated on the basis of Scotland’s share of household expenditure on goods and services subject to VAT, as estimated from the expenditure and food survey. GERS describes this in the following manner:
“The results should be treated with caution since they are based only on household expenditure estimates and not the share relating to the amount of VAT received from businesses registered with Scottish VAT offices or received from businesses trading within Scotland.”
The landfill tax has been allocated on a population share basis. The environmental data indicate that Scotland has more landfill per head than the rest of the UK, yet no geographical breakdown of the tax data is available. Again, this method may underestimate the revenue yield from that tax for Scotland.
On corporation tax, GERS says:
“This tax is exceptionally difficult to estimate for Scotland, due to both conceptual difficulties and a lack of data, therefore, the estimate of Scottish corporation tax should be treated with extra caution”.
In 2003-04 and in 2004-05, GERS allocated £2.4 billion in corporation tax, yet we know that in that latter year, the Royal Bank of Scotland alone paid £2.38 billion in corporation tax. We also know from a survey of 27 of Scotland’s top 50 companies in the same year that they
made a profit of £14.794 billion, which would have generated some £4.3 billion in corporation tax. We know from The Scotsman last week that last year Scotland’s 500 biggest companies scored a combined record of £23 billion profit, which would have generated a corporation tax yield of somewhere around £7 billion. So we know that the corporation tax figure is wholly wrong and cannot be trusted.
The principal conclusions in GERS are described as deriving a fiscal position for Scotland. They are
“subject to inevitable imprecision due to the need to estimate a number of elements of both expenditure and revenue. The calculation of expenditure for Scotland (specifically the non-identifiable and other expenditure components) cannot be carried out with the same accuracy as that for the UK as a whole.”
It is such inaccuracy that I hope the independent board and the Bill will help to resolve.
On the issue of expenditure I have several examples, because this point does not affect Scotland alone, as I shall explain. The non-identifiable expenditure allocations for Scotland include many things, such as, in 2003-04, a Home Office allocation of £2.521 billion from English prisons and offender programmes. The figure included as non-identifiable some millions of pounds that went to the English Tourism Council and the Greater London authority from 2001-02 to 2004-05. It included as non-identifiable inward investment grants made to English regional development agencies by UK Trade & Investment between 2001-02 and 2005-06, and the Courts Service spending by the Department for Constitutional Affairs of £547 million—half a billion. In an average year, more than £3 billion of English only spending—rightly spent in England alone—is allocated as non-identifiable. That has the impact on GERS of overstating Scotland’s expenditure by £262 million and understating English identifiable expenditure by more than £3 billion.
We want impartial and accurate statistics that present a true picture of Scotland at all times. They cannot just be impartial; they must be seen to be impartial.
I have some concerns about specific clauses. It appears to me—I suspect that my perception will be shared outside the House—that there will still be Treasury control and interference. Under clause 3, Scottish Ministers cannot directly appoint a member to the board: that can be done only by the Treasury. Under clause 27, Scottish Ministers cannot direct without the approval of the Chancellor. Under clause 45, Scottish Ministers cannot authorise disclosure of information to the board without Treasury approval. Under clause 49, Scottish Ministers cannot authorise disclosure by the board without Treasury approval.
On certain occasions, such as in the case of cost-sensitive or highly sensitive market information, such Treasury control may be appropriate, but it cannot be so in every case, and I intend to table amendments to those clauses. They may be probing amendments, but I may have to press them further.
The Bill contains much to welcome, including the separation of communication and political comment, the recognition of concerns about pre-release access and— notwithstanding my criticisms of enshrined Treasury powers—the ability at least of the Scottish
Executive to order statistics from the new board and have their efficacy guaranteed. We will not vote against the Bill tonight—I suspect that there will be no votes at all—but it raises real issues, to which we will return at a later stage. I look forward to hearing the Minister’s comments when he winds up.
| Next Section | Index | Home Page |