Skip to main content

London Population (ONS)

Volume 463: debated on Wednesday 25 July 2007

Motion made, and Question proposed, That this House do now adjourn.—[Mr. McAvoy.]

I might have cut things a little fine for my question this morning, but I have had ample leisure to repent while waiting for the debate this evening. I am grateful for the opportunity to raise an important issue for many Londoners.

It is appropriate to raise population statistics compilation in London with the Exchequer Secretary, who earlier today dealt with the motion about the chair of the Statistics Board. There is a link. In the earlier debate, several hon. Members referred to the general population’s lack of confidence in official statistics. I believe that the hon. Member for Slough (Fiona Mactaggart) made the point that one of the biggest causes of lack of confidence in official statistics is that they are sometimes demonstrably plain wrong. That is exactly the problem with the methodology that produces the official population statistics for London.

London Councils, which has been most anxious to hold such a debate, compiled compelling evidence that the methodology that the Office for National Statistics used to calculate the population of London is seriously flawed. Consequently, the official population statistic significantly understates the capital’s population. London Councils has repeatedly raised the matter with the ONS and the previous Treasury team of Ministers without success—hence its desire for a debate, and I am happy to do my best to oblige.

Lack of confidence in the London population statistics is not new. Westminster city council eventually successfully challenged the population estimates for its city and borough in the 2001 census, which had to be revised upwards after manifest errors were discovered. It was able to point out precisely what had gone wrong. Regrettably, the position has not improved. For reasons that I shall describe, the current methodology of the ONS is flawed and it now proposes to move to a different methodology, which is equally flawed, as can be demonstrated.

It may seem an arcane point, but official population statistics are important. They form part of the basis of important forward planning, such as where more homes, schools and health care facilities should go, long-term policy decisions about, for example, where new transport infrastructure should go, and operational decisions about police numbers for each borough and fire service cover. The risk assessment includes population levels and it is therefore important that the statistics are accurate. They are also important because they are part of the formula to allocate Government grant to councils for local services. If they do not reflect where the people who use the services are, resources will not be directed correctly. That leads ultimately to waste. The concern of London Councils, therefore, is that the funding cake for local government should be divided up on the basis of accurate figures, whatever its size. If it is not, the risk is that London might lose out unfairly. I should like to address that issue in a little more detail.

In summary, when London Councils compared the theoretically based estimations of population with detailed work that had been done on the ground in a number of London boroughs, it was able to demonstrate serious underestimation, in particular of the amount of in-migration to the capital. For example, the evidence shows an underestimation of 3,300 people in Enfield and of 2,000 people in Brent, as compared with the number of people who are demonstrably there on the ground in those boroughs. In Newham, it was possible to demonstrate that there were 750 more schoolchildren than were “officially estimated”—the key point is that, unlike the estimates, the schoolchildren are actually there. Similar demonstrations of inaccuracy have been made in Croydon. The problem is not just a London problem. Outside London, Slough has demonstrated an under-calculation of 6,000 people in its population.

I am grateful to my hon. Friend for mentioning Croydon. The borough has also seen a downward revision on the various ONS calculations and assumptions. However, that has come at a time when the number of migrants registered with GPs was more than 37,000 in the five-year period to April 2007 and when the number of national insurance numbers given out from 2002 to 2006 was almost 18,000. I also know from correspondence with the Jobcentre Plus office that national insurance number registrations accelerated to 7,000 in the first 11 months of the following financial year. Places such as Croydon are suffering from problems in their local government settlement, such that the rate at which they are delivering universal services—that is, non-social services and non-education provision—is only a quarter of the rate for other London councils. Does my hon. Friend agree that the figures for migrants are having an impact on the local government financial settlements that councils enjoy?

They certainly are. My hon. Friend makes a valid point. It is not good enough for the Government to say, for example, that migrants frequently move on quickly. That is not demonstrated by the evidence on the ground and nor does it deal with the point that, for a period, migrants live in London, which is often the first port of call. While they are there, they are using London services, and the cost of that has to be met in one way or another.

It is fair to say that those shortfalls are recognised by other responsible bodies, not just by the London boroughs. I remind the Exchequer Secretary of a piece in The Guardian in April, in which the Home Office Minister responsible for immigration said that

“at the very least the Office for National Statistics needs to improve its figures on which key local financing decisions are based, but it also means the tough enforcement of immigration laws, including the prosecution of employers of illegal migrants.”

That is an acceptance by one of the Exchequer Secretary’s colleagues in Government that the ONS needs to improve its act. Referring to the current method of calculating in-migration, which is based on the international passenger survey—a small sample survey of people arriving—the Governor of the Bank of England said last November:

“A survey that was designed to learn more about tourism and business travel is not the best source of data from which to learn about migration.”

That is a pretty unanswerable point. Lord Bruce-Lockhart, the chairman of the Local Government Association, has said:

“It is essential that the government gets the figures right on migration. That is currently not happening and local authorities are suffering as a consequence. Councils are finding it difficult to provide services to growing populations that are not recognised by government statistics.”

The Mayor of London and the London assembly have both complained about the inadequacy of the figures going back to the 2001 census. The Mayor has commented on the fact that the Lisson Grove estate, which contributed substantially to his majority when he was a Member of this House, seemed to have disappeared from the census calculations. It is not often that I agree with the Mayor of London, and I could not agree with him on that one.

Does the hon. Gentleman accept that this is not a matter of party political tension? It is a shared view across the parties and among independent people across Greater London. Is he also aware that in constituencies such as mine there is a turnover of those registered of between 25 and 40 per cent. a year? The reality is that, unless we have a five-yearly census or, even better, an annual opportunity to count, we will never begin to catch up and accurately measure the numbers in a way that can be reflected in an accurate apportionment of grant by central Government to local government.

The hon. Gentleman is absolutely right. There is consensus on this between all three parties in the Association of London Government, and this view was endorsed by its leaders committee in its most up-to-date report only a few weeks ago.

I have set out why this matter is important for the levels of funding for personal services, adult services and children’s services, the capitation formula for primary care trust funding and the dedicated schools grant. All those issues and more depend on getting this right. The criticisms have been well set out and, with respect, it is not good enough for the Government to retreat into the position that they and the ONS have so far adopted, saying that they have used the best available statistics. In this case, the best available statistics are demonstrably wrong.

The final piece of evidence that I should like to pray in aid is a survey by the City of Westminster which shows that some 40,000 new national insurance numbers were issued to residents of Westminster between 2002 and 2004. That represents a 17 per cent. increase on the 2001 base population. It might shock hon. Members to learn that that survey also shows that about 20,000 people arrive at Victoria coach station every week, predominantly from eastern Europe. For at least some of the time after they arrive in the UK, they will be in London.

The movement from the international passenger survey with the labour force survey will not, regrettably, make things better. It is suggested that the labour force survey got nearer to the estimate than the IPS in 2001, but if we go back to 1991, it was the other way round. So there is no guarantee that it is better. In fact, the labour force survey sample is even smaller than the sample used for the IPS. That is crucial, especially as we are to move to three-year local government finance settlements, starting in 2008. It is therefore all the more important that we get this right to start with, because the problem will be much more difficult to unpick if we are locked into a three-year settlement on the basis of inaccurate and unreliable figures.

The fault here is that the two methods fail to pick up short-term migrants, to whom I have already referred, or people who live in London for part of the week but are counted as resident elsewhere. Many people live in London for part of the week for their work, not a few of whom are connected to this honourable House, but their family home is regarded as elsewhere, so that is where they live for the purposes of the official population statistics. However, while they are here they use London refuse services, leisure services, street cleansing services, and so on. That is not recognised.

The figures do not take into account either the very real and well documented effect of population churn in large cities such as London. In the City of Westminster and some other central London boroughs, population churn of about a third—up to 34 per cent.—is well documented. That in turn creates particular costs, because it involves one set of short-term migrants being replaced by another. They are not picked up by the official statistics, but they are nevertheless using local services. The ONS has said that it will publish estimates of short-term migration later in the year, but that will be too late for the start of the three-year financial settlement in April 2008. Much more urgent action is needed, which is why I have raised the issue in the debate tonight.

I have set out in some detail what is wrong. London Councils does not want the Government to wash their hands like Pontius Pilate and say that it is all down to the quango—the ONS. The Government have responsibility. What we need to do is accept that current estimates are not working and are not fit for purpose, and accept that there is a lack of clarity and transparency. It is reasonable to ensure that we develop a methodology that is robust, up to date and fit for purpose. There must be proper consultation with local authorities before it is changed—that has not happened so far—and minimum standards of accuracy should be set out on a basis that can be agreed between the Treasury, the ONS and local authorities. Perhaps the incoming chair of the Statistics Board could, with a little push from Ministers, take that into account.

If only the Government and the ONS would listen to councils in London, accept what is happening and come up with a solution to restore both fairness and public confidence. That is the reason for this debate at this late hour—late in terms of the time and of the opportunity. We must get things right before the funding settlement cuts in. We contend that the Government cannot abdicate their responsibility on this issue. Ultimately, ONS has to be accountable. So far, regrettably, it has put its head in the sand in the face of compelling evidence. What is needed is for the House and the Government to put pressure on it to accept reality and change things in available ways while there is still time to do so.

First, I congratulate the hon. Member for Bromley and Chislehurst (Robert Neill) on securing this evening’s debate. I thank him for raising the important issue of the compilation of statistics on London’s population. The hon. Gentleman explained that this is such an important issue because ONS population statistics are used to determine local government funding. Equally, they are used to inform local planning in respect of services for the future.

It is precisely because of the importance of these statistics that ONS takes a great deal of care in making them as accurate as possible. ONS is currently involved in ongoing work to improve statistics. I will come on shortly to describe how those statistics are put together, and explain why some of the suggestions about flaws in the methodology are inaccurate.

I would like to start, though, by talking briefly about how these statistics are used, because the impression is sometimes created that local government funding depends entirely upon population statistics. I will not describe in detail the funding formula used for local government—we are all familiar with it, especially with how it impacts on our own areas—not least because it is the responsibility of the Department for Communities and Local Government rather than the Treasury. However, I will say that it takes account of other factors, such as local authorities’ ability to raise revenue from council tax, alongside the population statistics, which are clearly important in themselves.

There is also a process known as damping, which is used to ensure that all local authorities receive a minimum increase in grant, known as the floor. That floor means that even if population statistics were to show the population decreasing in an area, the local authority’s funding would still increase by a reasonable amount.

Nevertheless, population statistics are clearly important, and they should be as accurate as possible. As well as being used for local government funding, they are used to develop and monitor economic and fiscal policy and to understand social change, as well as any policy implications that stem from it. These statistics are calculated as accurately as possible, using a clear and carefully designed methodology.

One of the challenges for that methodology—it is only part of the task—is to factor in the impact of international migration on population levels. The hon. Gentleman has raised that matter in the House this evening. Migration is largely calculated on the basis of the international passenger survey. We have heard a number of things about the lack of accuracy in the IPS, but I should like to draw attention to areas where it is reasonably robust.

First, it has been claimed that IPS is restricted to the major airports and that it does not include Victoria coach station, for example, where 20,000 people can arrive. Actually, it sampled 16 airports, 11 seaports and the channel tunnel. Clearly, it does not sample Victoria coach station, but that is simply because coaches from other countries cannot get there without passing through a port or the channel tunnel. The IPS picks up the inhabitants of the coach on its way through to Victoria coach station.

Secondly, it is not the case that the international passenger survey’s sampling is carried out only during the working day, as critics have said. There is flexibility in when the sampling can take place, especially to include particular flights from particular countries that may arrive early at an airport.

Will the Minister accept, however, that she is missing an important point? It is important to have accuracy of statistics not just on who is coming into the country through the entry points, but in terms of calculating, for example, local need and grants when people are here. The IPS does nothing to assist in that, for the reasons that I set out.

The hon. Gentleman is right, and that is why the labour force survey offers extra insight, although I would be the first to say that it is not an actual count of all migrants wherever they are. Even if we could do that and afford to do it regularly, it would still provide only a snapshot of the situation, which, as he said, can be fluid.

The IPS is more comprehensive than is often suggested, but the ONS is committed to ensuring that population statistics are as accurate as possible. It has been working this year on improving its methodology. Plans to do that were published in April. The ONS has held meetings with a number of users, including some London local authorities, to discuss planned changes. Those meetings have led to refinements to the proposals, which were outlined in material published yesterday. There have been workshops and feedback has been given. I expect the dialogue to continue as changes to the methodology to produce more accurate figures are put into effect.

The improvements that the ONS is making to its methodology make the information more accurate and improve quality. In particular, migration estimates will now draw on the labour force survey in addition to the IPS, which will help to reflect what migrants do after they have entered the country and where they settle. There is some evidence that the IPS may be particularly inaccurate for well-known places such as Westminster, Oxford or Manchester because people often fill in a place that they have heard of when asked about their destination, and that is not always where they end up living.

Again, concerns have been raised about aspects of the labour force survey, including the suggestion that it does not take account of people in multi-occupied dwellings. Although it does not take account of people in communal establishments, such as halls of residence, it does cover all types of private households, including those that are in multi-occupied dwellings. While response rates to the survey for multi-occupied dwellings are lower than those for single household dwellings, the weighting of the sample to take account of age, gender and region goes some way to correcting that. There have been improvements and a process of ongoing dialogue with local authorities, including London local authorities. I expect there to be an ongoing liaison.

The improvements in the methodology improve in particular the regional and local distribution of international migrants in the statistics that the ONS produces. Combined sources of such statistics have strengths that increase the robustness and the likely accuracy of the results, rather than taking the results from a single survey. The improved methodology will now be applied to revise population statistics for 2002 to 2005 and to produce 2006 mid-year population estimates, all of which will be released in August.

The improvements will not change the trends demonstrated by the existing methodology’s estimates. In particular, figures based on the improved methodology will not show either population or migration levels decreasing. Across England and Wales, the original method showed an increase in population due to long-term international migration. The improved methodology increases that slightly, by 28,000. For London, the existing method implied an increase of 396,000 between 2002 and 2005 as a result of long-term international migration. The revised method reduces that increase to 336,000, but that is not the same as saying that migration is decreasing, and it is certainly not saying that the population is decreasing. Instead, the revised figures show that migration to London remains on an upward trend, but at a lower level than the estimates suggested.

I should re-emphasise that the changes are being made to improve the methodology and accuracy. As well as making the statistics more accurate this year, the ONS has planned further improvements for subsequent years. In particular, an interdepartmental taskforce on international migration statistics was set up in 2006 and has made a number of strategic recommendations for further improvements in the years to 2012. The ONS is already taking forward some of those recommendations, and will publish a full response later in the summer.

The ONS also published a revisions policy earlier this month, setting out the principles for further revisions to population statistics, and it has been working with local authorities and Government Departments to identify how new and existing information sources—such as GP and school registrations—can be used to inform migration estimates and address local issues in population estimation. That goes to the heart of what Members have been talking about.

However, that is not as easy as it might seem. The statistics published yesterday on national insurance numbers issued to overseas nationals caught the media’s attention and have been mentioned in our debate, but basing population estimates on national insurance numbers could be misleading, as some people apply for them but do not use them, while others need them only for a short period. In other words, national insurance numbers provide an inflow figure, which is not necessarily a net figure, and many of those who applied for national insurance numbers might have already left the country. National insurance numbers do not show the flow out; they only show the flow in.

That brings me on to my final point, on those who come to the UK only for a short period. The ONS population statistics that I have talked about rely upon the UN definition of a long-term migrant—someone changing their country of usual residence for at least a year. We believe that that is the most appropriate and reasonable definition to use and the most realistic way of accurately determining the population levels in different areas. We recognise that short-term migration is also important, as is churn, but such trends are difficult to pick up on accurately and in a timely fashion without spending one’s whole life getting snapshots of an ever-changing situation. That is why I am particularly interested in discovering how we can use other indicators and statistics to get a closer handle at a more disaggregated level on what is happening.

We believe that the annual definition is the most appropriate one to use, but the ONS has been working hard on producing innovative estimates, and will publish national figures later in the year. However, this is a new and difficult area and we cannot expect accuracy immediately as we feel our way towards a solution.

The Minister suggested that there might be other indicators. I mentioned GP registrations as an example. What other possible indicators might be useful?

There are several such indicators. The hon. Gentleman mentioned GP registrations. There are also school registrations, but that does not address the short-term or long-term issue. Each attempt to measure accurately at a disaggregated level must be looked at carefully for what it actually tells us. The national insurance numbers demonstrate only flows in; they do not give us a picture of the net situation, nor do they tell us about flows out. They can be more misleading than useful, even though they have been prayed in aid tonight as indicators.

Each attempt to take a look at what is happening in local communities must be analysed to establish its accuracy and its shortcomings as well as the information it provides. The plea I make is that that is what we all have to do as we try to feel our way to a more accurate method of measuring at community level precisely what is going on in a dynamic and open economy.

This has been an important debate, and I am grateful to the hon. Member for Bromley and Chislehurst for raising it. I look forward to there being ongoing dialogue as the ONS attempts to improve its methodology so that we can have more accurate and better policy.

Question put and agreed to.

Adjourned accordingly at twenty-four minutes past Eleven o’clock.