London's population and the 2011 Census - London Regional Committee Contents

2  The 2001 Census in London

  • As a national exercise, the 2001 Census was deemed broadly successful.
  • Estimated rates of non-participation in the 2001 Census were disproportionately high in London. The average response rate across all London boroughs was 85%, compared with 94% nationwide. In Inner London boroughs the estimated response rate was 78%.
  • Several London boroughs have told us that they considered the 2001 Census to have undercounted the population in several categories considered hard-to-count, in particular among ethnic minority groups.
  • There is evidence that the methods used to ensure an accurate count in 2001 were not equal to the challenges faced in some inner-city areas. A substantial number of these areas were in London boroughs.
  • The ONS acknowledges that London has "pretty much the full set of challenges" for preparation of an accurate address register for the 2011 Census.

Issues arising from the conduct of the 2001 Census

20. The Statistics Commission, the body which monitored the arrangements made for the 2001 Census, concluded in its report on the 2001 Census in Westminster that "the 2001 Census was, in most respects and on the evidence available, a success. It produced robust local estimates across most of the UK. But for a relatively small number of areas, particularly some inner city ones, there is now evidence that the methods used were not equal to the challenges they faced."[13]

21. It is acknowledged that several of the areas identified by the Statistics Commission were in London boroughs. The City of Westminster and the London Boroughs of Southwark and Wandsworth challenged the outputs from the 2001 Census and eventually secured an official revision to the figures reported in each borough.[14] Other boroughs argued successfully that their populations had been underestimated and secured additional funding for borough services as a result.[15]

22. Though he acknowledged shortcomings in some local authority areas in his evidence to us, Glen Watson, ONS Census Director, argued that the 2001 Census nationally had been broadly successful:

    "Most people considered the Census in 2001 to be a success. We counted 49 million people and coverage-adjusted—through the Census Coverage Survey—another 3 million people […] It was in the series of challenges that followed that we adjusted the population of 52 million upwards by 0.3 million. So that was very important for some local authorities—most notably for Manchester and Westminster—but for the vast majority [of local authorities] the results stood, and they still stand, and for a few more the results were adjusted by relatively small amounts. I do not want to play down the significance of the problems experienced in parts of London, in particular Westminster, but I want that context to be understood."[16]

23. While we acknowledge the overall success of the 2001 Census exercise in counting the population of England and Wales, it is important not to lose sight of the serious and acknowledged issues which arose in counting the population in London. We recommend that particular issues with a potentially serious effect on the accuracy of the 2011 Census in London are properly addressed.


24. The ONS, in its evidence to us, acknowledged that the most significant likely source of error in the 2001 Census in England and Wales was non-response or under-enumeration. In 2001, ONS estimates that 6% of the population did not respond to the Census. Although only a small non-response rate when compared with national government surveys, this non-response rate to a compulsory survey was greater than in 1991, when 4% of the population did not respond.

25. The estimated rates of response in London were disproportionately low as a result of the particular challenges facing the Census in London, which we discuss further below. As a result of this acknowledged low response rate in London, which was a contributory factor in significant undercounting, 2001 was the first Census after which data from a census coverage survey was used to adjust the final results to take account of under-coverage.[17]

26. The ONS told us that the response rate to the 2001 Census was significantly lower in Greater London than in the rest of the country. The average response rate across the whole of London was 85%, compared with 94% nationwide.[18] This was even lower in certain boroughs, especially in the inner city, the lowest of all being Kensington and Chelsea, where the 2001 Census response rate was around just 65%.[19] Several other inner London boroughs also told us that they had significantly lower Census response rates in 2001 than the national norm—Westminster had a 74% response rate, and Hackney 72%.[20]


27. The Census seeks to count all of those who are "ordinarily resident" on Census night. In 2001, it did not count short-term migrants (that is, people who move to a country other than that of their usual residence for a period of at least three months but less than twelve months). Several of the London boroughs which submitted evidence to us suggested that one of the flaws of the 2001 Census from their perspective was that it did not count their numbers of short-term migrants.[21] London is a traditional entry point for the majority of new migrants arriving in England and Wales, and its boroughs have disproportionately higher numbers of short-term migrants than other parts of the country. London also experiences significant short-term population churn through internal migration.

28. Several of the London boroughs told us that they considered the 2001 Census to have undercounted some of their harder-to-reach ethnic minority groups. Evidence supplied by Lambeth, for instance, suggested that the second most important factor associated with a higher Census non-response in 2001 was whether households consisted of occupants who were of the Black, Asian, Chinese or Mixed ethnic groups.[22] Witnesses also expressed concern that the 2001 Census had suffered from a low response rate in London boroughs because of the inability of enumerators to access all addresses.

29. The unique make up of London's population makes it particularly difficult to enumerate in a Census. London has a high concentration of those in hard-to-count categories. In evidence to the Committee, the ESRC's Census Study Programme stated that "London presents instances of extreme enumeration challenges."[23] Census Director Glen Watson acknowledged that "London has pretty much the full set of challenges" for preparing an accurate list of residential addresses for use in the Census.[24]

30. One recent assessment of hard-to-count categories in the Census lists the following groups as particularly difficult to reach via the Census:

  • renting privately;
  • where the occupants are of Black, Asian, Chinese or Mixed ethnic groups;
  • paying part rent/part mortgage;
  • containing a single person;
  • where the average age of the people within the household is between 23 and 29 and 29 and 34;
  • where the average age of the people within the household is 70 or over;
  • renting from housing associations or the council;
  • where more than two thirds of occupants had a different address one year earlier;
  • living in commercial buildings;
  • in an area with a higher Index of Multiple Deprivation Income score;
  • where the average age of the people within the household is between 60 and 69;
  • living in accommodation that is not self-contained;
  • living in a converted or shared house;
  • where the average age of the people within the household is under 23;
  • containing a single parent family; and
  • where more than two thirds of occupants aged between 18 and 29 are students.[25]

  • There have often been substantial discrepancies between the official population estimates derived from the 2001 Census baseline produced for London boroughs and boroughs' own estimates of their resident populations.
  • Boroughs argue that this lack of precision about the size of their resident populations has had a substantial effect on their ability to plan and provide resources for the public services they are required to deliver.
  • Evidence suggests that ONS estimates have failed to take into account a temporary—and therefore largely uncounted—population.

Mid-year population estimates

31. The Committee heard that one of the consequences of the 2001 Census's inability to fully record numbers of short-term migrants and ethnic minorities has been a substantial discrepancy between the official population estimates produced for London boroughs, derived from the 2001 Census baseline, and boroughs' own estimates of their actual resident populations drawn from a variety of other valid sources of administrative data.

32. Evidence submitted by the London Borough of by Southwark, for example, suggested that ONS estimates had failed to take into account a temporary—and therefore largely uncounted—population of some 21,000 residents, an estimate made by a comparison of ONS estimates and the number of people known to be registered with GPs in the borough.[26] Colin Barrow, Leader of Westminster City Council, suggested that his borough probably had at least 10,000 residents who had not been not counted because they were short-term residents.[27] Sir Robin Wales, the elected Mayor of Newham, told us that his research commissioned by his borough suggested that around 20,000 people were uncounted in Newham.[28]

33. Professor David Martin, Chair of the Census Study Group, told us that one source of data for preparation of the mid-year estimates was the International Passenger Survey.[29] This records the movement of people into the country, but does not provide information about movements around the country or subsequent departures from the country, and cannot therefore be relied upon for an accurate portrayal of population.

34. The relationship between mid-year estimates and Census outputs highlights the need for both sets of figures to be reliable. This issue is in part exacerbated by what Keith Dugmore of the Demographic User Group termed the 'inter-censal drift',[30] that is, the accumulated error in population estimates over the period since the last census. Whilst the outdating of statistics cannot be easily remedied, the fact that mid-year estimates are rebased according to Census figures means it is crucial to have an accurate Census in order to validate subsequent figures.

35. Boroughs argue that this lack of precision about the size of their resident populations has had a substantial effect upon their ability to plan and provide resources for the public services they are required to deliver, which include education, waste, transport and housing. Local authorities such as Hackney, as well as the Greater London Authority and the Government Office of London, use population estimates to plan capital investment for public services in the short, medium and long term.[31]

36. While there is no evidence to suggest that minority communities disproportionately use public services, their use of these services introduces a further degree of complexity to the considerations of local authorities planning them. Most notably, their use of services can often mean that information about them has to be supplied in a range of different languages. In Southwark alone, over 100 languages are spoken to the local authority's customer service centre, and 125 languages are catered for in primary schools in Hammersmith and Fulham.[32]

37. It is also important to note the effect on borough populations of short-term migrants to London from within the UK, who are equally difficult to count in official statistics but who also consume London borough services.

Effects on funding and service provision

38. Many of the London boroughs who provided evidence to the Committee considered that their populations had been undercounted at the 2001 Census. Several London boroughs had their population estimates reviewed following claims of under-enumeration in 2001. While only three (Westminster, Wandsworth and Southwark) had changes made to their population estimates as a result, other boroughs such as Islington have received some additional funding in recognition of discrepancies between mid-year estimates and actual population.[33]

39. The London Borough of Hounslow has estimated that it had lost around £4.5 million between 2005 and 2008 as a result of the use of inaccurate ONS population data by the Department for Communities and Local Government.[34] This is one example of the numerous estimates of underfunding submitted to us in evidence by individual London boroughs: there is no official estimate of the total sums in grant which local authorities in the capital may have foregone through the application of undercounted population figures to grant formulae.

40. The London Borough of Newham considers that, as a result of the undercount it estimates in the 2001 Census, the Borough has subsequently not received, after damping, £5.8 million of Formula Grant which it believes it ought to have received by virtue of the size of its population. Sir Robin Wales told us that "because all [Newham's] resources are based on that figure and then the mid-year estimates, we have had a decade of underfunding based on that poor Census data."[35]

41. Hackney and Newham consider the situation so serious that they do not now rely on data from ONS for planning purposes:

    "Hackney Council no longer has sufficient confidence in the ONS population statistics to rely on them for most areas of our work, although we are required to use them for many purposes."[36]

Newham, Hackney and Ealing now choose to use locally-held data sources such as GP registers, school registers and requests for National Insurance numbers to estimate their populations.

42. At least three London boroughs have engaged independent consultants to carry out population counts based on administrative data sources, for the purposes of comparison against ONS mid-year estimates. Brent, Hackney and Newham have all independently engaged Mayhew Associates to conduct studies of their populations in order to provide an accurate population count. The studies—which may include some short-term migrants not included in mid-year population estimates—found that Hackney's population had been undercounted by 13,515 in the 2007 mid-year estimate and Newham's by 20,500 in the same period.

43. Dissatisfaction with the accuracy of the outputs from 2001 in London is persistent. The High Ethnicity Authorities Special Interest Group (HEASIG), on which many London boroughs are represented, believe that the Department for Communities and Local Government should acknowledge the imprecise nature of the population data used in allocating funding and that those boroughs who can provide evidence of this should receive additional funding as a result.[37]

44. The representatives of London boroughs who gave evidence to us on 8 February were all critical of the perceived deficiencies of present population estimates based on the 2001 Census and equally critical of their use by the Department for Communities and Local Government for calculating local authority grant allocations. Colin Barrow told us that approaches to Ministers on the issue had had no effect, even though it was widely acknowledged that more accurate information from administrative data was available.[38] Sir Robin Wales characterised the position thus:

    "The ONS knows it is nonsense, so it says, "Don't use these figures." DCLG knows it is nonsense, but says, 'We have to use these figures, because they're all we've got.'"[39]

45. We are concerned to note the lack of confidence some London boroughs have in the official mid-year population estimates derived from the 2001 Census, both as a means of planning for services and as a basis for funding distribution. It is of substantial concern to us that some boroughs have had to commission their own population estimates based on administrative data sources for the purposes of planning service provision.

46. The Government is not at present convinced of the value of using administrative data for such purposes. While some London boroughs have indicated that GP registrations would be a valid source of population data to be included in the process for calculating mid-year population estimates, the Cabinet Office sounded caution:

    "Just on GP registrations, for example, one of the problems with those databases is that people register with a GP when they move to an area or not. It might be a number of months, even years, before they decide to go and register, because often people might not go to see a doctor until they are ill. So there are associated problems with databases held by London boroughs as well. I think it is a question of understanding those."[40]

The Government Office for London claimed not to be aware of the significant concerns raised by boroughs on the accuracy of mid-year estimates based on the 2001 Census, though they were well aware of concerns raised by boroughs—and London Councils—in respect of the changes to the mid-year estimates which had been made as a result of changes to the methodology related to migration statistics.[41] Chris Hayes, Director of the Government Office, told us that "I am not aware of any specific issues that [the boroughs] have with recent mid-year population estimates that have updated [the 2001 Census], apart from concerns about the extent to which factors such as migration are taken into account
[...] "[42]

13   Statistics Commission, Report No. 22, Census and population estimates and The 2001 Census in Westminster: Final Report Back

14   Ev 93 [Royal Statistical Society] Back

15   Ev 68 [Islington Council] Back

16   Q 125  Back

17   Ev 57 [Office for National Statistics] Back

18   Q 132 Back

19   Q 133  Back

20   Q 113 Back

21   Q 60  Back

22   Ev 62 [London Borough of Lambeth]  Back

23   Ev 103 [ESRC Census Programme] Back

24   Q 147 Back

25   Nargis Rahman and Shayla Goldring, Modelling Census Household Non-response, Office for National Statistics, 2007, cited by the London Borough of Lambeth at Ev 61-62. Back

26   Q 73  Back

27   Q 61  Back

28   Q 56  Back

29   Q 22 Back

30   Q 20 Back

31   Ev 78 [London Borough of Hackney]  Back

32   Q 93; Ev 71 [London Borough of Hammersmith and Fulham] Back

33   Ev 56 [Office for National Statistics] Back

34   Ev 143 [London Borough of Hounslow] Back

35   Q 56 Back

36   Ev 79 [London Borough of Hackney] Back

37   Ev 91 [Local Government Association's High Ethnicity Authorities' Special Interest Group] Back

38   Q 66 Back

39   Q 65 Back

40   Q 202 Back

41   Ev 74 [London Councils] Back

42   Q 201  Back

previous page contents next page

House of Commons home page Parliament home page House of Lords home page search page enquiries index

© Parliamentary copyright 2010
Prepared 31 March 2010