Beyond 2011 - The census and future provision of population statistics in England and Wales

Public consultation September 2013Bristol City Council Response DRAFT

General comments

Q1: What are your views of the different census approaches described in this document?Bristol City Council:

a) support undertaking a 2021 Census collecting information from all households;b) do not support a move to the estimation of population and population characteristics

using administrative and annual surveys as an alternative to a census;c) would like to see a combination of the two methods with a 2021 Census acting as a

benchmark for providing data on population and characteristics together with administrative data, enhanced through linking individuals, to provide more accurate and up-to-date population estimates and characteristics data following the census.

ONS should be striving for is good data every year and great data every 10 years, rather than having to choose between ‘good data every year or great data every 10 years’.

Bristol City Council’s view is that it would be a fundamental mistake to abandon the Census in favour of using untested and high risk methods based solely on administrative data/annual surveys. Improvements to administrative datasets and the addition of key population characteristics would be welcomed in addition to the carrying out of a census. The risks involved in moving away from a census currently outweigh the advantages.

The Beyond 2011 Independent Review of Methodology came to the conclusion that population estimates based solely on a Statistical Population Dataset (compiled from administrative data sources) ‘would not be able to meet the quality targets that ONS have set’. The significant unresolved issues with population estimates based on administrative sources are alone enough of a reason to continue to support the undertaking of a census to use as a benchmark for any new and improved methodologies.

The census has a proven ability to deliver detailed quality statistics for small geographic areas. The census provides consistency across the UK, comparability between areas, a single snapshot in time, accuracy and statistics available for very small geographies. The census is the fundamental and only benchmark on which to test alternative administrative datasets and methods given the absence of a statutory population register as maintained in many countries who rely on alternatives to a census.

Frequency vs accuracy vs geographyThe results of the first ONS Beyond 2011 consultation in January 2012 which attempted to determine user preferences in relation to frequency of demographic statistics versus accuracy and

geography resulted in no clear steer. This serves to emphasise the wide ranging uses made of demographic data and the need to improve all three aspects rather than sacrifice one aspect in favour of another.

The decision to opt for the administrative data/annual survey option significantly favours frequency over accuracy and geography at the expense of a significant loss of data. Given the increasing need to reliably target fewer resources to smaller population groups and smaller areas and to embrace localism and neighbourhood planning, accurate data for small geographies is vital. The requirement to be able to measure difference between groups (eg equalities groups), between areas (eg LSOAs) and also change over time for these groups and areas is increasingly important.

CostsThe difference in costs between the two options is minimal. The estimated cost of £625m or £1.10 per person per year for a full census is very cheap given the extensive use made of the data. The fact that it is comparable to other countries should be evidence enough that it is value for money.

Although ONS cost estimates show that the administrative/annual survey option is slightly cheaper, we would question this calculation given that this option involves the loss of considerable amounts of data including potentially commuting statistics, workplace population characteristics and household characteristics. These datasets are fundamental to policy development at the local level as well as nationally and the cost for this will need to be met in future if a census is not undertaken.

Uses and benefits of population and housing statistics

Q2: Please specify any significant uses of population and housing statistics that we have not already identified.Please refer to Bristol City Council response to earlier consultation on the uses of census data which is included as an appendix.

Q3: Please specify any significant additional benefits of population and housing statistics that we have not already identified.No response.

Impact of different census approaches on statistical uses

Q4: What would the impact be if the most detailed statistics for very small geographic areas and small population groups were no longer available? High, medium, low, or no impact?The potential significant loss of data for small areas and the data that is supplied being subject to large confidence intervals would have a very high impact on our work.

The current proposal for an annual survey would not provide the levels of accuracy that are required in order to effectively and confidently plan for our populations. In Bristol, areas of extreme deprivation are adjacent to areas of extreme affluence. Sometimes these different communities live within the same ward and the ability to identify detailed accurate counts of characteristics for small areas such as LSOAs is critical to resource allocation and service planning. Lower Layer Super Output Areas are our main means of analysing data and the use of LSOAs has become increasingly embedded in work across all departments. Cross tabulations of variables by age – mainly children, young people and older people - are our main means of identifying our target populations. The proposed annual survey would not be able to meet this need.

In the consultation as it stands, no consideration appears to have been given to some of the more detailed alternative population outputs from the census such as commuting statistics, workplace population characteristics and household characteristics. The census is currently the only source for these statistics and again the levels of accuracy that are required would not be available from the annual survey. Also, the loss of the wealth of multivariate data and the ability to produce derived variables (eg NS-SeC) is a significant concern if we were to move away from a census.

EXAMPLES OF THE POTENTIAL IMPACT OF THE LOSS OF SMALL AREA DATA IN BRISTOL

Example 1: Community Education and LearningCommunity Learning West receives £1.8 million Skills Funding Agency funding to deliver community learning across Bristol, North Somerset and South Gloucestershire local authorities. Funding is used to provide opportunities for developing employability and functional skills (English, language and maths), skills for community development and involvement, personal development, and also for family learning activities to support children to learn. Funding and activity is particularly targeted at learners who have few or no qualifications.

The three local authorities that make up Community Learning West have very different geographical and population profiles. Because of these differences LSOA level census data has been essential in targeting and planning community learning provision. Adult qualification level by ward and LSOA was the starting point for identifying areas of particular need. LSOA data has allowed us to ensure targeting of provision meets the needs of residents in these ‘hidden’ areas as well as providing for residents in more visible areas of need.

In addition to adult qualification levels our service has used the LSOA level census data to identify the number of BME residents, the location of faith groups, numbers of disabled people or those with long term limited illness and employment status information. We have used this information to set local area targets for engaging adults in community learning and encouraging progression to further learning, volunteering and work. By having this information it is possible to directly target communities requiring additional support and ensure national funding is able to have the most significant impact in relation to economic regeneration, education and community cohesion.

In 2012, Ofsted awarded Bristol City Council and Community Learning West with a strong ‘good’ (grade 2) judgement. Inspectors commented on the effective use of data and management information to target services effectively. In conclusion, census data in particular LSOA profile data allows us to utilise government funding to have the most significant impact on adult learners from disadvantaged communities across our region.

For a fuller description of this work see Appendix.

Using the ONS Confidence Interval Calculator to calculate the availability of data on qualifications based on the proposed annual survey option, in Bristol, even if the survey data were aggregated over a five year period:

Total number of people aged 16 and over with no qualifications at LSOA level: Mina Road LSOA had 217 people with no qualifications in 2011 – we only know this accurate

count every 10 years. Survey based on a 5 year average would estimate that there was between 130 and 304 people with no qualifications in the LSOA – this number would never be accurately counted again as there would be no census to do this.

Example 2: Bristol Local Plan Housing PolicySmall area data (LSOA) is required to support development management officers in their implementation of housing policies within the Local Plan. In particular, data on the composition of housing at a small area level is needed to assess the suitability of housing mix within development proposals and the contribution that different types of new residential development make to the housing mix of an area. This data has also been relied upon at planning appeals. Housing policies within the Local Plan could not be effectively implemented without this data.

Using the ONS Confidence Interval Calculator to calculate the availability of statistics based on the proposed annual survey option even if the survey data were aggregated over a five year period:

Total number of flats - LSOA level survey data aggregated over 5 years: no statistics provided for 60 out of our 263 LSOAs (23% of our LSOAs) 165 LSOAs (63% of our LSOAs) would have CIs greater than +/- 10% just 38 LSOAs (14% of our LSOAs) would have CIs less than +/- 10%

The number of flats in an area – currently broken down further in the census into purpose-built and conversions - is very basic key variable which, when accurate (ie not subject to +/-40% CIs), can be used as a baseline on which to monitor changes in housing type Using our monitoring of housing development, we are able to produce an annual time series for small area geographies based on actual counts rather than estimates from a sample survey.

Example 3: Housing StrategyBristol City Council use tenure information to look at the balance of tenures at the LSOA level to identify areas of concentration of social housing which also exhibit long-term worklessness, poor educational achievement, low income etc.

This data has been used to inform decisions on:

the allocation of LA affordable housing programme resources, the prioritisation of schemes for the development of new affordable housing by Registered

Providers for the allocation of government Affordable Housing Programme resources, the introduction of local allocation schemes, the balance of tenure in new affordable housing development , the design of housing need and aspiration surveys for specific areas of Bristol to inform the encouragement of balanced and sustainable communities in Bristol.

Information solely at the ward level does not provide the degree of detail required for these purposes. Without accurate LSOA level statistics, we would be unable to have a clear understanding of the characteristics and needs of communities at the very local level which enables the authority to target scarce resources most efficiently and effectively.

Comparison with 2001 Census small area data enables us to track, for instance, the rapid growth and concentration of the Private Rented Sector, and the implications this will have for future housing investment and development and meeting the acute shortage of affordable homes in Bristol. It is not

possible to track this development through the planning system as homes move between the owner-occupied and rented market.

As the focus of development moves toward new private sector delivered market rented homes it will be important to understand the factors driving decisions around private investment and the effect of this development on the character of smaller geographies and communities in Bristol. Accurate local detail, even at 10 year frequency, is more useful than less accurate but more frequent data. Most of the big social economic and environmental changes happen slowly, and only the detail can inform the changing picture over time.

Using the ONS Confidence Interval Calculator to calculate the availability of tenure data based on the proposed annual survey option, in Bristol, even if the survey data were aggregated over a five year period:

Total number of households privately rented - LSOA level survey data aggregated over 5 years: no statistics provided for 95 out of our 263 LSOAs (ie 36%) an LSOA which includes 631 privately rented households making up 54% of the stock would

have an 11% CI, so the data provided based on a survey would estimate between 564 and 698 households, ie between 48% and 60% of the total stock.

Total number of households social rented - LSOA level survey data aggregated over 5 years: no statistics provided for 118 out of our 263 LSOAs (45% of our LSOAs) even LSOAs with the highest number of socially rented households would get estimates

which could be as much as a 100 households different from the actual count.

All these examples serve to illustrate how the potential significant loss of data for small areas, and the data that is supplied being subject to large confidence intervals, would have a very high impact on our work. If there is no 2021 Census, then this effectively means that for a considerable number of LSOAs in Bristol there would no longer be any basic univariate statistics as well as no multivariate statistics. This is not acceptable.

Q5: What would the additional benefit be if more frequent (ie annual) statistics about population characteristics were available for areas like Local Authorities and Electoral Wards? High, medium, low, or no impact?The opportunity to identify changes in society more frequently is welcomed and would have a high impact on our work. That said, the current proposal for an annual survey would not provide the levels of accuracy that are required in order to effectively and confidently plan for our populations.ONS already provide annual population estimates at LSOA level and above broken down by single year of age and sex. These estimates currently rely on demographic components of change at LA level and administrative data for changes at smaller geographies. Increasing use of administrative data is welcomed in order to improve these small area estimates. The inclusion of additional key variables in administrative datasets in the longer term (eg ethnic group, country of birth, household income) would be a welcome step in order that some of the key characteristics of the population can also be estimated more frequently for small area geographies such as wards and LSOAs. In Bristol we already make extensive use of administrative sources which are available more frequently in order to supplement census results. Demographic changes experienced in Bristol between the 2001 and the 2011 Censuses were significant. In the absence of up to date census data,

we were able to access a variety of administrative datasets in order to get a better understanding of the changes that were taking place. These sources included:

National Insurance Registrations of Adult Overseas Nationals entering the UK Migrant Worker Registration Scheme Work Permits GP registrations School Census Students statistics

Whatever the results of the Beyond 2011 consultation, we will continue to use all sources available to paint a picture of our population.

Impact of different census approaches on historical research

Q6: Please specify any significant uses of census information for historical research that we have not already identified.No response.

Q7: What advantages or disadvantages for genealogical or historical research can you see from a move to a solution based on archiving administrative data sources?No advantages.

Managing risks

Q8: What are your views of the risks of each census method and how they might be managed?The risks outlined in the Beyond 2011 Consultation in relation to opting for a 2021 Census – namely reduced response rates, internet access and privacy – all also apply to the administrative data/annual survey option. All the additional arguably much more significant risks outlined – namely untested methods, loss of data and the stability of administrative datasets in a world of rapidly changing policy - are solely in relation to the administrative data/annual survey option.

Untested methodsThe administrative data/annual survey approach would depend upon new and untested methods requiring extensive development of new statistical methods and computer systems. Such a significant decision about the long-term future provision of demographic statistics should not be made at the current time given the level of uncertainties and risks involved.

The Beyond 2011 Independent Review of Methodology states that population estimates based solely on a Statistical Population Dataset (compiled from administrative data sources) ‘would not be able to meet the quality targets that ONS have set’. Further the ONS Beyond 2011 consultation itself states: ‘Our research has shown that producing statistics using administrative data is most difficult

for large cities, because the population changes most rapidly in these areas and administrative data tends to be out of date.’ Issues of under-coverage, over-coverage, duplications and omissions will therefore be concentrated in areas with the most transient populations. Until these issues have been fully addressed, a census is still vital as a quality benchmark.

Countries relying heavily on administrative data for statistical purposes are generally those in which there is a population register where records are tied to a personal number and a single address which people are legally required to keep up to date. The convenience of being able to change just one central record rather than having to contact numerous government and non-government organisations (as in the UK currently) ensures records are kept up to date and hence data is likely to be considerably more accurate and consistent between sources.

Discontinuity in statisticsDiscontinuity in statistics is inconvenient but if it leads to an improvement in methods and the quality of the outputs then this is generally acceptable. Population estimates using the demographic model have been improved considerably in recent years mainly due to the increased incorporation of administrative sources. The revision of the back time series to 2001 was possible in order to preserve continuity over the decade between censuses.

The proposed option of using administrative data in order to estimate the population by age and sex rather than using the demographic model would mean a fundamental shift in the way the population is estimated. This would not however replace the need for demographic modelling since such trends would still be required to produce population projections.

Reliance on administrative dataReliance on administrative data sources would require access to statistics that would always be susceptible to policy changes. The risk of loss of data and the need for a changing methodology to accommodate changes could lead to discontinuities in time series and loss of confidence in the estimates.

One example of the unreliability of administrative sources to improve population statistics that was anticipated by ONS, was the use of e-borders data to improve estimates of international migration. The e-borders data was planned to feed into the ONS Migration Statistics Improvement Programme (MSIP) which ended in March 2012. E-Borders originally had a target of achieving 100 per cent passenger coverage by 2014, however, following contractual issues, a new date for 100 per cent coverage was set at 2018 but again this has been delayed and now the expected timeline is stated as ‘in years to come’ due to issues around delivery.

The increased use of administrative data sources is welcome but total reliance on such sources at this stage without a population register or a census as a benchmark would be unacceptable.

Response ratesThe risks outlined in relation to the difficulty of achieving good response rates can be applied equally to both the census and the annual survey options proposed. Both options require that the census/survey is compulsory and both options rely on internet response. ONS will need to invest in ways to achieve high response rates.The good response rates achieved in the 2011 Census were in part due to the increased emphasis on local authority and community engagement in order to promote the benefits of the census and why people should respond. This was largely achieved due to the substantial input from Local Authorities and the economies of scale involved. Should there be a decision to move to an annual sample survey then the economies of scale in terms of publicity and promotion will no longer be achievable. More significantly, Local Authorities are unlikely to commit resources to community engagement in

relation to an annual survey given that the outputs are unlikely to fulfil their requirements for accurate small area data.

A further cause of low response rates is the fact that members of the public see responding to surveys as a burden. For both the census option and the annual survey option, there would be a legal requirement on people to respond. Although the census is often claimed to be a burden on people, the fact that it affects everyone at the same time just once a decade effectively helps to make it more acceptable. Moving to a compulsory annual sample survey, where ONS estimate that approximately 50% of the population will be sampled in a ten year period, may have the effect of making people feel singled out (which in fact they will be!) and less inclined to respond if not everyone on their street is being asked the same questions either on-line or on the doorstep.

In 2011 we were told by many people that filling in the census questionnaire was a family activity and educational for their children. In particular, the fact that the questionnaires might be used for genealogical research by future generations of their family was an added incentive. This incentive will not exist with an annual sample survey.

Non-response could therefore be more of an issue for an annual survey than for a census. However, as stated in the Beyond 2011 Independent Review of Methodology, without census data, there is little evidence in the programme of possible auxillary variables which could be used to correct for non-response bias’.

Internet accessAs acknowledged by ONS, a step change will be required in terms of internet response rates in order to make an on-line 2021 Census or an annual survey viable. Also, it will be important to ensure that people without internet access are not excluded in any way. Relative to other risks outlined however, this seems less significant given the rapid changes in technology that are taking place and the numerous ways in which people would be able to engage eg via home or publically available computer, tablet, smart phone etc.

Bristol City Council like many local authorities is investing in a ‘one council’ digital platform for Bristol residents, businesses and visitors to access council services and information. This will enable the council to realise substantial savings through channel shift by offering digital services ‘“so good, that people prefer to use them”, move customer interactions to self-service digital channels and away from in-person telephone and face-to-face interactions, reduce or eliminate unnecessary back-office manual process and duplication through greater automation and deeper systems integration and model best-practice in service design and agile delivery of compelling and robust digital services which can be rolled out across the council.

Currently in Bristol 85% of households (and rising) have access to the internet and of these 73% (and rising) are currently using it regularly. The internet by 2021 should be the main means of accessing and using services across local authorities and most organisations.

Privacy and public acceptabilityPublic acceptance of the use of administrative data for statistical purposes is a huge hurdle to be overcome and no commitment should be made to such a method until it can be ensured that this acceptability can be accommodated. The issue of privacy in relation to a census, can be addressed using targeted promotion and publicity at a single point in time once a decade. Administrative data matching however would be an on-going issue. Indeed, people may feel so strongly that it actually affects their actions including such things as choosing not to register or giving misinformation in order to avoid being counted.

Q9: Are there any other issues that you believe we should be taking into accountComplexity of Statistical OutputsAn annual survey would produce outputs that are much more complicated than the census data and administrative datasets that are currently used. The complexity of outputs - including data aggregated over a varying number of years depending on the geographical level, no consistency of outputs across the country and confidence intervals for individual variables - will make the interpretation of the statistics much more difficult.

This is a fundamental issue which needs to be given much more consideration before any final decisions are made. For the many users of socio-economic statistics the data would probably be almost inaccessible. Much time will need to be spent by ONS and analysts explaining what the data means, how to interpret it and presenting it, rather than getting on with actually using it. The significant costs of this additional work are unlikely to have been quantified and included in any cost-benefit calculation when choosing between the two options.

Census outputsOutputs from Census could be simplified in order to reduce costs and improve timeliness of published data. At the moment there are too many very detailed cross-tabulations and not enough simple cross-tabulations. A more flexible approach to outputs is required so people can choose what they wish to include in tables rather than being restricted to set standard tables.

Jayne MillsStrategic PlanningNeighbourhoods and City DevelopmentBristol City Council

Appendix: ONS 2011 Census Benefits Realisation Survey - Bristol City Council Case Studies - March 2013IntroductionIn order to support the ONS Benefits Realisation Project, rather than complete the questionnaire provided, we have pulled together a selection of case studies which demonstrate our need for small area data now and in the future. We have also attempted to provide a very approximate estimate of how much it would cost the Local Authority to undertake a census in Bristol ourselves, as well as including our thoughts about geography.

Census strengthsOnly a census can provide such a wide range of comprehensive and consistent statistics collected at the same time and across different geographies at the same time, allowing comparison between data without issues of data quality, which are inevitable when using individual administrative datasets.

The Census provides a baseline for trends and is currently the only means of rebasing and estimating population, household and socio-economic statistics that are collected through surveys during the inter-censal period. Further, a census will be a key requirement in order to evaluate any new methodology adopted.

Accuracy is extremely important as it gives us confidence in the data and the ability to identify trends and issues arising. Our preference is to use more accurate data even though it may be dated, although ideally we would obviously like both accurate and timely data. Since the census is compulsory this results in very high response rates which are not usually achieved via voluntary or sample surveys. As a result, accuracy of alternative measures is likely to be an issue in future.

Multivariate statistics and derived variables are a key feature of census data. The loss of the wealth of multivariate data and the ability to produce derived variables (eg NS-SeC) is a significant concern if we were to move away from a census.

Financial consequencesLocal authorities require the best information possible to work effectively and to avoid duplication of effort and waste of resources. If every local authority in the country had to try to acquire or develop small area statistics, the resource and waste implications are enormous. This is of increasing importance in a climate of public sector contraction.

The consequences of not being able to work effectively are social as well as financial.

Estimating the cost of a ‘Bristol Census’The Quality of Life in Your Neighbourhood annual survey provides an invaluable insight into life in Bristol and is now in its twelfth year. The results are used by the council, health service and other public sector partners to help plan local services, track change and improve the quality of life in Bristol. The survey is the council's main tool for providing neighbourhood level statistics and public perception information - this long running survey also allows people to see how things are changing over time.

Based on the Quality of Life model (A5 questionnaire, envelopes and 2nd class post) and with an 85% response rate, cost is £1.85 per return, so £791,000 total for whole population. This includes postage, printing, packing, 1 reminder, freepost returns (or online), data entry. A4 packs would increase the cost of printing, as would 1st class postage and take this total cost to approx. £1,000,000.

Our staff costs would include design (paper and online), logistics, provision of adjusted and aggregated data - at least 2FTEs, for a year plus another 2FTEs whilst the survey is collecting. This would add another £100,000. The promotional work/materials would also need to be included eg. leaflets, banners and general advertising (we don't do this for the Quality of Life survey). Also we rely on one reminder and the cost of enumerators/collectors and the whole cost of neighbourhood/census workers is a huge add on. Sarah McMahon, Consultation, Research and Intelligence Manager, Communications and Marketing, Bristol City Council

GeographyWhat minimum level of geography is required for small area statistics is very much dependent on local circumstances and how the data uis being used (ie public sector vs private sector) and so it is difficult to get concensus from all users.

In areas where LSOAs do not fit to wards or where data is not available for their current wards, Output Areas are likely to be the optimum geography. In areas where LSOAs fit to wards, LSOAs are more likely to be the optimum geography. In areas with large wards, OAs and LSOAs are likely to be the preferred geography.

Small area data is used extensively by Bristol City Council and our partners. In Bristol the main geography used to identify areas for resource allocation, neighbourhood renewal and funding bids are Lower Layer Super Output Areas. Wards are large in Bristol with an average population of 12,235. Areas of deprivation or need would easily be missed if statistics were only available at ward level and above leading to the ineffective and inefficient use of resources. Another advantage of LSOAs is that they are of equal size across the country and so can be used for benchmarking and making comparisons with other areas. This is not possible using ward level data as wards vary in size enormously from on area to another.

Output Area data although very useful for creating statistics for user defined areas are less useful and MSOAs are not used at all in Bristol as they have little local meaning spatially.

Bristol City Council Case Studies

The following case studies are based on information provided from departments across the city council. The case studies have been presented by the following topics:

A. Spatial PlanningB. HousingC. Education and LearningD. HealthE. Children and Young PeopleF. Economic Development and Inclusion

A. SPATIAL PLANNING

CASE STUDY: Local PlansThe National Planning Policy Framework stipulates that local planning authorities should ensure that Local Plans are based on ‘adequate, up-to-date and relevant evidence about the economic, social and environmental characteristics and prospects of the area’. Specifically it states that their Strategic Housing Market Assessments should take account of migration and demographic change, and should take account of the needs of different groups in the community. General requirements for information on the overall number and characteristics of the population are reinforced by the need for details on recent migrants and other specific sub-groups of the population in order to support the planning and delivery of core services and comply with specific obligations under the Equality Act 2010.

Census data are used to support a range of planning functions within local authorities, including:-

• undertaking predictive analysis;• structural planning and allocating resources; • developing, supporting and monitoring/evaluating outcomes of strategy;• risk profiling;• cost benefit analysis; and• conducting Joint Strategic Needs Assessments.

Key census variables include housing information, ethnicity, employment and travel to work. In addition to direct use of census information, extensive use is made of related outputs such as population and household estimates and projections to assess the future demand for services.

CASE STUDY: Neighbourhood Development PlansThe Governments Localism agenda has introduced an opportunity for local neighbourhoods to produce Neighbourhood Development Plans so devolving responsibilities to local areas. This is referenced in both the Localism Act (2011) and the revised National Planning Framework. Neighbourhood Planning is about making sure a community gets the development it needs for the future - through the building of homes, job opportunities, and leisure and community facilities, including schools, health services and shops.

Local Authorities will be expected to provide a substantial ready-made evidence base for plan making at the neighbourhood level. Neighbourhoods will need to gather information about the community in the neighbourhood area and these profiles will help in understanding current and potential future needs. In addition, information about population, health, employment and jobs, education and skills and groups of people with specific needs will help describe the ‘baseline environment’ as required by any Sustainability Appraisal.

The map provided shows those areas of Bristol that have either been designated as a Neighbourhood Planning Area or which have been proposed. The map illustrates the mismatch between the community defined areas and ward boundaries. Wards in Bristol are relatively large with an average population of 12,235. Some of the NPAs are very small compared to the ward that they are in and some wards even include more than one NPA. In order to be able to produce profiles of the people and households in the NPAs, LSOA or even OA level data is essential.

CASE STUDY: Local Plan Housing PoliciesSmall area data is required to support development management officers in their implementation of housing policies within the Local Plan. In particular, data on the composition of housing at a small area level is needed to assess the suitability of housing mix within development proposals and the contribution that different types of new residential development make to the housing mix of an area. This data has also been relied upon at planning appeals. Housing policies within the Local Plan could not be effectively implemented without this data.

Simon Fletcher, Strategic Planning, Neighbourhoods and City Development, Bristol City Council

CASE STUDY: Local Plan Health Impact AssessmentsSmall area data is valuable when we are doing health impact assessments, looking at the impacts of developments on small areas and considering health inequalities. Small area data provides a baseline and gives a picture of the type of population which is necessary when looking at health inequalities. Policy DM14 requires health impact assessments for all super major developments so there will be an increasing demand for small area data. It is also used in other neighbourhood based health work such as the Healthy Neighbourhood Check Walkabouts we have been doing across Bristol (Avonmouth, Lockleaze, Bedminster, St Judes/Old Market, Stockwood).

Stephen Hewitt, Healthy Urban Team, Bristol City Council/NHS Bristol

B. HOUSING

CASE STUDY: Housing Strategy and PolicyQ1 How has the analysis of small area information derived from 2001 Census data helped in the local authority’s decisions over the last decade in respect of housing?

We have used tenure information to look at the balance of tenures at the LSOA level to identify areas of concentration of social housing which also exhibit long-term worklessness, poor educational achievement, low income int al.

This data has been used to inform decisions on:

the allocation of LA affordable housing programme resources, the prioritisation of schemes for the development of new affordable housing by Registered

Providers (housing associations) for the allocation of government Affordable Housing Programme resources,

the introduction of local allocation schemes, the balance of tenure in new affordable housing development , the design of housing need and aspiration surveys for specific areas of Bristol to inform the encouragement of balanced and sustainable communities in Bristol.

Information solely at the ward level does not provide the degree of detail required for these purposes. We would strongly support the continued production of information at lower geographies, in particular the LSOA level.Q8 What plans have you got in place for analysis of 2011 Census data to inform changes in spending allocations in respect of housing?

We would anticipate using 2011 census data at the LSOA level to inform ongoing work on the allocation and direction of affordable housing resources, survey design, local allocation schemes as outlined above to encourage the development of balanced and sustainable communities in Bristol.Crucially by comparison with 2001 Census small area data we will be able to track, for instance, the rapid growth and concentration of the Private Rented Sector, and the implications this will have for future housing investment and development and the meeting of housing needs in Bristol. See comments at Q11 below.

Q9 If the 2011 Census was only able to deliver outputs down to Ward or MSOA level, what would you do differently, or be unable to do?

We would be unable to have a clear understanding of the characteristics and needs of communities at the very local level which enables the authority to target scarce resources most effectively.

Q11 If the 2001 Census had only delivered outputs down to Ward or MSOA level, what would you have done differently, or have been unable to do?

One example would be that tenure information at lower geographies enables the identification of those areas of the city where there has been a prolific growth and concentration in the private rented sector between census intervals. It is not possible to track this development through the planning system as homes move between the owner-occupied and rented market.

The private rented sector has become increasingly important for different income groups within the city, and rental levels vary considerably in different areas of the city, and are rising at different rates. With the acute shortage of affordable housing and increasing reliance on the private rented sector to house lower income households, information on the distribution and rent inflation trends within the PRS will be crucial in understanding the capacity of the sector to contribute to the servicing of housing needs of different income groups in the city.

New government-funded affordable housing development, and substantial proportions of existing social housing will in future have rent levels tied to market rents at the very local level. An understanding of housing markets at lower geographies, as well as characteristics of local communities will be essential in the effective targeting of scarce resources.

As the focus of development moves toward new private sector delivered market rented homes it will be important to understand the factors driving decisions around private investment and the effect of this development on the character of smaller geographies and communities in Bristol.

Q14 With options that would provide more frequent population numbers, say annually, but less detailed population attribute data at a geographically disaggregated level, how does the benefit of greater frequency compare with loss of local detail?

Local detail, even at 10 year frequency is more useful…Most of the big social economic and environmental changes happen slowly, and only the detail can inform the changing picture over time.

Kevin Mulvenna and Marie Price, Strategic Housing, Strategy and Commissioning, Neighbourhoods and City Development, Bristol City Council

CASE STUDY: Housing - Distribution of move-on accommodation from supported housing servicesAlongside essential homelessness prevention and supported accommodation services, Bristol has developed programmes to access private sector homes for those no longer needing support. Data at lower level geographies is helpful in:

identifying those areas of the city with a supply of 1 bed private rented homes identifying those areas of the city where more private access schemes are needed, and

where private renting supply of the right size supply would make this possible

We are able to cross-match this information with data we hold on the distribution and concentration of existing supported accommodation to assist in targeting and developing private access schemes outside of these areas.

Jonathan Didlick, Commissioning and Projects Officer, Strategy and Commissioning, Neighbourhoods and City Development, Bristol City Council

CASE STUDY: Housing – Targeted Local ActionWe regularly use the information at LSOA level to gain evidence for local targeted action and identify density of tenure of private sector housing and associated data such as deprivation, older people or vulnerable households.

For example resources can be appropriately targeted at areas who would most benefit from our intervention either through renewal activity and Home Action Zones for privately owned property or designation of discretionary licensing areas for private rented sector housing or enforcement action. We also use this information to select areas for energy efficiency initiatives. ECO areas (grants for affordable warmth) are based on LSOA so it makes sense to follow suit.

Targeted, local action rarely covers a whole ward which is too large geographically and therefore the data has to be broken down to smaller, manageable areas for action.

Most of our other data is either sample data which is extrapolated to ward level if were lucky or most commonly Local Authority level or from on our records which are not representative of the whole authority as often data is only available as a response to a complaint for example.

LSOA level Census data is a very valuable data resource as the data has been collected from every household, which is the only true statistically valid dataset at this level.

Jan Hamilton, Strategic Housing – Private Housing, Neighbourhoods and City Development, Bristol City Council

C. EDUCATION AND LEARNING

CASE STUDY: Community Learning WestCommunity Learning West receives £1.8 million Skills Funding Agency funding to deliver community learning across Bristol, North Somerset and South Gloucestershire local authorities. Funding is used to provide opportunities for developing employability and functional skills (English, language and maths), skills for community development and involvement, personal development, and also for family learning activities to support children to learn. Funding and activity is particularly targeted at learners who have few or no qualifications.

The 3 local authorities that make up Community Learning West have very different geographical and population profiles. Bristol is a large city with diverse communities. South Gloucestershire includes residential areas on the outskirts of Bristol as well as a large rural area and two market towns. North Somerset covers rural areas as well as the seaside town of Weston-super-Mare. Because of these differences census data, ward data and LSOA data has been essential in targeting and planning community learning provision. Adult qualification level by ward and LSOA was the starting point for identifying areas of particular need. Because of the size of the wards across the area LSOAs have

allowed us to identify pockets of deprivation and areas where up to 57% of adults have less than a level 2 qualification (5 GCSEs A-C).

If we look at this data at ward level we find that pockets of deprivation are hard to identify. For example within Bristol there are several wards that have high levels of deprivation across the board. These wards have historically received priority support and funding. We do however have lots of instances where the ward as a whole does not register as showing a high level of need but individual LSOAs within the ward show significant levels of deprivation.

LSOA data has allowed us to ensure targeting of provision meets the needs of residents in these ‘hidden’ areas as well as providing for residents in more visible areas of need. Within the Ashley ward for example we have one LSOA where 8% of the population do not have a full level 2 qualification alongside 28% in two other LSOAs.

Some other examples of variations between SOAs within wards:8% to 28% in Ashley, 28% to 44% in Bedminster, 5% to 26% in Cabot, 24% to 46% in Frome Vale, 12% to 40% in Horfield, 24% to 52% in Knowle, 21% to 46% in Windmill Hill (% of adults without a full level 2 qualification).

In addition to adult qualification levels our service has used the LSOA level census data to identify the number of BME residents, the location of faith groups, numbers of disabled people or those with long term limited illness and employment status information. We have used this information to set Community Learning West, Local Authority, Neighbourhood Partnership and local area targets for engaging adults in community learning and encouraging progression to further learning, volunteering and work. By having this information it is possible to directly target communities requiring additional support and ensure national funding is able to have the most significant impact in relation to economic regeneration, education and community cohesion.

2011/12 results show that 29% of learners were disabled, 22% from BME backgrounds and 12% were under 25 years old. 71% of our learners do not have a qualification equivalent to 5 GCSEs A-C. Across Community Learning West 57% of learners live in the 25% most deprived SOAs nationally. In Bristol 82% of learners live in the 25% most deprived LSOAs nationally.

In 2012, Ofsted awarded Bristol City Council and Community Learning West with a strong ‘good’ (grade 2) judgement. Inspectors commented on the effective use of data and management information to target services effectively:

‘Management is good. Strategic planning is very effective in developing andnurturing the consortium. Curriculum management and the use of managementinformation are good; arrangements to improve quality are well planned.’

‘Actions to widen participation are particularly effective. WEC targets its servicescarefully and a high proportion of learners come from the most vulnerablegroups. Staff have a good knowledge of, and links with, local communities andthose individuals who are hardest to reach. It offers carefully tailored courseswhich meet their needs well.’In conclusion, census data in particular LSOA profile data allows us to utilise government funding to have the most significant impact on adult learners from disadvantaged communities across our region.

Jenny Wilkes, Data and Communications Officer, Community Learning West, Bristol City Council

CASE STUDY: Communities and Local Government ‘Community-Based English Language Competition’The Department for Communities and Local Government is launching a competition to discover, and support, organisations than can deliver community-based English language programmes. The competition is being used to discover innovative ways of delivering community-based English language programmes that will help to integrate their participants into their local communities. They are looking for programmes that deliver genuine value for money, and that can be sustained over the longer term. The size of the total funding pot will be determined by the number of bids that meet the Department’s requirements, up to a maximum of £6m.

Bristol City Council will be making a bid for this funding based on 2011 Census data on language proficiency and new migrants.

For further information:https://www.gov.uk/government/publications/community-based-english-language-competition-stage-1-prospectus

Suad Abdullahi, English Language & Citizenship Project Manager, Bristol City Council

D. HEALTH

CASE STUDY: Bid to National Institute for Health ResearchEvaluating the effectiveness of the introduction of signs-only 20 mph speed limits in residential areas: impacts on road traffic casualties, wellbeing and social capital. A mixed methods study.

Attached is the 16 page Project description for a Main bid to National Institute for Health Research (NIHR) in which, for the quantitative work to look at road casualty statistics, we intend to use LSOA data. This is contained on pages 4-6.

Dr Adrian Davis, Public Health support to City Development, Bristol City Council

CASE STUDY: Joint Strategic Needs AssessmentThe JSNA is an assessment of the health and wellbeing needs of the local population, and in order to do this effectively we need to have a clear understanding of the underlying population demographics within Bristol. Whilst primarily we use ward-level data, there are many occasions where lower level data at LSOA is required to highlight the differences within wards.

There are often stark health inequalities within wards, so having wards as the lowest level available would not be sufficient detail. For example the highest and lowest Life Expectancy estimates in Bristol are in wards which lie next to each other, and in order to understand why we need to analyse the demographics and deprivation patterns within each ward. This ability to drill down further into the underlying issues at LSOA-level highlights the issues and allows us to better target service delivery.

Deprivation data at LSOA-level was used to inform the roll-out of NHS Health Checks in 2012, to target promotion to households in the most deprived areas where health outcomes were poorest.This same principle applies to analysing patterns of wider determinants of health, in areas such as child poverty and worklessness, where LSOA-level data within the JSNA is used by commissioners and also by partners and service providers to inform the planning of targeted service delivery.

In addition, health partners do not specifically focus on electoral wards, and LSOAs provide the building blocks for whatever geographies may be required. A considerable limitation to MSOA as

the potentially lowest level is that these don't map to wards, so we need LSOA in order to provide meaningful breakdowns across areas.

It is useful to have access to OA-level data for specific project analysis, though I accept that this level data is not routinely used within the JSNA process.

Nick Smith, Joint Strategic Needs Assessment Project Manager

E. CHILDREN AND YOUNG PEOPLE

CASE STUDY: Children and Young People’s ServicesUses of small area data:

Grant bid support – for example a recent Big Lottery grant looking at defined local area grants, based on IMD and other census information.

Modelling and forecasting of service demand – for example looking at small area population data and preferred service use to predict demand on school place allocations. In 2011, this work was worth an extra £19m in funding from Government: without our modelling data, this would not have been possible.

Equalities impact assessments for geographically-based projects.

Andrew Turvey, CYPS Information & Analysis Manager, Children and Young People’s Services, Bristol City Council

CASE STUDY: Early Years, Children and Young People’s ServicesWe in Early Years would be pretty much lost with out LSOA data. We currently use:

LSOA population estimates with IDACI to allocate about 85% of all Children Centre funding in a targeted way (aprox £6million)

IDACI by LSOA to allocate deprivation supplements to providers of the free early education entitlement for 3 & 4 year olds (£800k)

LSOA population estimates to ensure sufficient childcare is available, assess take up of Children Centre services, aid in developing/managing the childcare market and plan for the implementation of the free entitlement for 2 year olds (all statutory duties)

Christopher Amphlett, Project Manager, Learning Achievement, Early Years, Children and Young People’s Services, Bristol City Council

F. ECONOMIC DEVELOPMENT AND INCLUSION

CASE STUDY: Targeting small areasWithout small area data accurate targeting of people with, for example, low skills who are persistently workless will be impossible. As a result, dwindling resources will be spread more thinly.

As an example, in collaboration with other organisations, BCC runs regular job fairs. Advertising for said fairs is targeted at people living in areas of persistently high unemployment. Without small area (LSOA) data targeting would be at a ward level and as a result would either miss some of the targets or cost about twice as much, approximately £7,000 instead of £3,500.

In extreme cases, where 1 in 6, 7 or 8 small areas, that make up a ward, qualify for targeting, costs would rise over 6 fold and possibly as much as 8 fold.

In Bristol two examples are evident. Both the Henbury and Frome Vale wards have small pockets of significant worklessness (1 LSOA in each ward). In Henbury this amounts to about 830 in 5,200 and in Frome Vale 670 in 5,300 households. If statistics were collected at the ward level theses small areas would be missed.

Efficient targeting of other funding streams aimed at, for example, ethnic minorities (ERDF) or people with low skills would be, if not impossible, much more difficult.

Also producing accurate statistics for Enterprise Zones with boundaries that do not follow those of wards will be impossible.

Steve Usmar and Chris Hackett, Economy, Enterprise and Inclusion Team, Bristol Futures, Bristol City Council