Seminar Series End of the Census The UK population is undergoing drastic movement, with seachanges in mortality rates, life expectancy and how long individuals can hope to live in good health. In order to cater to this shift the Census is under review to see if there is a better way to manage and build the dataset Britain needs to support the changing nation. Examining the impact of a changing Census is part of a broader demographic series supported by Partnership Assurance Group plc (Partnership) entitled Population Patterns, which will look at the influences on and changes in UK demographics, particularly our ageing population. In provoking high quality debate about the way in which Britain is changing, how we record the changes and respond to it, ILC-UK and Partnership aim to influence better decisions in the way we plan for older age and support consumers throughout the changing landscape of their lives. The Population Patterns series, #populationpatterns, will look at the long-term challenges demographic change will have for government, especially age-related expenditure. Why does the Census need reforming? The Census has tracked population, housing, labour markets and a host of other factors that help to indicate change in UK life for the past 200 years. Since 1801 the Census has been completed every 10 years; with the last one being completed in 2011 at a cost of 480 million. Paper forms were sent out to the majority of homes in the UK, with some allowed to fill out the forms online for the first time ever. Those who failed to fill out the form by the deadline were followed up with a face-to-face interview from one of 30,000 Census collectors. Those who fail to fill out the Census survey face a 1,000 fine. The keeper of the Census, the Office of National Statistics (ONS), has always held the ability to reform the way and over which time period information is collected. The ONS is now exercising this right to adopt a reform of the Census and has consulted on radically changing the way in which details of UK households is obtained. There are four key drivers for reforming the Census, according to the ONS: Society changes rapidly and people move around User requirements have evolved Technology has advanced Access to other data sources has improved There is also the issue of cost; the Census currently costs 50 million every year but this cost is vital to the government departments that use the Census to allocate funding and resources in the areas they are most-needed. seeing retirement differently
What are the options? Currently the majority of the information for the Census is collected through postal submissions. For the 2011 Census, the ONS sent out 25 million household questionnaires and this was followed by target field follow-ups to identify households that had not responded. This process is costly in terms of time and resources for the government and the ONS has now put forward two options to shake-up the Census which it has requested comments on in its consultation paper. Option one: once in a decade This option most closely resembles the Census that is currently in use but with slight tweaks to the way information is collected. The Census would still be carried out once every decade, and would be mandatory but instead of sending out postal questionnaires, the survey would be completed online. For those who do not want to complete an online questionnaire, or do not have access to the internet, a telephone service or face-to-face survey option would be available. As now, there would be follow-up doorstep interviews and, like the 2011 Census there would be a followup coverage survey of 1% of the population to account for those who did not respond. Option two: administration data The second option is decidedly more radical, and less costly than the current iteration or option one. It centres not around a main survey but the use of administration resources from various governmental and local authority departments, including the Department for Work and Pensions (DWP), NHS, electoral register, birth and death rates, school registers and higher education details, all held in an anonymised form. This data would be coupled with an annual 1% of population survey to adjust for incorrect administrative data. There would also be a second survey covering 4% of households each year to collect data on details not held by administrative systems, such as characteristics of the population including ethnicity or languages spoken.
Pros and cons There are positives and negatives to both options put forward by the ONS and stakeholders were asked to put forward their comments and flag up any concerns they may have in the consultation. At the first Population Patterns seminar (#partnershippatterns), End of the Census organised by ILC-UK and supported by Partnership - leading statisticians, charities and experts from various fields put forward their hopes and concerns for the future of the Census. Option one: pros Understanding The first option is not a radical departure from the current Census and as such government departments, which use the information to develop budgets and allocate resource, and external stakeholders that depend on different facets of the Census will be provided with the same information they currently receive. Snapshot The once-in-a-decade Census provides a snapshot of the nation as it has done so for 200 years, allowing statisticians and historians to plot the changing demographics of the country. A once-in-a-decade Census will maintain continuity in the UK s records. Granular detail The 10-year Census provides a detailed set of information about the geographic spread of social conditions, pinpointing pockets of deprivation, as well as detailing relationships between individuals who are co-resident. It can specifically identify multi-generational households, and how they are related. The depth of the current Census allows the ONS to drill down into specific, granular detail on small geographical areas and demographics to provide an accurate measure of life in Britain for various age and social groups. We are still using the same data 10 or 12 years later, which creates an illusion of knowledge. Alan Teague, ONS. [The census has built an accumulative demographic history going back 200 years history will come to an end if we do not have a Census! Professor Heather Joshi. Option one: cons Out of date While the Census may cover a large number of fields and allow comparison of unrelated variables, it is largely out-of-date in its practical application. This has been the main criticism of the Census and one argument for reform. Often the data can take two years to collate and publish, meaning Census data can be up to 12 years old. Ageing population Undertaking a census once a decade has a particular impact on the understanding of later life demographics. Over a 10 year period, older generations can be greatly affected by mortality rates and changes in trends will not be spotted quickly enough. In ageing populations, like that of the UK, what is presumed to be known about older demographics can quickly drift away from reality.
In my view we have been doing a type of dual running for 30 years with our use of administrative data increasing over the years. Professor Peter Goldblatt Option two: pros Timeliness The main attraction of moving to an administrative data Census is the timeliness of the information provided. Instead of waiting a decade for up-to-date demographics the ONS has reported a correlation of Census data with administrative data after just five years, meaning trends could be spotted earlier and statisticians and government departments would be working on current, rather than historic, data. Cost Making use of existing departmental data and undertaking a 4% survey annually is cheaper than a detailed Census every decade the 2011 Census cost 480 million and the 2021 Census is expected to cost 800 million under the current system. The ONS has been increasing its use of administrative data over a number of decades, using information to fill in the gaps when the Census response has been poor. Expansion of data In theory the admin data Census could expand the information it looks at and provide a snapshot of areas which are not currently covered, including income data of which could be taken from PAYE and benefit payment systems. There could also be the option to carry out mini follow-up surveys if the main survey shows a particular trend that needs to be explored further. The Census is essential to the running of stable government, the paper Census is inefficient and streamlining needs to be done. Professor Ludi Simpson Option two: cons Data loss While there is the opportunity to expand data collected, there is also the concern that some data will be lost. The Census, for example, breaks down multi-occupancy residency and vital details about caring arrangements but administrative data does not. Neither is there a generic way to log ethnicity in administrative data, meaning information about the make-up of Britain could be lost. Access to data Even for the ONS it can be difficult to access certain databases, such as the DVLA, and difficulties have arisen in accessing higher education data. In order for the admin data Census to be a success, the ONS needs greater permissions to access information from a larger number of departments. Transition needed Moving to an admin data Census would require a period of transition whereby the new and old methods run alongside one another; a possibly costly exercise but one that will be necessary. The option of moving to a new system in time for the 2021 Census could be risky and there are concerns that the cost of getting it wrong have not been adequately considered. Other countries that have moved to an admin data Census have taken decades to do so.
Wider impact The insurance industry relies heavily on the information the Census provides about demographics, particularly the UK s ageing population. An unintended consequence of reforming the Census could be the loss of information that is vital to understanding Britons future lives, longevity, and health. The government uses the Census data for state pension planning and insurance companies use the information to understand how long pensions, annuities and other insurance policies will be in use for. Both administrative data Census and 10-year Census options impact the insurance industry. Moving to an administrative data Census would allow insurers to pick up trends in older age groups and mortality more quickly, but a 10-year Census allows more complete data to be collected albeit less often. Other private sector companies and charities will be similarly impacted by a change in the Census. For example charity Age UK relies heavily on the Census to plan for advice centres, using both details of older populations and transport. Points to consider The use of administrative data needs further trialling before widespread introduction. Increasing administrative data should be used alongside the Census. Abolition of the regular Census would impact data collection on carers and caring, multi-generational households and commuting and travel patterns. The regular Census could be abolished by 2031, but government must ensure funding for development of other longitudinal data sets and testing of administrative data. Government must open a nation-wide debate on data collection, including the use of commercial data, such as mobile operators, and information from the private sector. Discussion of privacy issues must be part of the debate and government should look at whether consumers or companies own data and whether the ONS should access it. Government must ensure as it outsources services, access to public data does not become private. ONS and government must continue recent trends towards open access of data; the more we know about our ageing society, the better we will be able to adapt.
The future of the Census The Census is a consistently evolving medium for tracking UK demographics. In future even more radical ways to collate information about the population may need to be introduced. These could include: Commercial data One criticism of government data is that it may not be as comprehensive as that collected by large companies that target customer data, such as Tesco and Facebook. It would not be beyond the realms of possibility for the ONS to make use of commercial data to help boost its own administrative data sets, although there would be cost implications, accuracy issues and public acceptability of such a move. Unique record A development of the online Census could be a unique, individual online record that must mandatorily be updated online to record changes in employment, benefits, education and other factors deemed important. Individuals would be legally required to upload information. Final thoughts Consultation on the Census gives the government a chance to reform the way data is collected in the UK and adapt it to better suit and track the rapidly changing population. However, it must not dismiss the importance of the existing Census nor the datasets collected across a vast range of demographics while looking at how it can be made more timely and what additional datasets which could be collected to drill down deeper into the social, spending and policy consequences of the UK s ageing population. The general consensus of the Population Patterns meeting was that a hybrid option of both option one and option two would be the most desirable outcome. The ideal Census would incorporate the granular detail of the 10-year Census while also utilising the timeliness of the administrative data Census. This would ensure that no vital detail on UK demographics would be lost while allowing trends and changes to be tracked more closely and picked up more quickly. The government is due to make a decision on the future of the Census in 2014 following a consultation with various stakeholders. Population Patterns Over the next 12 months ILC-UK supported by specialist insurer Partnership Assurance Group plc will undertake a series of events exploring the impact of demographic change on public policy. Seminar Series
ILC UK 11 Tufton Street London SW1P 3QB Tel : +44 (0) 20 7340 0440 www.ilcuk.org.uk Published in January 2014 ILC-UK Registered Charity Number: 1080496. seeing retirement differently