Beyond 2011 – A new paradigm for population statistics? Pete Benton, Beyond 2011 Programme Director Office for National Statistics, UK
Programme Purpose Established April 2011 To identify the best way to provide small area population and socio-demographic statistics in future Provide a recommendation in with cost-benefit analysis - design for implementation
Context Census – every 10 years for over 200 years The potential for change - Rapidly changing society - Evolving user requirements - Technological advances - Improved data sources - Improved data access Successful 2011 Census - The only approach possible in 2011
Approach Review user needs Consider possible census approaches Evaluate using published criteria, including: - cost - statistical quality - social and economic benefit - public acceptability - risk Publish findings in 2014 Final decision for Government and Parliament
Beyond 2011 : Statistical options Aggregate analysis 100% linkage to create ‘statistical population spine’ (Intermediate) Sample linkage e.g. 1% of postcodes Address register + Survey Administrative data options Traditional Census (long form to everyone) Rolling Census (over 5/10 year period) Short Form (everyone), Long form (Sample) Short Form + Annual Survey (US model) Census options Survey option(s)
Beyond 2011 : Statistical options 100% linkage Administrative data options Traditional Census (long form to everyone) Census options 4% survey every year
Two potential future approaches A census once a decade, primarily online Provides detailed statistics once every ten years with updated population estimates each year A census using existing government data and compulsory annual surveys Provides more statistics each year, but less detail than an online census once a decade
Administrative data and surveys Potential data sources: NHS Patient Register DWP/HMRC Customer Information System Electoral roll (> 17 yrs) School Census (5-15 yrs) Higher Education Statistics Agency (Students) Birth and Death registrations All sources anonymised prior to linkage
Major statistical challenges No population register Data quality Incomplete Out of date Duplicates / erroneous entries Little reliable population attribute data Coverage survey required Attribute survey required
2011 Patient register population counts compared with 2011 Census population estimates Patient register is … More than 13 % lower 8.5% to 13% lower 3.8% to 8.5% lower Within 3.8% 3.8% to 8.5% higher 8.5% to 13% higher More than 13 % higher
Tax register population counts compared with 2011 Census population estimates Tax system is … More than 13 % lower 8.5% to 13% lower 3.8% to 8.5% lower Within 3.8% 3.8% to 8.5% higher 8.5% to 13% higher More than 13 % higher
Linked admin data compared with 2011 Census population estimates Linked admin data is … More than 13 % lower 8.5% to 13% lower 3.8% to 8.5% lower Within 3.8% 3.8% to 8.5% higher 8.5% to 13% higher More than 13 % higher
Population Pyramids using linked admin data SPD 5
Population characteristics – potential outputs 4% random sample per year Gives Coefficient of Variation of 20% for 800 people using 1 year’s data Any geographic level Any cell in a cross-tabulation 320 people using 2 years’ data 230 people using 3 years’ data 130 people using 5 years’ data Additional topics from administrative data? E.g. Household income?
Public Consultation 23 Sept to 13 December Series of user and other events over next three months Consultation Document (and lots of supporting material and research papers) on ONS website Views welcome from anyone
Key risks of a census using administrative data and surveys Public opinion Data quality / technical challenge Changes in administrative datasets Managing the transition if moving to a census using administrative data Harmonisation of UK outputs
Next steps Review consultation responses Consider legislative implications Consider risks and potential transition plans Consider implications for wider statistical system (surveys; IT systems) … Conclusions in 2014