Estimating the number of rooms and bedrooms in the 2021 Census for England and Wales. An alternative approach using Valuation Office Agency (VOA) data

Transcription

1 Estimating the number of rooms and bedrooms in the 2021 Census for England and Wales An alternative approach using Valuation Office Agency (VOA) data Marie Haythornthwaite Administrative Data Census Team

2 Background Topic consultation 2016: ONS felt it was not appropriate to ask two questions on number of rooms in 2021 Census: - Medium score on user need - Respondent difficulty in understanding definitions (66% agreement for number of rooms in 2011 Census Quality Survey) Commitment made to explore Valuation Office Agency (VOA) data as an alternative source: - VOA are responsible for banding properties for Council Tax purposes - Contains variables for number of rooms and bedrooms 2021 Census White Paper to be published in Spring 2018: - Recommendation for rooms/bedrooms questions will be based on potential of VOA data and user feedback on impact of change - Consultation period (for feedback from users) is from 28 June 2017

3 Matching VOA data to 2011 Census G B 86.5% A 2011 Census 24,429,618 records post-edit & imputation (household spaces) 23,066,513 matched 2011 Census and VOA records (94.4% of census records) VOA dataset 25,232,551 records F D C E A B C D F Census addresses (pre with that built duplicate 2011) could after missing 2011 not UPRNs be (715,157) a assigned UPRN (73,760) (394,732, (44,222) a UPRN 1.6% (424,962, census records) 1.7% census E 86.5% = VOA Census records duplicates with a unique (A) match to that VOA could records not with be matched a unique to UPRN census that (1,332,899) records) G couldn t = Census be matched records with to census a unique (F) UPRN that could not be matched to VOA (543,411) 3

4 Definitional Differences Households vs. Addresses: Household = One person living alone or group of people living at the same address who share cooking facilities, living room, sitting room, dining area Can distinguish multiple households that occupy the same address / houses in multiple occupation (HMOs) Cannot identify HMOs from admin data because we do not have information about residents and relationships between them. UPRN assignment at address level 4

5 Definitional Differences cont. Number of rooms 5

6 Definitional Differences cont. Number of bedrooms 2011 Census: any room intended to be used as a bedroom when property was built, or any room that s been permanently converted for use as a bedroom VOA: rooms built as bedrooms, even if they are not being used as such. Excludes rooms incapable of comfortably holding a single bed (approx. anything less than 2m x 2m) 6

7 Distribution of differences in number of rooms between VOA and 2011 Census, addresses in England and Wales 7

8 Distribution of aggregated differences in number of rooms between VOA and 2011 Census, addresses in England and Wales 8

9 Distribution of differences in number of bedrooms between VOA and 2011 Census, addresses in England and Wales 9

10 Occupancy rating A census measure of whether a household's accommodation is overcrowded or under occupied based on census variables e.g. age, sex and relationship composition of the household. Occupancy rating = Number of Rooms/bedrooms Rooms/bedrooms required Occupancy rating for rooms: VOA at least one room less than census Rooms required is derived from census (so includes kitchens) Therefore, majority of addresses generally fall within overcrowded category when deriving occupancy rating for rooms using VOA data 10

11 Number of addresses by occupancy rating for bedrooms, 2011 Census and VOA, England and Wales. 11

12 Distribution of addresses by occupancy rating for bedrooms, 2011 Census and VOA data, England and Wales 12

13 Summary Good coverage of addresses in England and Wales (94% census records matched to VOA dataset) Agreement rate of 85% for number of rooms when considerations made for definitional differences and CQS findings Direct Agreement rate of 76% for number of bedrooms Similar results for occupancy ratings for bedrooms compared to 2011 Census Based on initial analysis, it is ONS view that with some further development, VOA data is of sufficient quality to replace one or both questions on rooms and bedrooms in the 2021 Census. 13

14 Next steps Publish initial report, supporting data and consultation seeking user views on 28 June 2017 Explore potential of VOA data to produce new, more regular statistics on housing data in an Administrative Data Census context Make recommendation for rooms/bedrooms question (based on potential of VOA data and user feedback on impact of change) in Autumn Census White Paper to be published in Spring

15 15