1 Establishing a registerbased statistical system Example Population

1 Establishing a register-based statistical system Example: Population and housing censuses in Norway Training workshop on censuses using administrative registers in Geneva 21 May 2012 Harald Utne, Statistics Norway harald. utne@ssb. no 1

Norwegian Censuses 1960 - 2011 • 1960: Last fully traditional census Census data used to establish CPR (in 1964) • 1970 Census: Mainly traditional Census used to – Check and update CPR – Establish Register of Education (in SN) • 1980 Census: Questionnaires shortened and sent by mail – Registers used to rationalise data collection: Mail out – mail back w No enumerators – Demographic variables taken from CPR 2

Norwegian Censuses 1960 – 2011 (2) • 1990 Census – Register data on demography, education, income, geographical characteristics – Questionnaires for data on labour marked, households and housing w Sample survey w Labour market register used in estimation • 2001 Census – All data on persons from registers – Data on families and households based on registers and questionnaires – Housing data based on questionnaires – Full coverage survey (all households) • 2011 Census: Totally register-based 3

Historical development, main trends: 1. Extensive use of census data in establishing registers – – Use of “census data” for administrative purposes: Multiple legal base for Census w 2. 2001 in Norway: Statistics Act, CPR, GAB Step-by-step development – Developing a register-based census may take time! 4

Register-based population and housing census system Other registers Education Social security Income P P P Unemployed persons P P Population Reg. P Job registers A B Business register A Addresses Buildings Dwellings P = Personal id number B = Business id number A = Numerical address • Units covered in base registers: Persons, dwellings/buildings and enterprises/establishment • • Derived units: Families and households Persons linked to – Dwellings – Workplaces (establishments) through job registers 5

Census data from registers – “tertiary use” Adm. registers Employees Wages Statistical registers in Employment Statistics Norway Census Income CPR Population Education CENSUS FILE 6

Census data system before 2001 Other registers Education Social security Income P P P Unemployed persons P P Central Population Reg. P Job registers A B Business register A Addresses Buildings Dwellings P = Personal id number B = Business id number A = Numerical address No register on dwellings No link person- dwelling, that is no household register 7

Norwegian 2001 Census Goals • Produce Census statistics for 2001 • Establish a basis for the 2011 register-based census – The Dwelling Address Project 8

The Dwelling Address Project • Major goals – Extend the Cadastre to comprise dwellings, not only buildings w Add dwelling number to street address w Main Street 8 A, H 0103 w Establish a link between dwellings and resident persons (in CPR) w Using data from census forms • Main problem: Multi-dwelling buildings • Participants – Statistics Norway (Census 2001) – Tax Inspectorate (CPR) – Norwegian Mapping Authorities (Cadastre) • Use “census data” for administrative purposes – Multiple legal base for Census 2001 9

Multi-dwelling buildings before Census 2001 ? H 0102 H 0101 H 0103 H 0104 Main street 8 A 10

Multi-dwelling buildings after Census 2001, goal H 0102 H 0101 Main street 8 A H 0101 H 0103 H 0104 BUT------ 11

Multi-dwelling buildings after Census 2001, result H 0102 H 0103 H 0101 H 0104 Main street 8 A H 0101 Only 55 % of persons in multi-dwelling houses linked to a unique dwelling address. Following up necessary! 12

Following up on the Dwelling Address Project • Statistics Norway playing an active role – pushing forward • Quality monitoring • Situation by end of 2010: 95 % of population linked to a dwelling 13

Comparing register-based and traditional censuses Advantages • Reduced response burden • Reduced costs – Specific census costs rather low • No need to hire and train a census staff every tenth year • Census statistics and subject matter statistics based on same sources – Comparability – Annual “census statistics” 14

Comparing register-based and traditional censuses Disadvantages • Only variables available in or derived from registers included • Some restrictions on definitions – Dwelling households, not housekeeping units – Legal place of residence, not usual place of residence • No collection tool for emerging needs – Sample surveys may be used, but not for small area statistics 15

16 Register-based censuses – the process Traditional Costs Relevant registers? Relevance Response burden Linking? Accuracy Annual statistics ? Investigation Quality Legal base? Coverage Public approval? Timeliness Partly registerbased Fully traditional New technologies Traditional and Sample surveys or Yearly updates Registers and Sample survey Registers and Traditional Reg. and ad hoc Sample survey Reg. and existing Sample survey Fully registerbased 16

Thank you for your attention ! 17