Use of the administrative data in Population Censuses

Use of the administrative data in Population Censuses Libourne, 4 - 7 June 2007 Kaija Ruotsalainen, Statistics Finland

Content 1 Administrative data and registers as sources of statistics l Advantages and disadvantages of register-based statistics production 2 Use of register data in statistics compilation Ways of utilising registers l Defining the families l Definition the main type of activity l 3 Data quality 4 New statistics made possible by the use of registers 7. 6. 2007 2

Advantages of register-based statistics production Total data can be produced annually n Lower costs (1980 Finnish Census cost EUR 34 million, 2000 Census EUR 840 000. n Because of the identifying codes it is possible to compile the flowstatistics n Possibility to produce more data (for instance multiple activities) n Possibility to link with other material n No subjective opinions => All the cases are handled in the same way n No additional burden to respondents n Fewer personal problems n Confidentiality advantages n 7. 6. 2007 3

Disadvantages of register-based statistics production Dependence on data suppliers n An amendment in legislation or other administrative change can cause changes to the data content of the register n Changes in data suppliers’ information system can bring problems, at least delays from the normal production rate n The coverage of registers may be defective for some data, although the register itself would contain all the units n Consistency problems may arise when linking information from different registers n All the information that previously used to ask on census forms not exist in the register n 7. 6. 2007 4

Use of register data in statistics compilation n Register data are utilised primarily in two different ways: 1) Obtain the information directly from the register e. g. person’s demographic data (age, gender, marital status, citizenship) or income data 2) Another way is to form new variables with the so-called register estimation method 7. 6. 2007 5

The register estimation Bases on different information in several registers n The aim is to estimate for each statistical unit the value of the target variable as close to the statistical concept and definition as possible n This is done by using all the existing data available and a set of decision rules to estimate the value of the statistical variable n 7. 6. 2007 6

The register estimation n The sources: partial coverage l overlapping l different degrees of quality l l supplement each others The definition takes place in steps n The sequence of the steps is based on three principles: n international definitions l quality of administrative sources l other reasons l 7. 6. 2007 7

Examples of using register data in statistics compilation defining the families n defining the person’s main type of activity n 7. 6. 2007 8

Definition of family (1) n n A family consists of a married, cohabiting or registered couple and their children living together; or either of the parents and his or her children living together; or a married, cohabiting or registered couple without children. Persons living in the household-dwelling unit who are not members of the nuclear family are not included in the family population, even if they are related, unless they form their own family. Brothers and sisters or cousins living together are not a family and do not belong to the family population. The same applies to people who live alone. 7. 6. 2007 9

Definition of family (2) n n Families living in residential homes are included in the family population. In contrast, persons who live in institutions are not included. A family can consist of no more than two successive generations. If the household-dwelling unit comprises more than two generations, the family is formed starting from the youngest generation. This means, for example, that a mother-in-law or father-in-law living with their child's family will not be included in the family population unless they live together with their spouse, in which case the old couple form their own family. 7. 6. 2007 10

Definition of child n n In the family statistics children comprise the following persons living with their parents: l biological children l adopted children and l biological children and adopted children of one of the spouses Foster children and children in the care of the family are not classified as children. 7. 6. 2007 11

e Ch sta te e Bu ck ild cod ing e En tra nc e Fl & at f div lat ide r al Re 004 004 004 k oc Bl ge lla 001 001 001 0001 0001 0001 Vi M un ici pa lity Computerised forming of families (1) 1 1 1 001 001 001 A 001 A A A Person Person 1 2 3 4 5 6 004 002 0001 1 002 A 001 A 004 002 0001 1 002 A 001 A Person 1 Person 2 Person 3 004 002 0001 1 002 A Person 1 • Population is sorted by domicile code (DC). DC (23 digits) identifies a dwelling Household 1 Household 2 Household 3 etc. 992 111 0001 1 019 A 020 A Person 1 Household N 7. 6. 2007 12

Computerised forming of families (2) n n n The computer program inserts into a table the inhabitants of one household-dwelling unit. The record of each person includes the personal identification codes (PIN) of his or her parents, married spouse and all born and adopted children. The computer program forms families by comparing the PINs of the persons living permanently in the same household-dwelling unit. 7. 6. 2007 13

Computerised forming of families (3) Household 1: 004 001 0001 0001 004 001 0001 1 1 1 001 001 001 A 001 A A A Household 2: 004 002 0001 1 002 A 001 A 004 002 0001 1 002 A 001 A Grandfather Grandmother Family 2 Son 1 Son 2 His wife Family 1 Their child Husband Family 1 Wife Her mother Not a family member 7. 6. 2007 14

Computerised forming of families (4) n n After forming a possible married family the program identifies as a cohabiting couple a man and woman living together in the same dwelling if: - both are aged 18 or over - their age difference is less than 16 years - they are not siblings The program also identifies reconstituted families by comparing the PINs of the children and parents of each person in the same dwelling. 7. 6. 2007 15

Definition of the main type of activity Labour force l employed l unemployed n Persons outside the labour force l 0 -14 -year olds l students l conscripts l pensioners l other persons outside the labour force n 7. 6. 2007 16

Defining the main type of activity (1) A cross-sectional data for the last week of the year n The validity of the data in the last week of the year is not totally certain for all data n Nearly 30 different registers are used n The classified population diminishes after each category until all the persons have been classified to some group n 7. 6. 2007 17

Defining the main type of activity (2) The decision rules were formed by utilising the parallel statistics of the 1985 Census for testing the decision in different ways and in different order n The aim was to come up with such decision rules that n 1) the numbers of persons in different groups would be as close as possible to the data in the questionnaire-based census 2) the proportion of those belonging to the same group would be as high as possible 7. 6. 2007 18

Defining the main type of activity (cont. ) STEP 1 n All persons under fifteen year of age => • 17. 5 % of the population (in 2004) 0 -14 -year-olds 7. 6. 2007 19

Defining the main type of activity (cont. ) STEP 2 n All persons over 74 years => n 7. 3 % of the population (in 2004) Pensioners 7. 6. 2007 20

Defining the main type of activity (cont. ) STEP 3 n All those who according to the Ministry of Labour's register on job applicants were unemployed the last weekday of the year n 5. 7 % of the population (in 2004) => Unemployed 7. 6. 2007 21

Defining the main type of activity (cont. ) STEP 4 n All those who were on their military service or non-military service during the last week of the year n 0. 3 % of the population (in 2004) => Conscripts 7. 6. 2007 22

Defining the main type of activity (cont. ) STEPS 5 - 7 n n n Persons who have only a valid selfemployment pension’s insurance during the last week of the year (4. 1 % ) Persons who have both a valid selfemployment pension’s insurance and employment pension insurance (employee relationship) and whose entrepreneurial income is larger than earned income (0. 2 %) Persons who according to registers do not have a valid employment relationship but whose entrepreneurial income is higher than EUR 810 and larger than earned income and the person is not retired (0. 2 %) => Entrepreneurs 7. 6. 2007 23

Defining the main type of activity (cont. ) STEPS 8 - 9 n n Persons who have a valid employment pension insurance and who have earned income (37. 8 %) Persons who do not have employment relationship data but who are employed with labour policy measures (0. 1 %) => Employees 7. 6. 2007 24

Defining the main type of activity (cont. ) STEPS 10 - 13 n n Persons who according to the student register study full-time in the autumn term (6. 0 %) Those 15 -year-olds who were not classified into any other group at earlier stages (1. 1 %) Persons in labour market training in the last week of the year (0. 3 %) Persons who have received study aid either in the autumn term or both in the spring and autumn terms (0. 2 % ) => Students 7. 6. 2007 25

Defining the main type of activity (cont. ) STEPS 14 - 15 n n Persons who at the turn of the year receive old-age, disability, unemployment or special farmers’ pension (14. 8 % ) All those whose pension is higher than their income from work and above EUR 580 and the pension is not a survivor's pension or part-time pension (0. 2 %) => Pensioners 7. 6. 2007 26

Defining the main type of activity (cont. ) STEP 16 n Persons whose earned income is higher than entrepreneurial income and over EUR 5600 per year and who were not classified into any other group at earlier stages => n 0, 8 % of the population (in 2004) Employees 7. 6. 2007 27

Defining the main type of activity (cont. ) STEP 17 n Persons who do not meet any of the above criteria => n 3. 4 % of the population (in 2004) Other persons outside the labour force 7. 6. 2007 28

Data quality Municipal pilot study based on 1980 Population Census n Register-based statistics in connection with 1985 census n Evaluation study of the 1990 census n Continuous quality assessment l Labour force survey as reference material l Two purposes: n Ø monitoring of the level of the results Ø monitoring of the extent to which the methods produce data classified in the same manner 7. 6. 2007 29

Monitoring of the extent to which the methods produce data classified in the same manner to identify errors in data processing n to identify situations requiring a change in decision rules n to check the level of results n 7. 6. 2007 30

Main type of activity of the population according to the registers and questionnaire in 1985 census (percentages) 7. 6. 2007 31

Main type of activity of the population according to the register based census and evaluation study 1990 (percentages) 7. 6. 2007 32

Main type of activity by questionnaire and by register 1985 7. 6. 2007 33

Main type of activity by questionnaire and by register 1990 7. 6. 2007 34

Employed according to the Labour Force Survey and Register-based Employment Statistics 1989 -2003 7. 6. 2007 35

Persons according to the Register-based Employment Statistics (RES) and Labour Force Survey (LFS) on December 2002 (persons) 7. 6. 2007 36

Persons according to the Register-based Employment Statistics (RES) and Labour Force Survey (LFS) on December 2002 (percentages) 7. 6. 2007 37

Unemployed according to the LFS and the RES 1989 -2003 7. 6. 2007 38

Unemployed according to the RES and LFS in December 2002 Target population: LFS sample in December 2002, 12 497 persons Unemployed in RES 886 Unemployed in LFS 489 Of which in LFS: Employed Students Pensioners Others Non response 118 15 33 187 188 Of which in RES: Unemployed in both statistics: 345 Employed 39 Students 66 Pensioners 4 Others 35 7. 6. 2007 39

Use of interview survey data with data from administrative sources (1) n The role of register data sampling l primary data collection l editing and imputing l weighting l quality control l n Basic principles informed consent l ethics l 7. 6. 2007 40

Use of interview survey data with data from administrative sources (2) long traditions n especially demographic data and income data n Why? n need for data of high quality n request for cost effectiveness n risk of growing non-response 7. 6. 2007 41

New statistics made possible by the use of registers For instance: n Flow statistics l employment flows l student flows l flows between branches of industry etc. n Placement statistics n Longitudinal researches 7. 6. 2007 42

7. 6. 2007 43

7. 6. 2007 44

7. 6. 2007 45

Employment rate of persons graduated of healthcare in the first stage of tertiary education (nurses, midwives, laboratorian etc. ) in 1989 -2003 by year of graduation 7. 6. 2007 46

Deaths and the history of different activities Pensioners Deaths 1971 -2005 by causes of death Completed educ. Family background Region Income level History of Employment 1950, 1970, 75, 80, 85, 87 -2004 Industry Housework Occupation Children Socio/econ Incomes Students Level of ed. Field of ed. Unemployed Births Immigr. 7. 6. 2007 47

7. 6. 2007 48

7. 6. 2007 49

More information http: //www. stat. fi/censusbyregisters 7. 6. 2007 50