Use of administrative data at Statistics Finland Ilkka Hyppönen Statistics Finland
Mar 27, 2015
Use of administrative data at Statistics Finland
Ilkka Hyppönen
Statistics Finland
11.8.2005 2Ilkka Hyppönen
Structure of presentation:
Statistics based on registers
Use of administrative data
Enterprise statistics
Other econ stat
Population and social statistics
Other statistics
General considerationsSystems considerations
11.8.2005 3Ilkka Hyppönen
Register based statistics production
Use of administrative data
is closely linked with
Register based statistics production philosophy
11.8.2005 4Ilkka Hyppönen
Business Register
Population Register
Buildings and Dwellings Register
Population, IDs, classifications, some basic data
Various base statistics
Concepts Data
Statistical systems (like SNA)
Statistics' production based on registers
Statistics basedon registers
11.8.2005 5Ilkka Hyppönen
Business Register
Population Register
Buildings and Dwellings Register
Organization numberTax authorities
Establishment numberStatistics Finland
ID-systemsin 1990’s (1980’s)
Personal Identification number
Population register center
ID-systemsin 1960’s
Building / apartment number
Population register center
ID-systemsin 1980’s
Statistical Base registers
Statistics basedon registers
11.8.2005 6Ilkka Hyppönen
Register of educationalinstitutions
Register ofqualificationsand degrees
Organization numberTax authorities
Establishment numberStatistics Finland
ID-systemsin 1990’s (1980’s)
Personal Identification number
Statistics Finland
ID-systemsin 1960’s
Statistical Base registers
Statistics basedon registers
11.8.2005 7Ilkka Hyppönen
ID number = identifier schemes essential
Unique identifier schemes Wide usage in administration, in enterprises, in pension
schemes, in hospitals, in schools etc. Some kind of BASIC register (mother register) for assigning ID-
numbers (in administration)
Without unique ID’s use of administrative data difficult or impossible: - definite identification impossible - double counting
Statistics basedon registers
11.8.2005 8Ilkka Hyppönen
Statistical units
employee
Legal unitEstablishment
Employer
age, sex, wages, occupation, living address Through links: Qualification / degree, ... activity, location of work
place, ...
activity code, location, ...
Statistics basedon registers relationship
link for multi-establishment firms from direct survey
Building
Educationalinstitution
11.8.2005 9Ilkka Hyppönen
Statistical units and attributes (data) derived through links
Statistics basedon registers
For population statistics: for persons: NACE, location of work place, size of enterprise where works etc.
For business statistics: for enterprises: Number of employees, wages per employee,
structure of work force: sex, occupation, education, ..
For educational statistics: For educational institutions / curricula: ex-students: where they work, occupations, level of income ...
11.8.2005 10Ilkka Hyppönen
PreconditionsUniversal ID-chemes-Persons-Organisations-Buildings, apartments
Acceptance by the people, businesses and administration
WIDE use
Strict confidentiality in statistics
Up-to-date legislation - statistical law - personal information protection
Well developedIT infrastructurein administration
Possibility to use in statistics,
alsoto
COMBINE
Statistics basedon registers
11.8.2005 11Ilkka Hyppönen
Use of administrative data in statistics
11.8.2005 12Ilkka Hyppönen
Main reasons for using admin data are - reduction of response burden - reduction of costs of statistics - to have total populations ---> more detailed classifications are possible ---> more reliable “totals”The Finnish Statistics act: It is compulsory to use existing data (if suitable). State government and social security institutions are obliged to deliver the data they have to Statistics Finland
Reasons
Use of administrative data
11.8.2005 13Ilkka Hyppönen
The administrative concepts are very seldom exactly the same as the statistical concepts
Essential is, whether administrative concept correlates closely enough to the statistical concept. If so,
the development or the state of the social phenomena can be described using administrative dataor the statistical variable can be estimated from the administrative variable / variables
It is essential to change the way of thinking
Use of administrative data
11.8.2005 14Ilkka Hyppönen
Use of administrative registers and data in statistics:
About 94 % of INPUT data at StatFi comes from administrative sources (as measured in number of
stat units times number of variables)
Typically there is some direct data collectionin every business statistics
Typically direct data collection isONLY from large enterprises
For local government units,direct data collection is typical
Use of administrative data
11.8.2005 15Ilkka Hyppönen
Data collection in 2004 for official statistics (includes all statistics)
Total number of data collections: 189 - administrative data: 73 - interview 8 - other direct data collection 108 - on paper 35 - on WEB forms 41 - in electronic form (files etc) 32
paper, WEB and electronic form are double counted to some extent
Use of administrative data
11.8.2005 16Ilkka Hyppönen
By statistical area
Enterprise statistics
Use of administrative data
Enterprise statisticsOther economic statisticsPopulation and social statisticsOther statistics
11.8.2005 17Ilkka Hyppönen
Number of enterprises Direct collection Administrative
data
Structural Business Statistics 8000 180 000
Business Register 32000 over 250 000
Short Term Business Statisticsturnover 2000 250 000wages and salaries 350 90 000
Even for enterprises in direct collection some data are taken from administrative sources
Direct collection vs. use of administrative Datafrom tax authorities in statistics on enterprises
Use of administrative data: Enterprise statistics
11.8.2005 18Ilkka Hyppönen
Common accounting data surveys with Bank of Finland and the Financial Supervision Authority --> Administrative data where Statistics Finland has had a considerable influence
Financial statistics
Use of administrative data: Enterprise statistics
11.8.2005 19Ilkka Hyppönen
VAT value added tax declarations data (monthly) --> turnover (STS) ---> estimates of turnover class (Business register)
Employers wage payment data (monthly) --> wages and salaries (STS) --> estimates of number of employees (Business register)
Company tax (yearly accounts) --> turnover etc. (SBS, Business register)
Employers declaration on wages and salaries paid for each employee (yearly) --> estimates of man-years (Business register)
Customer register of Tax authorities --> names, addresses, ... (Business reg.)
Individual tax forms --> income, expenditure, assets, … (agricultural enterprises)
Use of tax data in business statistics
Use of administrative data: Enterprise statistics
11.8.2005 20Ilkka Hyppönen
Building permits, starts, completions --> floor area of building permits, new orders of housing construction (STS)
Location co-ordinates (Business register)
Use of population register centre data for enterprise statistics
Use of population statistics data for enterprise statistics
Use of employment statistics data (occupation, education) in estimates of man-years (Business register)
Use of administrative data: Enterprise statistics
11.8.2005 21Ilkka Hyppönen
Enterprise group relationships from public accounts of the groups Manual data
Use of trade register data for business register
Use of vehicle register data for goods transport statistics on roads
The sampling unit is a heavy goods transport vehicle The sampling frame is the vehicle register It gives names and addresses and data on the vehicle
Use of administrative data: Enterprise statistics
11.8.2005 22Ilkka Hyppönen
Other economic statistics
By statistical area
Use of administrative data: Other econ stat
11.8.2005 23Ilkka Hyppönen
Prices of dwellings (property transfer tax data)
Real estate prices (National Land Survey)
Telecommunications (partly admin data)
Patenting (patent register)
Use of energy (partly private sources)
Use of administrative data in other economic statistics
Use of administrative data: Other econ stat
11.8.2005 24Ilkka Hyppönen
State and local government (pension institutions) Private employers’ organisations (for about half of the number of wage earners)
Use of wage statistics of employers’ organisations
Use of administrative data: Other econ stat
11.8.2005 25Ilkka Hyppönen
Population and social statistics
Underlying all statistics on persons and households, is all the combined data from population register, register of buildings and dwellings taxation of income and property of persons, pension schemes, register of qualifications and degrees, and employment statisticsData in these registers / data files need not to be surveyed directly
Also underlying, where relevant, is the Business Register
By statistical area
Use of administrative data: popul and social stat
11.8.2005 26Ilkka Hyppönen
Population and housing
Population statistics (population register) Building and dwelling statistics (buildings and dwellings
register) Statistics on housing conditions (dwellings register,
population register)
Cause of death statistics (death certificates)
Use of administrative data: popul and social stat
11.8.2005 27Ilkka Hyppönen
Employment statistics
pension insurance schemes (private, central and local government etc.)
pension registers taxation registers (employer-employee data, etc.) register of unemployed job-seekers military service register student registers register of qualifications and degrees Business Register, register of government units
Use of administrative data: popul and social stat
11.8.2005 28Ilkka Hyppönen
Employment statistics (cont.)
Register of buildings and dwellings
Plus a direct survey on enterprises with multiple establishments and on government units to establish employee -- establishment link
Use of administrative data: popul and social stat
11.8.2005 29Ilkka Hyppönen
Population census
Is produced in Employment statistics Population statistics Buildings and dwellings statistics and partly combining data from these statistics
andoccupation data (mostly administrative data / wage statistics data, partly surveyed directly)
Use of administrative data: popul and social stat
11.8.2005 30Ilkka Hyppönen
Justice, education, culture
Statistics about justice and crime (16 different statistics / data from Ministry of Justice information systems)
Election statistics (data from Ministry of Justice information systems)
Education statistics (mostly direct data collection, partly administrative data)
Cultural statistics (data from various authorities)
Use of administrative data: popul and social stat
11.8.2005 31Ilkka Hyppönen
Other social statistics
Income distribution statistics (various admin data combined with survey data)Income and property statistics (tax data)Household assets (tax data, survey data)
occupational accident statistics (accident insurance institutions)
Use of administrative data: popul and social stat
11.8.2005 32Ilkka Hyppönen
Other statistics
By statistical area
Use of administrative data: Other statistics
11.8.2005 33Ilkka Hyppönen
Use of waste data of the The Compliance Monitoring Data System of the Finnish Environment Institute
Waste statistics / manufacturing, Air emissions
Motor vehicle stockMotor vehicle new registrations
Use of administrative data: Other statistics
Use of Motor Vehicle Register data
Use of Police incident recording systemroad traffic accidents
11.8.2005 34Ilkka Hyppönen
Coordination, cooperation
Use of administrative data: general
Meetings at DG levelwith ministries and otherauthorities
Register pool permanent co-operationbetween major registerholders
Co-operation officersat Statistics Finland
11.8.2005 35Ilkka Hyppönen
Coordination, cooperation (cont.)
A major achievement:
SBS have a common form on yearly
profit and loss account and balance sheet
with the Tax authorities
To influence the contents of administrative data -- e.g. classification of buildings -- inclusion of statistical classifications e.g. NACE -- other contents
Use of administrative data: general
11.8.2005 36Ilkka Hyppönen
Problems
Concepts --> administrative
Data contents --> - only those relevant to the authority in question
Slow (typically)
Not under our own control --> strong dependence --> need for co-operation
Problems and advantages of administrative sources
Advantages Total populations
--> representative--> detailed classification of units--> also small area statistics
Only marginal costs
No response burden
Is deemed rational by the society
Use of administrative data: general
11.8.2005 37Ilkka Hyppönen
Problems
Administrative simplification efforts --> reduce data contents--> reduce periodicity
Final VAT, Intrastat, ...
General attitudes--> against registration of persons
EU / harmonisation may lead to changes in administration --> changes in data systems AND lines of action in administration
Future
ActionsIncreasing co-operation with
administration
Probably increasing direct data collection (speed, data contents)
Increasing methodological work (like imputation for missing variables)
Use of administrative data: general
11.8.2005 38Ilkka Hyppönen
Present situation
Identifiers of statistical units
Common identifiers used in all data systems
Data systems
Basically “separate system for each statistics” ; “stove-pipe” approach
Classifications
Basically taken from base registers
Statistical units
Basically taken from base registers
Information system architecture
Data copied from one system to
another Basically, “shared” data is copied to
each data system (usually no update anomaly)
Business register has about 300 users (of 600 persons employed in statistics divisions)
In population statistics, price statistics, labour force survey, national accounts, business statistics, etc.
Use of administrative data: systems
11.8.2005 39Ilkka Hyppönen
Present situation
Use of administrative data
Administrative data are acquired by the “responsibility area” which primarily uses the data
This organisational unit is the “owner” of these data -- data security, support for other uses etc. Other users apply for a permit to use the data or the data derived from the original data
Information management
Use of administrative data: systems