EUROPEAN COMMISSIONEUROSTATDirectorate F: Social statistics and Information SocietyUnit F-3: Living conditions and Social protection Statistics
Community Statistics on Income and Living Conditions (EU-SILC)
Workshop �Monitoring social Europe�
16-17 November 2005University of Mannheim
Introduction� The successor of the ECHP (1994-2001)� EU reference source for:
! Comparative income distribution! Social exclusion statistics
� Key tool for policy makers to monitor the Lisbon Strategy - primary data source for the computation of the: ! structural indicators of social cohesion! Laeken indicators! indicators of pensions adequacy! sustainable Development indicators of
poverty and of ageing
Introduction
� Covering 25 MS from 2005 plus Norway, Iceland and Turkey (Bulgaria, Romania from 2006)
� Two dimensions! A cross-sectional dimension" Yearly data - First priority
! A longitudinal dimension" Yearly data - More limited
Introduction
� Key features of EU-SILC:
# Integration into the National Statisticalsystem
# Timeliness# Quality: accuracy, consistency,
comparability Production of quality reports and methodological studies
# Flexibility
Introduction� Harmonised definitions (income �
household)
� Minimum methodological requirements (probability sampling, fieldwork..)
� Methodological recommendations
� Characteristics of the dataAll household and personal data linkable in the cross-sectional component and in the longitudinal component (No need for longitudinal micro-data to be linkable with the cross-sectional micro-data)
Introduction
� Based on a legal basis (EP and Council Regulation complemented by Commission Regulations)
� Some degree of flexibility
# Sampling design# Data compilation (registers/interviews)# Questionnaire
2003: Launching on the basis of gentlemen agreement in 7 countries (B, DK, EL, IRL, L, A, N)
2004: Launching under the Framework Regulation in 13 EU (all old Member States except D, NL and UK plus Estonia) plus ISL and N
2005: Launching under the Framework Regulation in D, NL, UK and in all new Member States except Estonia
Time reference
5
Primary annual component: domains coveredHousehold information
Basicdata
(including degreeof urbanisation
(X,L)
Income (X,L)Total household
income (gross and disposable)
Gross income atcomponent level
HousingDwelling type,
tenure status and housing
conditions (X,L)Amenities (X)
Housing costs (X)
Social exclusionHousing and non-housing
related arrears (X,L)Non-monetary deprivation indicators (X,L)
Physical and social environment (X)
Labourinformation (X)
Child care
Primary annual component: domains covered -Personal information
Demographic data(for persons aged
under 16 (X,L), for persons aged 16+ (X,L)
and for former householdmembers (L))
Income (X,L)(gross personal income,
total andcomponents
Basic data(X,L)
HealthHealth status and
chronic illness and condition (X,L)
Access to health care (X)
Education (X,L)(including highest
ISCED level attained)
LabourBasic information on current activity
and current main job, inclon last main job for unemployed (X,L)
Basic information on activity status during income reference period (X)
Total number of hours worked on current second/third jobs (X)
Detailed labour information (X,L)Activity history (L)
Calendar of activities (L)
� Introduced every four years or less frequently
� One module per year only in the cross-sectional component
� Starting from 2005
� 2005: inter-generational transmission of poverty
� 2006: social participation
� 2007: housing conditions
Modules in EU-SILC
Household sample size - cross-sectional component (minimum effective)
2250 Iceland
3250 CyprusLuxembourg
3750
IrelandLatvia
NorwaySlovenia
4000 FinlandLithuania
4500AustriaPortugalSweden
4750
5000 Netherlands
6500 Spain
7250 FranceItaly
7500 UnitedKingdom
8250
Household sample size for 27 countries: 127.000
Germany
BelgiumGreece
Czech RepHungary
4250DenmarkSlovakia
3500 Estonia
3000 Malta
6000 Poland
Household sample size - cross-sectionalcomponent (minimum effective)
3750 Iceland
6500 Luxembourg
7250 Denmark
7500 CyprusSweden
8750Belgium
NetherlandsAustria
9000
10000 Czech RepGreece
10500 Portugal
13500 France
13750 UnitedKingdom
14500
Sample size of persons aged 16+ for 27 countries: 260.150
Germany
LithuaniaSlovenia
8000 Ireland
6750 Finland
62500 Norway
10250 Hungary7000 Malta
7750
Latvia7650
Estonia
11000 Slovakia
15000 Poland1550016000
ItalySpain
Implementation at national levelSurvey design
Sample of persons:
� 4-year rotating panel (DK, ISL, SL)
� 8-year rotating panel (NO)
� Two separate cross-sectional andlongitudinal operations (FI, SE)
Sample of households:� 9-year rotating panel (FR)
� 4-year rotating panel (all the other MS)
SILC: recommended design2004 2005 2006 2007 2008
1
2
3
4
3
4
1’
2’
4
1’
2’
3’
1’
2’
3’
4’
$Selection of four sub-samples or
replications
$From one year to another, three of
the four replications are
kept
2
3
4
1’
Implementation at national levelIntegration in National Statistical System
A complete new operation in all countries with the exception of:
� UK: integration in General HouseholdSurvey
� FI, SE: integration of the cross-sectionalcomponent respectively in IncomeDistribution Survey and in Survey of Living Conditions (ULF)
Implementation at national level
Sampling frame
� National Population Register (BE, DK, IT, LT, AT, FI, SE, IS)
� Census 2000/01 updated (CZ, ES, FR, CY,HU, PL, PT)
� Mikrocensus (DE)� Population register and census (LV)� National social security file (LU)� Sub-sample of LFS (NL)� Register of dwellings from water
corporation (MT)
Implementation at national level
Data sources
� Interviews and registers (DK, NL, SL, FI, SE, IS, NO)
� Interviews only (all other countries)
Implementation at national levelPeriods of data collection
� First semester of the year in the majority ofcountries (CZ, DK, DE, EE, EL, ES, FR, CY,LV, LT, HU, AT, PL, PT, SL, SK, FI, IS, NO)
� Second semester (BE, IT)� Around the middle of the year (LU, MT,
NL)� Continuous survey over the year (IE, SE,
UK)
Cross-sectional component:� End of February (N+2):
Community micro-data files collected in year N available for scientific purposes
� December (N+1):Social cohesion structural indicators based on cross-sectional sample of year N included in Spring report of year (N+2)
Data availability
� Longitudinal component:Community micro-data files for data collected up to year N available for scientific purposes by the end of July N+2
� End of July 2007 ⇒ 2004 and 2005 (countries starting in 2004)
� End of July 2008 ⇒ 2004-2006 or 2005-2006� End of July 2009 ⇒ 2004-2007 or 2005-2007 � End of July 2010 ⇒ 2005-2008
Data availability
� End of June (N+2): Intermediate comparative quality report relating to common cross-sectional EU indicators of year N produced by Eurostat
� End of June (N+3): Final comparative quality report covering both cross-sectional and longitudinal components of year N produced by Eurostat
� Eurostat will organise methodological studies to estimate the impact on comparability of using national data sources
Evaluation of data quality
� Access to confidential data in Eurostatpremises or release sets of anonymised micro-data for scientific purposes under the conditions laid down in Regulation (EC) N° 831/2002 of 17 May 2002
Access for scientific purposes
Cross-sectional UDB containing the most recent wave of SILC
# March 2006: UDB SILC 2004# March 2007: UDB SILC 2005 # March 2008: UDB SILC 2006 # �..
Three types of release
Longitudinal UDB containing all the trajectories ending with most recent
wave of SILC
# August 2007: UDB SILC 2004/053 rotation groups 2004 and 2005
# August 2008: UDB SILC 2004/05/062 rotation groups 2004, 2005, 20061 rotation group 2005 and 2006
# August 2009: UDB SILC 2004/05/06/071 rotation group 2004, 2005, 2006, 20071 rotation group 2005, 2006, 20071 rotation group 2006, 2007
Three types of release
Longitudinal UDB containing all 4 years trajectories available at the time of release
# August 2009: UDB SILC 2004-20071 rotation group 2004-2007 trajectories
# August 2010: UDB SILC 2004-20081 rotation group 2004-2007 trajectories1 rotation group 2005-2008 trajectories
# �
Anonymised measures for research release
A set of global recoding aiming at reducing the risk of disclosure of the database under appropriate threshold
# To maximize usability for international comparisons
Local suppressions not considered as a large-scale alternative
Anonymisation of data for release to researchers
# Region limited to NUTS 1 code# Month of birth recoded into quarter of birth# Top coding (80 and more for age)# Number of rooms (top coded 6 and more)# Country of birth recoded into 4 modalities
(Local/EU/non EU/world)# First citizenship recoded into 4 modalities
(Local/EU/non EU/world)# Second citizenship dropped# Highest ISCED level attained: regrouping of
coded 5 and 6# NACE recoded into one letter code
Set of global recoding: examples
Conclusion
� All your questions / comments are welcome