Introduction to EU-LFS and EU-SILC - UK Data Service · Minimum required sample size Achieved...

August 2011 European workshop to introduce EU-SILC and EU-LFS data

Introduction to EU-LFS and EU-SILC

Frank Espelage – Eurostat F2 (EU-LFS) [email protected]

Fabienne Montaigne – Eurostat Laboratory (EU-SILC) [email protected]

mailto:[email protected]

mailto:[email protected]

August 2011 Introduction to EU-LFS and EU-SILC 2

Introduction to EU-LFS and EU-SILC

1. Political background

2. Organisation of EU-LFS and EU-SILC

3. Microdata access: legal background and technical outlook

4. EU-LFS: design, processing and dissemination aspects

5. EU-SILC: design, processing and dissemination aspects


1. Political background

Europe 2020 is the EU's growth strategy for the coming decade

Goal: “Smart, sustainable and inclusive growth”

5 headline targets agreed for the whole EU by the Council in June 2010:

1) 75 % of the population aged 20-64 should be employed2) 3% of the EU’s GDP should be invested in R&D3) The “20/20/20” climate/energy targets should be met4) The share of early school leavers should be under 10% and at least 40%

of the younger generation should have a tertiary degree5) At least 20 million people should be lifted out of the risk of poverty and

exclusion

Progress monitored through the European Semester, and particularly throughthe Joint Assessment Framework (JAF) which includes a list of indicatorsbased on EU-LFS and EU-SILC



Regulatory framework at European level:

Clear specifications of deliverables

Obligation of Member States to deliver data

Harmonisation but

Changes require consensus negotiations

Procedures take time

Implementation:

Surveys implemented nationally

Microdata collected and processed nationally

Microdata transmitted to and validated by Eurostat

Indicators computed and disseminated by Eurostat



The National Statistical Institutes are responsible for the survey at national level:– Flexibility in the implementation (design, data sources,

questionnaires …)– Surveys can be supplemented to serve national purposes

Ex-ante output harmonization is reached by:– Definition of target concepts/variables to be measured– Minimum requirements (survey design, sample representativeness

/ sample size, fieldwork, imputation, timeliness, calculation of weighting factors, quality reporting …)

– Guidelines and explanatory notes (updated regularly)– Standardized output (individual data)– Monitoring by Eurostat



Council Regulation on Community Statistics (like EU-LFS, EU-SILC):

general guidelines for receiving, processing and disseminating confidential data obtained for Community statistics, including also the access for scientific purposes

access to microdata via Eurostat may be granted by National Statistical Institutes

the Commission should be assisted by the Working Group on Statistical Confidentiality in defining the conditions governing access for scientific purposes to confidential data

Implementation Regulation: distribution of DVDs as one option for accessto confidential data for scientific purposes (anonymised microdata obtainedby modifying confidential personal records to minimize the disclosure risk).Both EU-LFS and EU-SILC use this distribution channel.



EU-LFS and EU-SILC anonymised microdata are available free of charge.

Details on conditions of microdata access (eligible institutions etc.) canbe found on the Eurostat website: http://epp.eurostat.ec.europa.eu

Recent discussions between Eurostat and Member States on the wayforward focus on decentralised access also for EU microdata:

Network of safe centres run by National Statistical Institutes

Analysis of synergies with data archives (CESSDA)

Remote access as part of the medium term strategy

Creation of public use files (anonymised records, neither direct norindirect identification) also under discussion

http://epp.eurostat.ec.europa.eu/


4. EU-LFS: design

PeriodicityQuarterly and annual results

Participating countriesEU27, IS, NO, CH, HR, MK, TRFirst data in database for 1983From 2000 full coverage of EU27

Reference populationPopulation in private householdsPeople in collective households and institutions may also be covered, but are not included in all countries

Sample sizeMinimum precision criteria for key indicator (unemployment) + further requirements. Sampling rates vary between 0.2% and 3.3%About 1.5 million people per quarter


4. EU-LFS: design

Topics covered in the core

• Demographic background and household composition• Labour status, employment characteristics of the main job,

atypical work, hours worked, second job, previous work experience, search for employment, methods used to find work

• Self-declared main labour status (optional)• Education and training• Situation one year before• Income from the main job (deciles)• Technical items related to the interview

EU-LFS is the main source for labour market analyses inEurope


4. EU-LFS: design

Topics covered in the annual modules (since 1999)

• Accidents and work and work-related health problems: 1999, 2007, 2013

• Entry of young people into the labour market: 2000, 2009• Work organisation and working-time arrangements: 2001, 2004,

2015• Employment of disabled people: 2002, 2011• Lifelong learning: 2003• Reconciliation between work and family life: 2005, 2010• Transition from work to retirement: 2006, 2012 • Labour market situation of migrants and their immediate

descendants: 2008, 2014


4. EU-LFS: design

Structure of the data files

- Quarterly data allowing derivation of ILO labour status and main results on education, occupation, economic activity, hours worked, full-time/part-time, temporary employment etc.

– Yearly data extending the quarterly data by structural variables, for instance reasons (for FTPT, job search etc), income, atypical work, situation one year before survey, details on education etc. Yearly sample can be subsample of quarterly samples

– Ad hoc modules should be surveyed for respondents in the yearly sample to allow exploiting the full wealth of information


4. EU-LFS: design

EU-LFS is a continuous sample survey with evenly spreadreference weeks, but offers some flexibility regarding the useof registers for certain variables (mainly used by Nordiccountries or for income information)

National flexibility in the implementation:• Sampling design• Mode of data collection (PAPI, CAPI, CATI, also CAWI in the

pipeline)• Use of imputation• Etc.

Countries use different rotation patterns, but the EU-LFS isnot a panel !


4. EU-LFS: processing aspects

Data transmission deadlines according to regulation:

Quarterly data: 12 weeks after the reference quarter

Yearly data for year N (in particular yearly weights in case of a subsample for yearly variables): together with quarter 4 data in March N+1

Ad hoc module data for year N: end of March N+1


4. EU-LFS: processing aspects

Data validation:

Syntax: valid codes, correct routing between variables

Logic: plausibility of relationship between members of the same household; coherence between variables

Currently no comparison with previous results at individual record level

Comparison with previous quarterly/yearly results for some 50 calculated main figures, e.g. employment, unemployment, hours worked, …

NSIs are asked for explanations of detected irregularities/ correction of errors/confirmation of main results


4. EU-LFS: dissemination aspects

Availability of results using

Quarterly data main indicators available on the Eurostat website on average 4 weeks after transmission deadline

Yearly data annual averages: usually at the same time as quarter 4 results except for yearly variables: mid N+1 for year N

Ad hoc module data separate processing; requires in general yearly files for countries with yearly subsamples; most data for year N confirmed mid N+1; second half N+1 special evaluation of each module; first results on the website: not before N+2


4. EU-LFS: dissemination aspects

Anonymisation of data for year N

3rd quarter N+1 for quarterly and yearly data

year N+2 for ad hoc module data

Consequences:

Release of year N+1 contains core data until N and ad hoc module data until year N-1.

Latest available data anonymised and disseminated, including all revisions for previous years Eurostat received since the last release

Country coverage: EU27, IS, NO, CH (from 2011 release)


4. EU-LFS: anonymisation

General aggregation and anonymisation criteria

Agreed between Eurostat and the NSIs of the Member States and endorsed by theWorking Group on Statistical Confidentiality, the most important ones being:

– Age (5-year groups) – Nationality / Country of Birth (up to 15 categories since 2009 release)– NACE (1-digit)– ISCO (3-digit)– No tracking of persons across datasets / time possible

Furthermore aggregation of, for instance, marital status, years of residence greater than10, professional status, number of persons working in the local unit, duration of searchfor employment, hours worked > 80, plus further country specific aggregations, e.g. forgeographical information


5. EU-SILC: design

Participating countries (by first year of implementation)2003: BE, DK, EL, IE, LU, AT, NO 2004: EU-12 + EE, NO, IS 2005: EU-25 + NO, IS, TR 2007: EU-27 + NO, IS, TR, CH + tested in Croatia, FYROM, Serbia

Reference population= All private households and their current members Excluded: people in collective households and institutions

Sample sizeMinimum precision criteria for key indicator, to produce results both at country and EU level (precision of at-risk-of-poverty rate of 1%)


0

2000

4000

6000

8000

10000

12000

14000

16000

18000

20000

22000

Ger

man

y

Uni

ted

King

dom

Fran

ce

Italy

Spai

n

Pola

nd

The

Net

herla

nds

Swed

en

Belg

ium

Cze

ch R

epub

lic

Hun

gary

Gre

ece

Portu

gal

Aust

ria

Den

mar

k

Finl

and

Nor

way

Slov

akia

Lith

uani

a

Irela

nd

Latv

ia

Slov

enia

Esto

nia

Cyp

rus

Luxe

mbo

urg

Icel

and

Mal

ta

# H

ouse

hold

s

0.0%

0.2%

0.4%

0.6%

0.8%

1.0%

1.2%

1.4%

1.6%

1.8%

2.0%

2.2%

2.4%

2.6%

2.8%

3.0%

3.2%

3.4%

Minimum required sample size Achieved sample size Achieved sample size compared to total population


5. EU-SILC: design

Topics covered in the core• Information on sampling design, household composition• Basic characteristics of household members• Income, education, labour information, health• Housing conditions and material deprivation

Topics covered in the annual modules2005: Intergenerational transmission of poverty;2006: Social participation; 2007: Housing conditions;2008: Over-indebtedness and financial exclusion;2009: Material deprivation;2010: Intra-household sharing of resources;2011: Intergenerational transmission of disadvantages;2012: Housing conditions; 2013: Well-being


w1 w2 w3 w4

w1 w2 w3 w4

w1 w2 w3 w4

w1 w2 w3 w4

w1 w2 w3

w1 w2

w1

2003 2004 2005 2006 2007 2008 2009 2010 2011

1

2

3

4

1'

2'

3'

4-year rotational panel(except FR 9 years, NO 8 years, LU panel)


5. EU-SILC: design

Structure of the data files:One year’s data consists of four files :– Household register (D)– Household data (H)– Personal register (R)– Personal data (P) only for those aged 16+These files can be mergedLongitudinal and cross-sectional files cannot be linked

National flexibility in the implementation:• Source = survey and/or registers (DK, NL, SI, FI, SE, IS, NO)• Sampling design• Mode of data collection (CAPI, PAPI, etc.)• Fieldwork: one-shot (≠ length) or continuous survey• Income reference period: fixed or moving 12-month• Imputation method


5. EU-SILC: processing aspects

Data transmission deadlines according to regulation

• Cross-sectional files: Extreme deadline for transmission of micro-data to Eurostat is 30 November (N+1) for Member States where data are collected at the end of year N or through a continuous survey or through registers, and 1 October (N+1) for other Member States

• Annual ad-hoc modules transmitted with and included in the cross-sectional files

• Longitudinal files: Transmission of files to Eurostat within 15 months of the end of data collection. Deadline is 31 March of year N+2 (for all countries).


5. EU-SILC: processing aspects

Data validation

• Syntax and logic checked both by NSIs and Eurostat• Comparison of key variables and indicators with previous

years’ values plausibility checked and if necessary explanation/confirmation asked for

• NSIs asked for explanations of detected irregularities and correction of errors where this is possible

• Indicators calculated by Eurostat and bilaterally validated before publication

• Cross-sectional indicators updated every month and published at the latest by 15 December of N+1


5. EU-SILC: dissemination aspects

Issuing of microdata/release policy• Cross-sectional files: have to be issued by end February

of year N+2• Longitudinal files: have to be issued by end July of year

N+2. Include the four most recent years (if applicable)• Revisions for cross-sectional and longitudinal files:

issued with the micro-data of the new operation, for both the cross-sectional and the longitudinal components

• Extended documentation now included on the CDs• Country coverage: EU27+IS,NO (from 2011 release),

CH to be included from 2012 release


5. EU-SILC: anonymisation

Data for the CD is anonymised as follows• Region: changed from NUTS2 to NUTS1 • Survey design: Most variables related to survey design withdrawn or

randomised• Time of interview: day of household/personal interview withdrawn,

month of household/personal interview recoded into quarter• Age: month of birth variables recoded into quarter, other age

variables top-coded to 80+• Housing: dwelling type=other type of accommodation recoded as

missing, number of rooms top-coded to 6+• Origin: country of birth and citizenship 1 regrouped in four categories:

survey country, EU, non-EU, world; citizenship 2 withdrawn• NACE classification regrouped• Additionally there are some country-specific provisions


EU-LFS and EU-SILC: further information

General background information can be found in the EU-LFSand EU-SILC dedicated sections on the Eurostat website:

http://epp.eurostat.ec.europa.eu/portal/page/portal/ employment_unemployment_lfs/introduction

http://epp.eurostat.ec.europa.eu/portal/page/portal/ income_social_inclusion_living_conditions/introduction

http://epp.eurostat.ec.europa.eu/portal/page/portal/employment_unemployment_lfs/introduction

http://epp.eurostat.ec.europa.eu/portal/page/portal/employment_unemployment_lfs/introduction

http://epp.eurostat.ec.europa.eu/portal/page/portal/income_social_inclusion_living_conditions/introduction

http://epp.eurostat.ec.europa.eu/portal/page/portal/income_social_inclusion_living_conditions/introduction


Thank you for your attention

Introduction to EU-LFS and EU-SILC - UK Data Service · Minimum required sample size Achieved...

Documents

Transcript of Introduction to EU-LFS and EU-SILC - UK Data Service · Minimum required sample size Achieved...