EUROSTAT
Directorate F: Social Statistics
Doc. Eurostat/F/13/DSS/02/3.5.EN
Item 3.5.
Regionalisation and improved timeliness of EU-SILC data: state of play
MEETING OF THE EUROPEAN DIRECTORS OF SOCIAL STATISTICS
Luxembourg, 18 and 19 November 2013
BECH building, room AMPERE
1
Item 3.5 Regionalisation and improved timeliness of EU-SILC data: state of play
1) Background information
During the last meeting of the DSS that took place on 4th-5th March 2013, it has been agreed that the ESS has to move towards delivering data on poverty and inequality at the regional level, as well as to improve timeliness of EU-SILC data. These issues were also addressed from a technical point of view during the Living Conditions Working Group in June 2013.
Improving the availability of regional data is an urgent policy need coming from DG Regional Policy (DG REGIO), in the context of streamlining the EU 2020 Strategy into this domain of action through complementing GDP with poverty and inequality data for monitoring purposes (2014-2020) and fund allocation (starting 2020) of the Cohesion Policy. Reliable data (preferably at NUTS 2 level, and in exceptional cases NUTS 1 for larger Member States) is needed starting 2018 for fund allocation from 2020, while at least estimations of regional indicators are necessary from reference year 2014.
Moreover, more timely data are needed for better policy monitoring, especially in times of rapid socio-economic changes. This is a strong policy need, coming especially from DG Employment and Social Affairs (DG EMPL), and a plan with concrete actions towards achieving it has been devised by Eurostat and the ESS.
For meeting these two important policy needs it was agreed to define and implement concrete measures at national level. Moreover, improving both timeliness and the regional dimension may require the use of additional resources. For the latter purpose, DG EMPL and DG REGIO indicated their willingness to support financially these actions together with Eurostat.
In this context, in order both to prepare in advance the actions to be carried-out and to earmark accordingly EU funding, Eurostat asked Member States to provide already by the end of June 2013 a draft National Action Plan and a more detailed plan by end of October mainly. The latter applied mainly for the regional dimension in a limited number of countries with a high number of NUTS 2 and even NUT1 regions but might involve important costs.
Hence, this note reports on the plans and the kind of measures set-up by Member States.
2) Synthesis of the national action plans
2.1 General information
First of all it should be noted that nearly all Member States submitted actions plans, except Ireland and Italy. In addition some countries didn't provide information in October as their action plan didn't need any update. The work planned is impressive, despite difficult staff and budgetary constraints (in some cases unfortunately the plans were reduced or rescheduled between June and October due to such issues).
Actions already started in numerous countries and even already achieved first result in some cases. Of course, not all countries will be able to achieve already in 2014 the targets of the plan on timeliness agreed with DG EMPL but all propose to implement significant improvements as soon as possible. Concerning the regional dimension, also different methods are proposed and need to be further investigated but, provided the results of these studies are positive, they would be also implemented in a relative limited timeframe.
More generally, it should be noted that both regionalisation / timeliness needs and the EU-SILC revision plans (in particular for the extension of the longitudinal component) involve in a lot of countries changes in their production chain, e.g., by collecting data earlier during the year, by revising sampling methods or by using more and better data from administrative sources. In 2 Members States a complete change of survey is envisaged, while in some others data collections methods will be hugely adapted. In 6 Member States the implementation of CAWI will be tested or is already planned for implementation and other countries are moving / will move from PAPI to CAPI or CAPI to CATI.
Eurostat would like to thank the Member States for the efforts already made and the willingness to achieve so important improvements in the availability of enhanced comparable data on living conditions in general and poverty and social exclusion in particular in Europe.
2.2 Regional action plans
Belgium will develop and France (NUTS 2) already developed and achieved good results with a methodology based on small areas estimates. UK will also carry out a study based on small areas estimates at NUTS 2 level.
For its part Germany will develop estimations procedures (NUTS 1) and Portugal will adapt the sample in parallel to the change of the rotational pattern, both over the period 2014-2018.
For the short term Spain and Austria indicated they would prefer to use 3-year moving averages. On the medium term Spain would increase by 50% the sample size. UK will also investigate the use of 3-year moving averages at NUTS 1 level.
Greece will explore the increase of the sample size or the use of administrative sources.
Croatia will better balance the sample among regions while Hungary will review the survey sampling and the Netherlands will develop a new weighting model.
Poland and Romania will investigate different methods.
Finally Denmark would like to develop a national variance method using Jackknife in order to take into account the specific Danish calibration in SILC and assess more accurately the need for improving or not the precision of regional estimates.
The other Member States either are not concerned (no NUTS 1 and 2) or have already accurate estimates from SILC at regional level.
2.3 Timeliness action plans
This domain is those where the highest number of actions is envisaged with in some cases already huge improvements achieved.
Concerning the early delivery of material and economic constraint data and potentially all non-income data by the end of the year of data collection N (or in a first time beginning of N+1):
In addition to the 9 countries that participated in the test in 2013 on 2012 data (for BE however the exercise will be done again only from the 2015 data), 12 other countries plan to do it from 2013, 2014 or 2015 data. This will however be sometime progressively (for the first time delivery in the first quarter N+1 and then in December N / January N+1). Other countries develop or will develop statistical methods that will allow them to participate in future to the early submission.
Concerning the final delivery of cross-sectional data in June N+1 instead of November N+1:
13 countries plan to achieve it progressively in the next 2 or 3 years. Some countries will implement it already from 2014 on 2013 data. Others target the availability / delivery of all cross-sectional data in the first quarter of N+1.
Other activities are planned in some Member States on current monthly income, by testing and modelling. In some cases this will also be used in the context of the above timelier submission of SILC cross-sectional data. Tests will be concentrated in 2014 (or are already on-going).
Finally, work on micro simulation models is also being developed in few countries and the UK proposes to work on and test a module on coping strategies.
3) Call for proposal
As announced earlier, Eurostat plan to launch a call for proposal in early 2014 covering 3 domains: regionalisation, timeliness and tests for the revision of EU-SILC. As said above, in addition to Eurostat, both DG EMPL and DG REGIO will provide funding for this Eurostat's call. In total, provided that budget is made available as expected, around 2.6 million euros will be available for grants (including activities for testing the EU-SILC revision).
1