Agenda Item 2

Agenda Item 4.2

• Agenda Item 4.2

• JVS seasonal adjustment

• [e-mail]

Transmission of SA data

• Transmission of SA data

• SA quality reports sent by the NSIs

• Eurostat's validation of the SA data

• SA data for the European aggregates

SA data required for individual/aggregated NACE sections B-E, F, G-I, J, K, L, M-N, O-Q, R-S

• SA data required for individual/aggregated NACE sections B-E, F, G-I, J, K, L, M-N, O-Q, R-S

• All Member States except Finland now provide SA data for both job vacancies and occupied posts

• Finland will be able to provide SA data from June 2017 due to a methodological break in time series

• No SA data transmitted by Norway

LAMAS agreed in January 2016 on filling the template for SA quality report and transmitting it to Eurostat

• LAMAS agreed in January 2016 on filling the template for SA quality report and transmitting it to Eurostat

• Eurostat developed a plug-in for JDemetra+ version 2.1 producing the information required

• The SA quality report was sent by all Member States except Greece, France and Luxembourg

• No report from Switzerland

JDemetra+ officially recommended, since 2 February 2015, to the members of the ESS

• JDemetra+ officially recommended, since 2 February 2015, to the members of the ESS

• Eleven countries use JDemetra+ version 2.0 or higher

• A number of countries still use the old versions of the Demetra software

• Austria and the United Kingdom use the X12/X13 software whereas Portugal uses SPSS/TRAMO-SEATS

Varies across the countries from 4 years to 15 years

• Varies across the countries from 4 years to 15 years

• Some countries do not transmit regularly the whole SA time series: CZ, HR, EL, IE, NL, PT and SE

• All countries are reminded to provide full SA time series by each data transmission i.e. not just the last quarter(s)

Covering the times series for job vacancies and occupied posts in the NACE sections: B-E, F, G-I, J, K, L, M-N, O-Q and R-S for enterprises 1+

• Covering the times series for job vacancies and occupied posts in the NACE sections: B-E, F, G-I, J, K, L, M-N, O-Q and R-S for enterprises 1+

• STEP 1: Eurostat checked for each series if it had been seasonally and/or calendar adjusted

• By comparing SA with NSA data
• By consulting the SA quality report

STEP 2: Eurostat processed non-seasonally adjusted time series in JDemetra+ version 2.1

• STEP 2: Eurostat processed non-seasonally adjusted time series in JDemetra+ version 2.1

• TRAMO-SEATS or X12/X13 method used
• Direct approach applied for B-E, G-I, M-N, O-Q and R-S
• Results compared with SA data provided by the countries

• STEP 3: Eurostat checked for residual seasonality in the SA time series provided by the countries

Seasonal adjustment: major discrepancies between the results of the countries and Eurostat seen for BG, HR, HU, PL, PT and RO

• Seasonal adjustment: major discrepancies between the results of the countries and Eurostat seen for BG, HR, HU, PL, PT and RO

• General recommendation not to apply calendar adjustment

• Residual seasonality detected in times series of DK, DE (now corrected by Germany)

Methodology

• Methodology

• Assessment

• Dissemination plans

Three main strategies possible:

• Three main strategies possible:

• - the direct approach

• - the centralised indirect approach

• - the decentralised indirect approach

• The direct approach is the only option that can be implemented in the short run (see next slides)

Seasonal patterns rather not the same across EU Member States which would support using indirect-centralised or decentralised–approach

• Seasonal patterns rather not the same across EU Member States which would support using indirect-centralised or decentralised–approach

• Difficult to justify the centralised indirect-approach as Member States are required by the JVS regulation to provide SA data

• The centralized indirect approach could be considered after some NSIs improve their national seasonal adjustment practice/methodology

• Therefore, the direct approach is the only option that can be implemented in the short run

The decentralised indirect approach gives the highest volatility of SA data

• The decentralised indirect approach gives the highest volatility of SA data

• Lower volatility for both the centralised indirect approach and the direct approach

• The centralised indirect approach not selected (against the subsidiarity principle)

• This supports implementating the direct approach in the short term and re-assessing the decentralised approach in the long term

SA data for the EU aggregates could be published in December 2016

• SA data for the EU aggregates could be published in December 2016

• Disseminated for NACE aggregates: B-S, B-N, B-F and G-N

• In the course of 2017, Eurostat will assess if SA figures could be given more visibility

For the countries that have not transmitted their SA quality reports to do so before the end of this year

• For the countries that have not transmitted their SA quality reports to do so before the end of this year

• To take note that, from 2017 onwards, the SA quality reports should be included in the standard quality report due in August each year

• To ensure that full SA back series are transmitted to Eurostat by each quarterly transmission

For the countries: BG, HR, HU, PL, PT and RO to take note of Eurostat’s findings as regards possible caveats in their seasonally adjusted data

• For the countries: BG, HR, HU, PL, PT and RO to take note of Eurostat’s findings as regards possible caveats in their seasonally adjusted data

• To take note of the general recommendation not to apply calendar adjustment for JVS

• To take note that seasonal adjustment should be done in accordance with ESS guidelines and (preferably) using JDemetra+ version 2.1

• To take note of Eurostat's plans and methods to publish SA data for European aggregates

• Thank you for your attention!