EUROSTAT
Directorate D: Economic and Regional Statistics
Unit D-1: Key Indicators for European Policies /
Euroindicators
Working Group
Luxembourg, 3th and 4th December 2007
- WELCOME – ADOPTION OF THE AGENDA
1.1 Welcome
Mr. Wurm (Eurostat), the chair of the Working Group welcomed participantsand introduced new Euroindicators staff.
1.2 Adoption of agenda
The agenda was adopted as structured. However, due to practical considerations the ordering of the points was re-arranged. The minutes are structured as detailed in the original agenda.
- WORK PLAN FOR EUROINDICATORS AND EUROTREND
2.1Main achievements of 2007 and future activities for 2008
Gian Luigi Mazzi (Eurostat) presented the main achievements of 2007 concerning PEEIs/Euroindicators special topic, outlined in Doc 196/07.
2.2 Update on governance structures around PEEIs (PEEIs Steering Group activities)
Nikolaus Wurm (Eurostat) presented the Report on PEEIs Steering Group activities, outlined in Doc 197/07.
- EUROINDICATORS/ PEEIs SPECIAL TOPIC
3.1The new PEEIs webpage
Rosa Ruggeri Cannata (Eurostat) presented the new PEEIs webpage, released recently by Eurostat, outlined in Doc 198/07.
The ECB, the OECD and MemberStates (UK, ES, IE, DE, NL) welcomed and expressed a great appreciation for the release of the new selected PEEIs webpage as a very good tool to access data, metadata and news releases dates stored in one place. The ECB and IE (CSO)suggested extending the list of selected PEEIs for the missing indicators from the original PEEIs list. Eurostat stated that all PEEIs will be integrated as soon as they are available. The OECD informed that a similar project is under construction. Moreover they are cooperating with IMF to provide information for some PEEIs for the OECD and other countries. TheOECD asked if other countries apart Japan, USA and Canada will be included in PEEIs list of countries.The UK(ONS) asked about the public access to the webpage andabout direct access to the complete series from the database. Eurostat stated that the access to the complete series is possible only through the Euroind database but direct access to the data tree will be reconsidered in the future. IE expressed approval for the TGM tables. NL (Statistics Netherlands)asked about the presence of the business cycle clock in the PEEIs webpage. Eurostat stated that the link will be added when the tool will be officiallyreleased.
3.2 Improvements in the Euroindicators/PEEIs predefined tables
Agnieszka Politewicz (Eurostat) presented the changes concerning Euroindicators/PEEIs predefined tables, outlined in Doc 199/07.
3.3 Business Cycle Clock: a new dissemination tool
Werkhoven Thomas (Eurostat) presented a Business Cycle Clock, outlined in Doc 200/07.
DE (Bundesbank)noted that not all the cycles are typical as in the case of their economics. Eurostat confirmed that the German subcycle was taken into consideration andis supported in the graph.
PT (INE-Portugal)suggested introducing the information on the trend and on the weight of cyclical components. PTalso raised the problem of precision due to the use of time series long just enough to cover two cycles. PTrecommended to enable user to export the data to excel files. Eurostat stated that the matter of information about trend was already ongoing issue. As for the precision of the cyclical component, a starting year 1990 should be enough. Dissemination of the cyclical data is not yet foreseen as the aim of the tool is to support users in understanding the economic processes just from the graphical point of view.
3.4Methodological pages
Gian Luigi Mazzi (Eurostat) presented improvementson PEEIs/Euroindicators special topic methodological pages, as outlined in Doc 201/07.
- EUROIND DATABASE
4.1 Annual assessment
Juystyna Gniadzik (Eurostat) presented an annual quality assessment of the Euroind database and short term indicators, outlined in Doc 202/07.
PT(INE-Portugal) asked if the database covers the different versions of data in terms of vintages and suggested that the type of revision could be also a criteria to asses the quality of the data.Eurostat explained that Euroind is adatabase where you can check the information available for the given moment. Beside Euroind, Eurostat wants to have soon a real time data base containing all the vintages which would be used to perform an accurate revision analysis. The real time database will be reconstructed on the bases of daily backups of Euroind. Eurostat will reflect on the problem of simplifying the Labour Cost and External Tradedomains.
IT (ISTAT) asked about the methods for checking for the presence and the interpretation of strange values. The UK(ONS) noted that strange values can make analysis difficult and asked about the existence of systematic cause for strange values and the possibility to establish the reason for them. Unit D1 stated that in the past the practice was to disseminate to all production units the results of the monthly analysis in order to check the plausibility of the results for the presence of strange values. Unit D1intends to improve the cooperation between, Production Units and MemberStates in order to receive detailed feedback from the MS, which would allow correcting the automatic algorithms by taking into consideration specificity of the data through domains. The use of TERROR (application of TRAMO software) to detect anomalies in database is in the test phase and the implementation is foreseen for the future to replace raw algorithms.
4.2 Recent and future changes
Rosa Ruggeri Cannata (Eurostat) presented the new PEEIs webpage, released recently by Eurostat, outlined in Doc 203/07.
FR (INSEE) suggested adding Conventional Earnings to the Euroind database.Eurostat explained that the collection of Conventional Earnings was stopped when the use of Labour Cost Index as a harmonised measure was agreed. Unit D1 will ask to include this indicator into the database. FR (INSEE)asked about the reason whythe series in the Euroind database that started in 80', now are starting in 90'. Eurostat explained that the series were shortened due to methodological changes. The back recalculation is a very useful instrument that will give users the possibility of having sufficiently long time series to perform analysis. Eurostat proposed to perform the comparison not only with the IFS but also with the BIS database. The ECB was asked to help in getting the authorisation to use the BIS database. The use of BIS database to perform the exercise of back calculations would allow toimprove the coverage and the structure of back-calculated data. TheECB agreed tocheck the possibility of cooperation between the BIS and Eurostat.
4.3 Improving Euroind metadata
Juystyna Gniadzik (Eurostat) presented the improvement of the Euroindicators metadata,outlined in Doc 204/07.
DE asked (DESTATIS) if the implementation of the Code of Practice is planned for PEEIs page. Eurostat explained that the Quality profiles which was a metadata summarywould be the first step towards theCode of Practices. The UK (ONS) suggested that, when possible,existing country profilesshould be linked with the Eurostat profiles in order to provide more in-depth information.
4.4 The new Eurostat historical database
François Libeau (Hendyplan) presentedthe new Eurostat historical database, outlined in Doc 205/07.
- PUBLICATIONS
5.1 Improvement of the Eurostatistics publication
Agathe Rosin (Eurostat) presented theimprovements of the Eurostatistics publication, outlined in Doc 206/07.
5.2 The new PEEIs Statistics in Focus
Gian Luigi Mazzi and Agnieszka Politewicz (Eurostat) presented the concept of the new publication on PEEIs, outlined in Doc 207/07. Eurostat will inform users on the basis of 22 PEEIs about the situation of different relevant non-EU economic partners. Eurostat requested big effort from the OECD to identify the most appropriate proxies for the PEEIs especially in the non-OECD countries and non-European countries.
NL(De Nederlandsche Bank) asked if Eurostat would include EFTA countries into the publication. Eurostat stated that due to the standard format of the publication, there were a limited number of pages of the publication. Therefore it was not yet decided which countries will be included into the publication. TheECB expressed appreciation for the idea of comparing EU countries with Japan and USA and suggested to review metadata and sources. Eurostat stated that the publication would be completed with very synthetic metadata informing users of main differences and similarities. OECD explained that USA and Japan didn't have the Economic Sentiment Indicator, nevertheless it should be possible to find a proxy. OECD informed that the Bureau of Labour Statistics is developing a version of CPI, which will replicate the methodology of the HICP.
5.3 Other online publications
Agnieszka Politewicz (Eurostat) presentedthe Selected readings online publication, outlined in Doc 208/07.
DE (Bundesbank) commented that Eurostat should not undertake the problems concerning financial markets or business cycle analysis. Eurostat explained that the idea is only to supply a reading guide which is classified by subtopic or author.OECD offered to inform Eurostat about methods on how effectively insert the publication in the Google engine.
- SEASONAL ADJUSTMENT
6.1 Presentation of Joint Eurostat/ECB activities in the field of seasonal adjustment
Gian Luigi Mazzi and Rosa Ruggeri Cannata (Eurostat)presented the Joint Eurostat/ECB activities in the field of seasonal adjustment, outlined in Doc 209/07.
HU(Hungarian Central Statistical Office) informed about the “Seasonal adjustment methods and practices” manual available on their website.
6.2 ESS guidelines on seasonal adjustment
Gian Luigi Mazzi and Rainer Muthmann(Eurostat)presented theESS guidelines on seasonal adjustment, outlined in Doc 210/07.
The guidelines were widely recognized by the Working Group as a valuable documentand participants asked whenthe final version would be available and how it will be disseminated.
FR (INSEE)recommended to the participants of the WG not to use the current version of the guidelines as an official document as it is was not yet officially endorsed. ECB asked about the predicted date of implementation of the guidelines. Eurostat stated thatbefore the implementation, the guidelineswould be presented to the CMFB and the SPC. After the official endorsement it will be also agreed with the OECD how to inform non European countries about the official version of the guidelines.
6.3 Tools for seasonal adjustments
Cristina Calizzani (Eurostat)presented a new version of Demetra 2.1 and the strategy for for seasonal adjustmenttools development, outlined in Doc 211/07.
FR, PL and DE expressed their appreciation for the new version of Demetra. PL(NBP) asked about the statistical background and programming approach to the new software. Eurostat stated that the new tool will be based on the object oriented technology but it is not yet decided wheatear to use X-13 or X-12.
VII. REVISION ANALYSIS
7.1 Joint Eurostat/OECD taskforce on Performing Revisions Analysis for Sub-Annual Economic Statistics
Richard McKenziee (OECD) presented the idea behind the Joint Eurostat/OECD taskforce on Performing Revisions Analysis for Sub-Annual Economic Statistics, outlined in Doc 212/07.
PT (INE-Portugal)asked about the statistical importance of revisions in the project of the taskforce. The OECD informed that the tools that would be provided to perform revision analysis wouldalso provide measureon statistical significance of revisions.
DE (Bundesbank) noted thatmetadata are needed for the revisions analysis in order to understand current situation and to provide the information containing the reasons for revision. The OECD stated that the guidelines will cover data and metadata required to build areal-time database for performing revisions analysis. Metadata must be linked to the vintages where the changes in methodology, scope or to the statistics process were made.
7.2 Impact on revisions from timeliness improvements in Swedish Retail Trade
Richard McKenziee (OECD) presented the impact on revisions from timeliness improvements in Swedish Retail Trade, outlined in Doc 213/07.
IT (ISTAT) asked about the impact on the statisticalunit when applying new legislation. The OECD stated that it was an interesting project which increased timeliness of Swedish data and it could be an example to follow by other countries.
7.3 Nomenclature for classifying revisions to short-term statistics (ONS, UK)
Robin Youll(ONS, UK) presented the nomenclature for classifying revisions to short-term statistics, outlined in Doc 214/07.
IE (CSO) asked about the usefulness for the Member States of the tool for revisions. Moreover IEinformed about their revision practice, which is quite similar to the ONS one and stated that there was no need to reduce the content of the proposed nomenclature.The UK (ONS)would provide both a tool and advice on how toestablisha database on which the tool could operate.
DE (Bundesbank) asked if the limit between the revisions and new data was considered specifying the interest in additive models problems. The UK (ONS)informed about the work done on multiplicative models for seasonal adjustment. It was found that the cross terms were quite small and that the reasonable additive analysis can be performed by just looking at the tree main components. From the user perspective revision is the difference between two series: one that were produced till the given momentand one that are in use from now. Eurostat stated that additive seasonal adjustment works only in the case of additive decomposition as only in this case the constraint on orthogonality between components holds. Major revision creates a breakin the time series of revisions. We could assume the appearance of new time series, so the vintages starts after the break and we start to build a revision analysis after the break on the basis of new vintages. Else we try to remove the break using time series techniques and analysing together the period before the break and after the break. In the last case we need to make a hypothesis that the process before the break is the same as the process after the change.
7.4 Elements for a revisions policy of infra-annual statistics
Gian Luigi Mazzi (Eurostat)presented the principles on which the PEEIs revisions policy should be based outlined in Doc 215/07.
The presented principles for the Eurostat revision policy received great support from the WG participants. Nevertheless there were some comments how to improve these principles. DE (Bundesbank) noted that frequent revisions could create a problem of credibility for statistical institutions. DE asked if there is a need to update the European aggregates each time there is a revision at national level. Eurostat stated that according to principle no.7, there should not be a revisions between two consecutive releases, with the exception of the revisions changing radically the economic policy of the countryor due to errors connections.
ECB asked about the follow up of the principles. Eurostat stated that revisions policy has to be discussed internally. Only after that, it will be decided how to enlarge it from the Eurostat context to the ESS one.
VIII. EUROTREND ACTIVITIES
8.1 Activity report of the Flash consortium
Dominique Ladiray (FR/INSEE)presented theactivity report of the Flash consortium, outlined in Doc 216/07.
ES asked if the GDP flash estimate(t+30) will be published in the Euroindicators webpage and if the national flash, if produced, would be used in the production of Eurostat flash estimate. FR (INSEE)suggested that if the national flash estimate existed, it should be introduced to the database. Eurostat confirmed this position.
PL (NBP) informed that the use of factor methods for short time series of post transformation countries can produce misleading results. FR (INSEE) replied that probably the short time series will not be used to perform flash estimation due to the resulting lack of stability of statistical methods. DE (Bundesbank) expressed concerns on the relation between flash estimates and official statistics. Eurostat stated that its task was to produce official statistics. Consequently the flash estimates have to be based on real statistical observations and must not only use purely mathematical models. It was further mentioned that the purpose of the current grant exercise of the consortium led by France was mainly to investigate the feasibility of earlier flash estimates and to look for possible methods and available indicators.
8.2 Eurostat paper on back-calculation policy
Rosa Ruggeri Cannata (Eurostat), presented Eurostat policy on back-calculation, outlined in Doc 217/07.
IE (CSO) asked which long time series were not available in the database and would be covered by the project of back calculation. IE was also interested in proxies forvariables which would be used for calculating GDP.Eurostat stated that no series are available back from 1970 due to changes in methodology, in definitions or in the base year.Generally the most appropriate proxy for a variableshould be the previous version of the same indicator. In case of GDP, the previous version of GDP available under ESA 79 was used to perform this exercise.