Presentation on theme: "EU KLEMS Growth and Productivity Accounts"— Presentation transcript:
1EU KLEMS Growth and Productivity Accounts Mary O’Mahony (University of Birmingham and National Institute of Economic and Social Research)Presentation at CMFB MeetingLuxembourg, 28 June 2007This project is funded by the European Commission, Research Directorate General as part of the 6th Framework Programme, Priority 8, "Policy Support and Anticipating Scientific and Technological Needs".
2What is in the March 2007 version of the EU KLEMS Database? Systematic data collection based on national accounts and complementary official sources (LFS and other surveys)Long time coverage , with greatest detail for post-1995Harmonized methodologies on industry classification, capital and labour input, deflation and aggregations (e.g. market economy, market services, ICT producing vs. using)Decomposition of capital (ICT and other) and labour input (skills, age, gender)Broad coverage of EU countries:Growth accounts coverage of “old” EU-10 (excl. GR, IR, LU, SE, PT) plus 5 new member states (incl. PL, SK, HU, CZ and SI) & Euro areaLimited coverage of other 5 other “old” EU countries and 5 new member states (CY, MT, LT, LV and EE)Also comparisons with U.S. and Japan
3What the data tell usSector Contributions to Labour Productivity Growth in Market Economy Confirm Existing ViewsThe EU-US growth differential is not in manufacturing and minor in ICT productionbut huge in market servicesCountry contributionsIreland, Greece at top end, Finland, Sweden also highSpain, Italy – very lowSources of growthEmployment rising in EU; average skill levels also rising but at decreasing rate since 1995Slightly smaller contribution from ICT to growth compared to USBut EU falls behind US in multi factor productivity growth, again mainly in market services.Labour Productivity growth shows some signs of increasing in EU in recent years - If changes persist need to understand quickly country/industry location and sources of growth
4Milestones on EU KLEMS Database March 2006 – May 2007 End of March 2006 and November 2006: restricted release of preliminary analytical database, round of feedback from NSIs and othersEarly March: green light from DG ECFIN and DG Research for release at Commission premises in BrusselsMarch release is analytical module of EU KLEMS (uses feedback from NSI’s but is only research database, not official statistics)
5Statistical vs. analytical modules of database Analytical module of the databaseCore of the EU KLEMS databaseUses “best practice” techniques in area of growth accountingFocuses on international consistencyAim is full coverage (country * industry * variable) for revision periodConsider alternative or pioneering assumptions (measurement of skill levels, construction of capital services).
6Statistical vs. analytical modules of database Statistical module of the database:Developed parallel to the analytical moduleData consistent with those published by NSIsMethods according to rules and conventions on national accounts, supply and use tables, commodity flow methods, etc. (SNA 1993, ESA 1995) or at least supported by NSI’sStatistical module meets statistical standards of NSI's and Eurostat and can eventually be incorporated in their present publication programme and in New Cronos.
7Feedback from NSI’s: Many comments focused on differences national accounts versus EU KLEMS (I) Differences include:Revision datesAggregation methods, within country and EU totalsNational accounts complemented with other sources (e.g censuses)Adjustments for FISIM (e.g. UK)EU KLEMS - headcounts (adjustments from jobs for AT,ES, UK)Hours for self-employedleave and unregistered work has been added to adjust NA data to actual hoursPlus general adjustments to ensure plausibilityGenerally complementary, national accounts and growth accounts provide very useful check resultsConceptual differencesBEL Hours for self-employed have been calculated and added to hours of employees.DNK Total employment excl. leave (EUK) vs Total Employment incl. leave (EUROSTAT)CYP Ratio COMP/VA applied to VA for consistency.Total hours for employees (Eurostat) have been compared with (estimated) hours for total employment (EUK). We made the assumption that self-employed work the same hours as employees.CZE Strange employment and hours figures have been adjusted.EST In EUKLEMS two periods delivered by NSI have been linked in There are now complete series in EurostatESP For output we use non-revised supply and use tables. For employment the difference is in heads vs. jobs.FRA Compensation before 1999 has been estimated with F level data. For hours unpublished sources have been combined with NA dataHUN Output and compensation data for has been estimatedEmployment in mining has been redistributed with other data, because the classification did not fit to EU KLEMSIRL Eurostat has used more recent and more detailed data, especially for hoursLTU Employment and hours have been constructed with several sources to overcome the data gapsLVA Output and compensation have been constructed with several sourcesNLD The split up of some industries in employment differs between both sources.POL Compensation in Agriculture differs, due to own calculations. Employment has been linked in 2000, which also leads to differences.PRT Compensation, employment and hours have been linked with older sources in 1999SWE Services industries have been classified differently in EurostatSVN The split up of some mining industries happens with different sources, adjustments have also been made for a few other sectors.Hours have been calculated with some additional calculations.UK FISIM has been distributed in output, employment is in persons
8Future work on Analytical module (until end of project) Update to 2005Incorporate revisions to national accountsExpand data for some countries, e.g. Ireland, Greece, PortugalMain focus on Levels estimatesLabour productivityLevels of capital intensity and skill intensityLink to microdatafirm dynamics – entry/exit, concentration, firm agelabour force data – migration, training
9Development of analytical module: the future of EUKLEMS Analytical module: Central development will remain necessary for time being, probably by EU KLEMS consortium (or sub-group). Requires:Continuous updates (of 31-industry growth accounts and 60-industry labour productivity accounts)Extension/improvements in terms of measurement of levels, non-market industries, intangible assetsDependent on financial support from European Commission and other parties
10Development of statistical modules Development of statistical modules is separate and depends on own resources from NSI's.Feedback from NSIs suggest support for development of KLEMS growth accounts in conjunction with national accounts but in many cases statistical module cannot go beyond official national accounts:As published by NSI or transferred to EurostatWill leave many statistical modules rather empty, and will not go beyond labour productivityLimited or no resources for most NSI’s to work on productivity numbersMost NSI’s (but with some exceptions) cannot promise implementation of KLEMS account in their statistical system
11Reactions from NSI’s (summarized in EU KLEMS Statistical Progress Report, May 2007) Options for statistical module: According to Eurostat/EU KLEMS letter (29 November 2006):Integrate analytical and statistical moduleUse (modified) subset of data from analytical module in statistical moduleNo data from analytical module in statistical module1. Acceptance of analytical module as statistical, but with flaggingCzech Republic, Denmark, Finland, Slovenia2. Modified subset of dataEstonia, Italy3. No Analytical module data except official data (either own country official or data transferred according to ESA)Austria, Belgium, France, Germany, Hungary, Netherlands, Poland, Spain, SwedenNot clear or in progressGreece, Ireland, Portugal, Slovak Republic, UK
12Strategy with regard to statistical module Follow-up contact between consortium members and NSI on details of implementation:Which series precisely to include:1. Non-official but accepted data (grey)2. Not accepted data (should be removed)3. Official data that can stay in (black)4. Own data of the statistical office that should replace EU KLEMS data (blue)Check to include investment, capital stocks and labor composition data in public versionWhat time frame for publication?Who does what? (NSI, consortium member, joint effort)The development of the statistical module is crucial to move EU KLEMS beyond a research databaseEU KLEMS consortium partners are willing to provide advice and helpMany downloadsLittle bit of circle reasoning