SME Statistics OECD Workshop SME data and methodologies

SME Statistics OECD Workshop SME data and methodologies in the EU - item 5 Paul Feuvrier / Eurostat 17 September 2003 1

SME Statistics Workshop OECD l l l 1 The European legal framework for SME statistics 2 Data availability 3 Methodological issues 4 Statistical Disclosure Control (statistical confidentiality) 5 Quality indicators 17 September 2003

SME Statistics Workshop OECD l 1 The European legal framework for SME statistics – the SBS-Regulation in short, data broken down by NACE Rev. 1 4 -digit level • by NACE Rev. 1 3 -digit level and size class (SME statistics) • by NACE Rev. 1 2 or 3 -digit level and Region (NUTS) • other data sets – environmental protection expenditure – purchase of energy products – etc. . . • 17 September 2003

SME Statistics Workshop OECD l SME data = part of the SBS Regulation Module on definitions – Business Register Regulation – Statistical Units Regulation – 17 September 2003

SME Statistics Workshop OECD l Four specific modules Services (common module) - 5 variables – Manufacturing - 10 variables – Trade - 5 variables – Construction - 10 variables – l For each module – specific size class breakdown • more detailed breakdown for small businesses size bands 17 September 2003

SME Statistics Workshop OECD l Size bands simplified – from reference year 2002 onwards Commission Regulation No 1614/2002 • mainly to prevent too many confidentiality problems • 17 September 2003

SME Statistics Workshop OECD l 2 Data availability – – New. Cronos = Eurostat dissemination database Reference and updated database • – theme 4/sbs/sizclass – Annual enterprise statistics broken down by size classes Historical database • • • theme 4/hist_theme 4/SME “old” joint OECD/Eurostat data collection no longer updated, as a New. Cronos warning clearly puts it 17 September 2003

SME Statistics Workshop OECD 17 September 2003

SME Statistics Workshop OECD EU-15 totals data by size class available for reference year 1999 and 2000 – Germany < 20 not available beforehand… • Acceding countries totals ACC (10 countries) available in New. Cronos end of September 2003 • EU-25 available in New. Cronos on 1/5/2004 • 17 September 2003

SME Statistics Workshop OECD l 3 Methodological issues – – Eurostat performs quality checks before any data release Implementation problems encountered at national level • Level of detail of size class data – Difficult to implement for some small countries • One size hardly fits all… – Still a single size class breakdown is necessary (and relevant) for computation of reliable EU-15 totals 17 September 2003

SME Statistics Workshop OECD – Completeness The existence of cut of values used to be the most serious problems related to SBS – Data related to enterprises with more than x persons employed used to be available • instead of > 1 person employed requested by the Regulation • Depending on the availability of a Statistical Business Register including small businesses – and on its quality… • Removal of cut of values = main achievement of SBSRegulation – Germany, among others, made dramatic efforts top comply with the Regulation and cover <20. . . • 17 September 2003

SME Statistics Workshop OECD – Remaining problematic countries as regards cut of value • • • Greece – >10 persons employed in all industries Ireland – >20 persons employed in Construction Hungary – >5 persons employed in all industries That’s it for reference year 2001! – Other countries >1 person employed in all industries More problems for historical data. . . 17 September 2003

SME Statistics Workshop OECD – Implementation problems encountered at national level • small and medium-sized enterprises are defined as enterprises which have: – fewer than 250 employees, – an annual turnover not exceeding € 50 million – an annual balance-sheet total not exceeding € 43 million. – Which are independent • in particular not affiliate of any enterprise group. 17 September 2003

SME Statistics Workshop OECD l 4 Statistical Disclosure Control (statistical confidentiality) Member States send confidential data to Eurostat. . . – …so that Eurostat is in a position to calculate European totals. . . – … and to release them as soon as the protection is not broken – 17 September 2003

SME Statistics Workshop OECD – Cells of detailed tables contain information either on a single or very few respondents • – --> implementation of measures to limit disclosure risk Main method implemented by Eurostat so far = Cell suppression Assessment of disclosure risk connected to each step • Suppression of sensitive cells – primary suppression • Secondary suppression – primary suppression not enough to prevent disclosure because of the additive relationship between the cells • 17 September 2003

SME Statistics Workshop OECD 17 September 2003

SME Statistics Workshop OECD l 5 Quality indicators 17 September 2003

SME Statistics Workshop OECD – Variance takes into account survey design • non response • misclassification • – To that respect, CVs clearly higher for small businesses than for large ones, Accuracy of SME data not as good as the one of large businesses statistics • Mainly because small businesses are not completely enumerated – take some stratum • Large businesses are – take all stratum • 17 September 2003

SME Statistics Workshop OECD • Other problems posed for large businesses – industrial heterogeneity • secondary activities – statistical coverage of enterprise groups 17 September 2003

SME Statistics Workshop OECD l Conclusions – Eurostat database on SMEs of high quality • • • timeliness – data available at T+20 accuracy – CVs available availability completeness – coverage of small businesses almost fully achieved comparability – legal framework for concepts and definitions 17 September 2003

SME Statistics Workshop OECD – Yet some challenges ahead for Eurostat work on SMEs: Still some cut off values to remove. . . • Reflection on a distinction between – independent SMEs – SMEs affiliates of groups • Improving the quality of R&D variables for SMEs – where the relevant statistical unit is probably more the enterprise group than the enterprise • Improving the consistency between SBS - SME data and SBS - Business Demography data • 17 September 2003