SME%20Statistics%20OECD%20Workshop

1
SME Statistics OECD Workshop

• SME data and methodologies in the EU - item 5
• Paul Feuvrier / Eurostat

17 September 2003
1
2

• 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
3

• 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
4

• SME data part of the SBS Regulation
• Module on definitions
• Business Register Regulation
• Statistical Units Regulation

17 September 2003
5

• Four specific modules
• Services (common module) - 5 variables
• Manufacturing - 10 variables
• Trade - 5 variables
• Construction - 10 variables
• For each module
• specific size class breakdown
• more detailed breakdown for small businesses size
  bands

17 September 2003
6

• Size bands simplified
• from reference year 2002 onwards
• Commission Regulation No 1614/2002
• mainly to prevent too many confidentiality
  problems

17 September 2003
7

• 2 Data availability
• NewCronos Eurostat dissemination database
• Reference and updated database
• theme4/sbs/sizclass
• Annual enterprise statistics broken down by size
  classes
• Historical database
• theme4/hist_theme4/SME
• old joint OECD/Eurostat data collection
• no longer updated, as a NewCronos warning clearly
  puts it

17 September 2003
8
17 September 2003
9

• EU-15 totals data by size class available for
  reference year 1999 and 2000
• Germany lt 20 not available beforehand
• Acceding countries totals ACC (10 countries)
  available in NewCronos end of September 2003
• EU-25 available in NewCronos on 1/5/2004

17 September 2003
10

• 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
11

• 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 gt 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
  SBS-Regulation
• Germany, among others, made dramatic efforts top
  comply with the Regulation and cover lt20...

17 September 2003
12

• Remaining problematic countries as regards cut of
  value
• Greece
• gt10 persons employed in all industries
• Ireland
• gt20 persons employed in Construction
• Hungary
• gt5 persons employed in all industries
• Thats it for reference year 2001!
• Other countries gt1 person employed in all
  industries
• More problems for historical data...

17 September 2003
13

• 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
14

• 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
15

• Cells of detailed tables contain information
  either on a single or very few respondents
• --gt 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
16
17 September 2003
17

• 5 Quality indicators

17 September 2003
18

• 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
19

• Other problems posed for large businesses
• industrial heterogeneity
• secondary activities
• statistical coverage of enterprise groups

17 September 2003
20

• Conclusions
• Eurostat database on SMEs of high quality
• timeliness
• data available at T20
• accuracy
• CVs available
• availability
• completeness
• coverage of small businesses almost fully
  achieved
• comparability
• legal framework for concepts and definitions

17 September 2003
21

• 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 RD 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