The French Business Register : from a quality approach

1
The French Business Register from a quality
approach ..to a statistical register

2
Description of SIRENE

• SIRENE
• a national business register in which all
  businesses and their local units are registered
• since 1994 a unique identification number, at
  the core of the inter-administrative system
• Content of SIRENE
• identification data civil status of businesses,
  legal form, the date it was set up and when it
  ceased trading
• address of local units
• a code identifying the principal business
  activity carried out, number of employees
• as SIRENE is used in the areas of statistics and
  trade economic classification variables (list
  drawn up after consulting statisticians using the
  system)

3
Inventory of users and requirements
Users Some requirements
Inter-administrative use (administrative partners involved in managing the register and businesses) Unique identification Exhaustiveness (even units which are registered only for the needs of a partner) Legal termination
Dissemination of information Address, Activity code
Statistical use business demography basis for sampling survey managers Exhaustiveness but reference groups Identification Stratification data Economic termination
4
An administrative approach of the quality

• Quality is taken into account in day-to-day
  management
• a workstation with automated processing
  capabilities (receiving standardized
  administrative information via EDI) which
  guarantees uniform processing
• checks into the data-processing flow
• Quality is taken into account as an independent
  component (quality operations)
• at the initiative of register managers
• at the request of those in charge of surveys
• The register managers bring into play quality
  with the limitations of an interadministrative
  register
• ? Quite administrative approach of the quality

5
Volumes of quality operations

• Updates
• each year 100,000 legal units with their activity
  code updated-validated
• 3.5 million units (legal units or local units)
  with employee statistics
• 2 million businesses with turnover statistics
  updated
• Checks
• 15,000 units checked at the request of
  statisticians (wrong address for example)
• 15,000 possible duplicate entries examined
• 50,000 units possibly misidentified as
  operational examined
• Surveys
• in 2007, 200,000 surveys-questionnaires sent for
  the introduction of the NACE

6
Links between  quality operations  and Eurostat
quality report ? 1/2

• Relevant component
• the identification of statistical users and of
  their expectations taken into account
• some quality operations based on the request of
  statisticians
• Accuracy component
• meaningless
• but the quality operations improve the
  accuracy of the sampling plans.
• Timeliness and punctuality
• register daily updated by administrative partners
  (sometimes insufficient ?)
• mass updates data are included as soon as
  available
• discrepancies examined as quick as possible.

7
Links between  quality operations  and Eurostat
quality report ? 2/2

• Accessibility and clarity of the information
• frozen versions of the business register (for all
  users or for specific needs) produced quarterly
  and yearly
• each month an extract of the register for the
  staff responsible of the dissemination of
  information
• for all statisticians a specific tool, to get
  all the data of a list of units, at a given date
• Comparability of statistics
• few statistics spring from the register except
  business demography.
• the day-to-day management of the register and the
  advanced level of automation provide
  comparability of statistics (in the time)
• Coherence
• maybe a lack of the register ?

8
Towards a new statistical registerProspective
1/2

• Review the use of SIRENE as a central tool for
  business statistics comes up against its twofold
  role
• legal vs economic cessation
• exhaustiveness vs reference groups
• Recent overhaul of the annual business
  statistics one of its outputs to come is a
  statistical register
• Aims of the statistical register
• based on the inter-administrative register
• and/but closer to the requirements of the
  statisticians.

9
Towards a new statistical registerProspective
2/2

• New concepts
• the definition of statistical units
• the addition of specific data
• the definition of reference populations
• A new quality management
• statisticians will manage quality
• a quality operation checking all mistakes in
  a population
• more statistical updates
• more automatized updates
• gt classification of sources (to solve
  inconsistencies between several sources)
• Maybe the definition of indicators to measure
  quality of this statistical register