The introduction of base year 2005 and NACE Rev. 2 in STS

1
The introduction of base year 2005 and NACE Rev.
2 in STS
Report on progress since the 2006 STESWP meeting

• OECD Short-term Economic Statistics Working Party
  (STESWP) 25-27 June 2007

2
Introduction base year 2005 NACE 2 in STS

• STS WG Dec-07 mandate to TF
• Task Force STS Backcasting
• 1st meeting 26 April 2007
• participants from 10 Member States and OECD
• discussion on estimation of 2005 weights
  according to new NACE and backcasting

3
Introduction base year 2005 NACE 2 in STS

• Main TF conclusions
• Weights 2005 according to NACE 2
• Problems appear to be more or less solvable (most
  to rely on dual coded register)
• Provisional weights in time (June 2008)
• Some countries may wish to revise weights around
  2010
• Backcasting of STS time series
• Countries less advanced priority so far on
  rebasing
• More attention needed in 2008

4
Introduction base year 2005 NACE 2 in STS

• Main TF conclusions (cont.)
• Countries who have not started yet their
  preparations are urged to do so!
• STS statisticians should co-ordinate themselves
  with experts from related domains and from
  horizontal groups (Structure Statistics,
  registers, IT, etc.)
• Gaining experience from non EU countries could
  help
• STESWP group the main opportunity for this

5
Backcasting issues to resolve

• Choice of backcasting method
• Macro versus micro method?
• Criteria quality, resources (time, skills,
  access to data, financial)
• Little use of benchmarking/interpolation
  techniques
• Macro and micro method!

6
Backcasting issues to resolve

• Length of the backcast time series
• Users need to maintain current length of time
  series
• Target time horizon minimum 2000 for backcasting
  STS
• Longer backcasting for higher level aggregates
• Trade-off between length and quality which
  approach?
• NSO are best placed to carry out this job!

7
Backcasting issues to resolve

• Past base years linking of series
• Most will calculate conversion coefficients for
  2005 only
• Can the introduction of past base years weights
  improve the quality of our backcasting results?
• Is linking an issue?
• Gaps in data availability
• Some (few) activities were not or not
  sufficiently covered under the old classification
• Need to identify cases, assess importance,
  develop solutions!

8
Introduction base year 2005 NACE 2 in STS

• Next steps
• Eurostat manual on backcasting to become
  available soon
• 2nd meeting of TF STS Backcastingin October
  2007
• focus on backcasting
• TF work continues through 2008
• STS WG in December 2007
• draft Commission Regulation on backcasting

9
Issues for discussion

• STESWP members experience with deriving
  concordance matrices from dual coded registers?
• Comparability with those generated from micro
  data?
• Problem areas / biases that are likely to occur?
• Reweighting past base years with concordance
  matrix and linking
• How to identify weak points in the concordance
  matrix and adjust these back in time
• Which points in time would adjustments be made
  (e.g. at base years?)
• What does this imply for linking
• Do STESWP members have experience in producing
  longer time series for higher level aggregates?
• What level of aggregates are produced (e.g. 1
  digit, 2 digit)