Revision of National Accounts Time Series in the Netherlands

1
Revision of National Accounts Time Series in the
Netherlands

• Ruurd Schoonhoven
• ( RSNN_at_CBS.NL)
• Gert den Bakker, Bram de Boo, Ruben van der Helm
• Statistics Netherlands

2
Time series revisions in the Netherlands

• Systematic time series revisions
• - after benchmark revisions of 1977, 1987, 1995,
  2001
• for supply and use (and I/O) tables, sector
  accounts and labour accounts
• Different methods were used .
• . with different costs (time, resources)
• . and with different quality of results
• How to proceed in future?

3
Suitability criteria for time series revisions

• Resources
• Time
• Quality
• Transparency
• Flexibility

4
Approaches to time series revisions

• 1. Annual backward calculation
• 2. Benchmark years and interpolation

5
I. Annual backward calculation

• I.1 From microdata, with new concepts and
  classifications (from scratch)
• I.2 Adjustments superimposed on balanced
  pre-revision data
• I.3 Linking pre-revision (or revised) growth
  rates to results for revision year

6
II. Benchmark years and interpolation

• II.1 Benchmark years from scratch
• II.2 Benchmark years by superposition of
  adjustments on balanced pre-revision data
• Intervening years by interpolation with
  pre-revision (or revised) growth rates

7
The 2001 time series revision

• Conceptual revisions superimposed on pre-revision
  data year by year (I.2)
• Statistical revisions interpolated between
  benchmark years (II.2)
• Balancing of the results, year by year from 2000
  back to 1969/1980
• Strict process, good quality results, but (again)
  required much time and resources..

8
Requirements for futuretime series revisions

• Harmonisation of revision policy within EU
• Users want revised time series faster
• Budget reduction

9
Points of consideration (1)

• EU regulations I/O tables back to 2000, main
  aggregates back to 1980
• Main user detailed gt30 yr time series

10
Points of consideration (2)

• From scratch or using pre-revision results
• Working level of detail
• Compiling the time series in parallel with the
  revision year

11
Points of consideration (3)

• How to avoid (reduce) the need for year-by-year
  balancing?
• Automatic balancing methods

12
Discussion

• International harmonisation of revision policy
  for time series?
• Inventory of methods currently used in different
  countries