Methodological issues in annual chainlinking the output measure of GDP - PowerPoint PPT Presentation

1 / 22
About This Presentation
Title:

Methodological issues in annual chainlinking the output measure of GDP

Description:

Fixed base chain-linking: out-of-date weights x. judgement to chose ... Effect of new 5 year weights. for fixed base chain-linking (on growth from 1994 to 1997) ... – PowerPoint PPT presentation

Number of Views:31
Avg rating:3.0/5.0
Slides: 23
Provided by: ONS1
Category:

less

Transcript and Presenter's Notes

Title: Methodological issues in annual chainlinking the output measure of GDP


1
Methodological issues in annual chain-linking
the output measure of GDP
2
The message...
  • From Blue Book 2003, growth estimated using
    annual chain-linking
  • Eurostat legislation and improvement in our
    growth estimates.
  • details of the methodology in development.

3
Summary of presentation
  • Current method
  • What is annual chain-linking?
  • Eurostat requirements
  • UK decisions
  • Future developments

4
Current method by industry contribution
  • For each industry
  • compile constant price series to show growth
  • currently we reference to 1995100
  • series might look like this

5
Current method of adding up
  • To calculate a series showing growth for a group
    of related industries
  • need to weight each industry series by how big a
    share it is of overall economy
  • repeated until reach overall GDP
  • weights are constructed using data from a fixed
    year or fixed base
  • weights updated every 5 years, chain-linked
    sub-series of 6 to 9 years

6
Annual chain-linking
  • Annual chain-linked figures use most recently
    available weights
  • process produces pairs of years on each weight
    structure
  • pairs of years are chain-linked together to
    produce a continuous series - process same as for
    fixed base

7
Differences
  • Fixed base chain-linking
  • out-of-date weights x
  • judgement to chose link year x
  • possible big revisions from new weights-
    rebasing x
  • but components for several years add-up ?
  • Annual chain-linking
  • weights from previous year ?
  • every year a link year ?
  • updating year of weights not relevant ?
  • fewer years add-up x

8
Improvements to relevance of weights (percentage
of all industries)
9
Effect of new 5 year weightsfor fixed base
chain-linking (on growth from 1994 to 1997)
10
Choice of link year
11
Modelling the effect on growth estimates A
  • Post and telecommunications

12
Modelling the effect on growth estimates B
  • Transport, storage and communications

13
and effects on even faster-falling prices
14
European requirements
  • Principle 1 aggregate from elementary level
  • Principle 2 use base-weighted indices (or Fisher
    indices)
  • Principle 3 if base-weighted, use weights from
    the previous year.

15
Types of Indices
p price q volume p t q t weight
Base-weighted (Laspeyres)
Fisher
Base-weighted - need weights from base
year Fisher - need weights from now
16
Decision for monthly and quarterly data
  • Quarterly data linked on quarter 4 and then
    bench-marked to match annual
  • Monthly data linked on mean of quarter 4 months
    and then bench-marked to match annual

17
Decision for Seasonal Adjustment
  • Chain-link
  • Benchmark
  • then Seasonally Adjust
  • Chain-link
  • then Seasonally Adjust
  • Benchmark

?
18
Investigation of the elementary levelDifference
in growth figures (intermediate level start
minus detailed level start)
19
Below annual chain-linking level
20
Future developments
  • Annual chain-linked data for GDPo in Economic
    Trends October 2001.
  • Work on consistent input level.
  • Prototype systems for Expenditure measure
    components.
  • Long-term move to constant price balancing at a
    lower than overall level.

21
The message...
  • From Blue Book 2003, growth estimates from
    annual chain-linking
  • Eurostat legislation and improvement in our
    growth estimates.
  • details of the methodology in development.

22
Improvements to relevance of weights (percentage
of all industries)
Write a Comment
User Comments (0)
About PowerShow.com