Aucun titre de diapositive

1
Methodology of CEPIIs Database on Trade in
services
OECD-Eurotat Meeting of Experts on TIS 26 - 29
April 2004
Isabelle Rabaud, Research fellow at CEPII and
lecturer at University of Orléans
2
Purposes of CHELEM-Database

• CHELEM-Database provides coherent long term time
  series on
• International Trade,
• GDP,
• Balance of Payments.
• In the version of CHELEM-BOP of summer 2004, data
  from 1967 to 2002 will be available

3
TIS in CHELEM-BOP

• 13 items of Trade in services
• 4 aggregates
• Transportation services,
• Other services,
• Commercial services,
• Total services.
• Data are in current dollars.

4
SOURCES

• Balance of Payments database of the IMF is the
  primary source of information
• Completed by national data for Taiwan, for
  instance
• For Hong Kong, data on early years, prior to 1998
  are repatriated from the old CHELEM-BOP database.

5
Method to insure coherence of the data

• First, inconsistencies between each aggregate
  (transport and Other services and the sum of
  its components are corrected.
• Second, missing data for the latter years or
  inter-period of less than four years ahead are
  estimated.
• Third, for the beginning of period, data
  registered in BPM4 nomenclature, including
  memorandum items, have been reallocated to the
  new classification recommended by BPM5.

6
Consistency of the aggregates (1) usual causes

• If all components are null, the aggregate is
  reallocated to a residual item
• 233 Transport, n.i.e or
• 982 Services not included elsewhere ,
• If gap(Aggregate - S components) lt 0 for credit,
  the aggregate is recalculated as the sum of
  components.
• If gap 1 component, the aggregate is
  recalculated as the sum of components.
• If S components lt aggregate, residual item
  initial value gap.

7
Consistency of the aggregates (2) causes
specific to transport

• Decomposition by mode seldom available before mid
  1980s or 1990s
• Breakdown by mode is estimated in 3 steps
• Broad geographic areas are defined as in DOTS
  with distinction of enclosed and non-enclosed
  countries,
• For a given country, for each missing year, the
  product of the aggregate by the average share of
  the mode of the area to which it belongs is
  reallocated to each mode.
• For early years, the average shares of some areas
  are not available then the aggregate is entirely
  reallocated to the residual item.

8
Consistency of the aggregates (3) causes
specific to services

• No reallocation is made for other services.
• For total services
• From 1985 to 1990, for Japan, each service is
  estimated by the product of the aggregate and the
  average share during 1991 to 1993.
• For Korea, the average share from 1999 to 2001
  estimates 2002.
• For Vietnam, the average share of each service in
  Asia is used.

9
Estimation of missing data

• Missing data for the latter years or inter-period
  of less than four years ahead are estimated
• For each item and each year, the share of each
  country in global transactions is calculated
• For each item, each missing value is then
  calculated as the share of the country in global
  operations during the preceding year.

10
Matching BPM4 and BPM5 classification