Transcript Agenda item 8 Implications of the new treatment of goods for processing in the supply and use tables Discussion Joint UNECE/Eurostat/OECD Meeting on National Accounts Geneva,

Agenda item 8
Implications of the new treatment of goods
for processing in the supply and use tables
Discussion
Joint UNECE/Eurostat/OECD Meeting on
National Accounts
Geneva, April 24, 2008
Bent Thage
Production units and institutional units
• Dual classification
• Production units (Establishments and industries)
(Down to the generation of income account)
• Institutional units (Enterprise units, instit. sectors)
(All accounts)
• What does ”change of ownership” mean when
dealing with production units (even their
intermediate consumption is not a transaction)?
Institutional characteristics
• Also production units have institutional
characteristics
• Establishments with the same kind of
output may have different input ratios
and structures
• Because of (for example)
• More or less vertical integration
• Renting of owning capital equipment
• Different degree of outsourcing
What can be done about it?
• In general these institutional
characteristics cannot be removed
• Therefore the notion of ”technological
structures” has always been misleading
• Therefore the text qouted (p.11) from
SNA, Rev 1, Par. 14.2 is also
misleading, and may be changed.
• What we have is a table of (interindustry)
transactions, even in the case of a
product x product symmetric input-output
table.
However…..
• The choices we make when compiling
the supply and use tables should be
made with the subsequent analytical
uses in mind
• The choice of industrial detail
• Redefinitions (for example agriculture
and trade), grossing up etc.
• While retaining as much as possible the
micro-macro links.
• And consistency within the NA system
How big is the problem?
Denmark
2005
2006
2007
Billions DKK
Imports for
processing
1.489
5.439
5.557
Exports from
processing
2.003
5.762
8.160
Exports for
processing
1.436
2.067
2.151
Imports from
processing
1.693
2.448
2.480
Total exports
of goods
496.207
538.248
548.747
The balancing problem
• Foreign trade statistics (goods) will
remain on a border-crossing basis
(Intrastat special case)
• Processing industries and client
industries report on a net basis
• Is this balancing problem different in
nature from many other, that we have to
deal with?
• Difficult to justify additional data
collection, but a few additional questions
in existing surveys may be OK.
Analytical issues I
• Present treatment exaggerates
import intensity and export
performance of goods producing
industries
• Understates the value of
international trade in services
• Calculate import content of exports
to get the correct picture?
Direct and indirect imports contents
Direct and indirect import contents of final uses.
Denmark
1975
1985
1995
2000
2004
Private consumption
23,24
24,19
20,67
22,94
23,25
Government consumption
10,65
10,99
10,08
10,15
10,76
GFCF. Machinery and equipment
50,91
56,32
52,98
56,85
54,68
GFCF. Transportation equipment
66,27
42,15
43,06
56,22
53,29
GFCF. Buildings and structures
19,15
20,62
17,10
16,89
17,51
Other GFCF
34,69
31,04
21,58
32,27
27,51
-31,26
67,38
64,89
58,48
82,96
34,58
39,67
38,31
43,18
44,06
Changes in inventories
Exports of goods and services
Analytical issues II
• The (too large) import coefficients
leads to an understatement of
impact coefficients
• Productivity measures (par. 14-15).
Are there any adverse effects?
Labour productivities
Tabel 7.14 Average annual increases in labour productivities 1975-2004. Denmark
Direct
-------------------------
Direct and
indirect
Per cent
Global
-------------------------
1. Agriculture etc.
6,05
5,94
5,30
2. Manuf. industries
2,81
3,20
2,60
3. Energy-, water etc..
3,16
1,63
1,99
4. Construction
1,56
1,61
1,59
5. Trade etc.
1,49
1,48
1,40
6. Transportation, tele
3,53
4,13
4,14
7. Financial and business
serv
0,94
0,34
0,17
8. Government and pers.
services
0,75
0,75
0,78
Adjustments of data in new treatment
• Merchandise exports and imports must
be adjusted for the value of ”goods for
processing” (Requires additional
information – international agreements)
• The net values of the ”removed” exports
and imports must be compared to the
processing costs from client units and
revenue data from processing units.
• Must take place at the lowest level of HS
Additional data collection
• For client units: Question on value of
goods of own manufacture that are sent
abroad for processing, the postprocessing value, and the fee paid.
• For processing units: The processor
will not know the market value. Can give
information on the type of goods
processed, and the cost of the
processing service.
Impact on the supply and use tables
• Present recording a mixture of net and gross
treatment
• Role of the supply and use table:
- Balancing in an integrated framework
- Derive GDP and double-deflated data
- Database to derive structural measures
• New treatment and the ”technology” question
• Drawback that (some) forward and backward
linkages will disappear.
The oil refinery example
• The refining of crude oil is done by a contract
processor (an oil refinery)
• The oil refinery has only ”refining services” as
output
• The linkages between:
- crude petroleum extraction
- oil refinery (in Canada)
- Production of petroleum by-products (start)
will be interrupted.
• In which branch is the client classified?
Multiregional supply and use tables
• Linkages not only across production processes,
but also across regions via inter-regional trade
matrices
• A net treatment is build into this system as no
”gross” information is available
• If goods are sent to other regions for processing
there are limited possibilities to document
technological dependencies between industries
and regions
• In the oil refinery case an imputation was
exceptionally made to overcome this problem
Conclusion
• The existing treatment have important analytical
and data-collection disadvantages
• The new treatment poses certain
disadvantages for the supply and use tables
• Suggest that both the ”gross” and the ”net”
treatment should be implemented and data
presented to users on both bases
• Question: In that case, why not implement only
the ”gross” version in the supply and use tables,
and make the necessary adjustments to the
SNA, rev1 and BPM6 definitions for the
aggregates (Perhaps only in the BOP)?