Title: Chinas InputOutput Survey and Tabulation Method Comments 13th OECD NBS Workshop on National Accounts
1Chinas Input-Output Survey and Tabulation
MethodComments13th OECD NBS Workshop on
National AccountsNovember 40- December 4,
2009Haikou, China
- Contact nadim.ahmad_at_oecd.org
2General Observation
- IO and SU tables are extremely important
components of the national accounts framework and
on-going work and developments in China are
welcomed and strongly encouraged.
3Detailed Comments
- Paper describes a number of survey sources used
to populate the output, intermediate consumption
and value-added cells in the IO table. - 52 questionnaires categorised by 3 types
- Census questionnaires
- Focused questionnaires
- Typical questionnaires
4Detailed comments
- Census questionnaires
- Investment composition of fixed assets30m Yuan
- Composition of manufacturing costs of the
large-sized industrial enterprises above
designated size 5m Yuan - Gross industrial output and sales volume and
output of major industrial products - Investigation information of costs and expenses
of industrial enterprises
5Detailed comments
- Points of clarification
- Not clear which enterprises are included all?
or just manufacturing? - How are small and medium enterprises (lt5 m Y)
captured? - What use is made of the business register? And
how are survey results grossed-up to the entire
population? What percentage of enterprises (as a
per cent of value-added or output are surveyed?) - How is the business register updated?
6Detailed comments
- Focused questionnaires basic units of medium
and small sized industrial enterprises above
designated size of construction and service
enterprises - Points of Clarification
- Are these questionnaires for the construction and
service industries? Or are they questionnaires
that target expenditures on these activities by
small and medium enterprises?
7Detailed comments
- Typical questionnaires compositions of
transportation expenditures and travelling
expenses etc - Points of Clarification
- More detail required on these questionnaires?
Its not entirely clear what they are and how
they are used.
8Commodity by commodity
- Tables are produced on a C by C basis.
- Questions
- Given the information available to produce C by C
tables it should be possible to produce
corresponding SU tables. Is the reason they are
not produced because it is assumed that IO
commodity coefficients (cost structure) can be
used to estimated the intermediate consumption
and value-added of enterprises not directly
surveyed (but which have been able to provide
estimates of output)? - If this is the case, it is important to better
understand at what level of industry
classification it is applied as it presupposes
strong homogeneity in production processes
which is questionable when comparing large and
small enterprises. - How is value-added allocated when secondary
products are by-products?
9General questions
- Are estimates of own-account production made?
(software, construction etc) - How is the informal economy (and unincorporated
enterprises) measured? - How are estimates of margins on products derived?
- How detailed are the tables (commodity
classification)?
10General questions
- Are estimates of own-account production made?
(software, construction etc) - How is the informal economy (and unincorporated
enterprises) measured? - How are estimates of margins on products derived?
- How detailed are the tables (commodity
classification)?
11Responding to NBS questions
- Which level of the 5 levels of CPC should be used
in commodity balancing for compiling accounts of
goods and services and for balancing of further
supply and use? Could we say the 4th or the 5th
level of CPC must be adopted? - The SNA is not prescriptive in this regard. It
depends entirely on the information available in
a country. The general principle is that the
compilation of tables (especially those that
require assumptions based on observed
coefficients in surveyed enterprises) should be
as detailed as possible .The publication level of
tables however need not (and rarely does) follow
the compilation level. Generally these are less
detailed depending on the quality of balancing,
microdata and confidentiality constraints.
12Responding to NBS questions
- In US, information from economic census is very
rich, with detailed output data of goods and
services, and the complete output data by
detailed commodity can be gathered. However, for
countries where there is not economic census yet,
or just part of establishments has been
investigated, how are the complete output data by
commodity compiled, or how are the complete
output data by commodity imputed? - A business register is a pre-requisite. Most
business registers will also provide an
indication of size either via turnover or
employment. And this information can be used to
stratify enterprises into different bands. But it
is essential that at least some enterprises are
surveyed in each of these bands.
Tax/administrative records provide an additional
useful source for compiling a BR.
13Responding to NBS questions
- In addition, must the objects of investigation be
establishments? Could the enterprises be objects
of investigation, while the industry to which the
enterprise is attached is determined by nature of
the enterprises central product? In other words,
for better adopting the suggestions by SNA, what
are the most essential requirements? - The SNA still prefers establishments. However in
practice there is an increasing tendency to
collect information from enterprises (and this
forms part of the 2008 SNA research agenda). Many
EU countries for example already compile their
tables using enterprise information which is
useful for analytical applications such as
productivity measurement, RD, innovation etc.
14Responding to NBS questions
- How should auxiliary activities be captured? Are
different questionnaires needed? How are
auxiliary activities/expenditures allocated to
commodities? - Para 5.9 of the 1993 SNA defines ancillary
activities. The costs of these activities must be
allocated to using establishments/commodities and
these can be allocated on the basis of
(preferably) value-added or output shares.
15Responding to NBS questions
- Wed like that you introduce us the main
information used and compilation procedures in
compiling supply and use tables through case
studies of a developed country. How do you
account the value added and its composition by
industry when you compile the use table? Do you
use investigation data of establishments to
calculate the value added by production approach?
Whether is the operating surplus derived by
aggregate of value added minus the sum of the
three components of compensation of employees,
net taxes on production and depreciation of fixed
assets? Or other calculation methods are adopted? - See the previous US BEA presentation.
16Responding to NBS questions
- When supply table and use table are compiled,
which level of classification by industry and by
commodity is used at least? - This varies enormously by country. It is
important to note too that the level of detail
used for compilation is usually different to that
used for publication. In the EU countries are
currently required to provide tables at the NACE
60 level but many produce tables at a more
detailed level.
17Responding to NBS questions
- How the basic price, the producers price and the
purchasers price are reflected in input-output
accounts? Which kinds of taxes are included in
each of them? - The 1993 SNA provides clear guidance here. In
most countries IO tables are produced at basic
prices and so will include other taxes and
subsidies on production. Tables compiled at
producers prices will additionally include any
taxes/subsidies on products (e.g. VAT) that are
not separately invoiced to the purchaser. - Purchasers prices include all taxes/subsidies
on products. - Supply-use tables in theory should show
value-added and output at basic prices and
consumption at purchasers prices.
18Responding to NBS questions
- How do you deal with value-added tax? Is the
sales tax added to the output in basic price, and
the purchase tax added to the intermediate input?
Or is the value-added tax that should be turned
over to the state added to the output in basic
price, while without consideration of the
purchase tax in the component of intermediate
input? - The answer depends on the valuation basis used
for the tables. IO tables produced at basic
prices should have a separate row for any VAT
incurred by enterprises (which is rare for market
enterprises but not for non-market and small
enterprises). In other words for IO tables at
basic prices output should not include VAT
neither should any of the transactions between
industries/commodities.
19Inter-relationship between Chinas IO estimation,
production-based GDP and expenditure based
GDP13th OECD NBS Workshop on National
AccountsNovember 40- December 4, 2009Haikou,
China
- Contact nadim.ahmad_at_oecd.org
20General Comments
- Recommended reading 8th Workshop on National
Accounts - 5. Accounting frameworks for constant price
measures- GDP Balancing in Constant and Current
Prices, Mr. AHMAD Nadim, OECD - gt Paper in English (PDF)gt Paper in
Chinesegt PowerPoint presentation
21General Comments
- Recommended reading 11th Workshop on National
Accounts - Supply-use tables techniques for consistent
estimates of missing data - AHMAD, Nadim (OECD)
- PPT doc. english