The%20OECD/NBS%20project%20

1
The OECD/NBS project Improving the
Monthly/Quarterly Statistics of China

• Marco Marini (Istat)
• Francesco Zollino (Bank of Italy)

OECD/NBS Workshop on National Accounts Haikou,
China November 30, 2009
2
Outline of the presentation

• Motivation of the project
• Activities done
• Main results
• Open issues for future work

3
Motivation 1/2

• Providing training and consultancy to the NBS in
  the challenging project of improving the quality
  of short-term indicators and quarterly GDP of
  China
• Setting up the general framework in line with the
  international best practices
• Addressing the links between source data and the
  quality of monthly/quarterly data

4
Motivation 2/2

• Reviewing guidelines and procedures for seasonal
  and calendar adjustment
• Featuring data requirements and methods for the
  purpose of a sound analysis of short term
  economic developments of China

5
Activities doneSummary

• Contributions mainly focused on methods and
  procedures
• Training
• Consultancy
• Three workshops organized in 2009 at the NBS
• Compilation of QNA estimates and seasonal
  adjustment (June 15-18)
• Short-term indicators, QNA and business cycle
  analysis (June 22-25)
• Assessment of procedures and preliminary results
  (November 23-26)

6
Activities doneCompilation of QNA

• Main features of an indirect approach to the
  compilation of quarterly accounts the case of
  Italy
• Source data in QNA
• GDP production components
• GDP expenditure components
• Estimation methods used
• Temporal disaggregation
• Univariate and multivariate benchmarking
• Seasonal and calendar adjustment

7
Activities doneShort-term indicators and QNA

• Quality requirements for source data in order to
  improve accuracy and reliability of GDP estimates
• Specific issues for transition economies
  (discrete vs. accumulated data, public vs.
  private sector)
• Review of main issues related to short-term
  indicators currently available in China
• Assessment of advantages and pitfalls of basic
  sources sample survey vs. administrative systems

8
Activities doneSeasonal adjustment

• Guidelines on Seasonal Adjustment in the European
  Statistical System
• Presentation of tools for Seasonal Adjustment
  TRAMO-SEATS (Bank of Spain), X-12ARIMA (US
  Census), Demetra (Eurostat)
• Regression-based method for the adjustment of
  calendar effects
• Revisions Policy of SA QNA estimates in Italy
• SA Tutorial on selected Chinese time series

9
Activities doneShort-term analysis and
forecasting

• Tools and purposes of business cycle analysis
• Experience in the analysis and the forecasting of
  main macro-economic variables
• Forecasting key short-term indicators the case
  of industrial production in Italy
• Forecasting GDP through short-term indicators

10
Main results 1/2

• Full understanding of the indirect approach for
  the compilation of QNA
• Acquiring expertise in the practice of seasonal
  adjustment
• Pre-treatment of calendar effects (especially
  Chinese moving holidays)
• First experiments of seasonal adjustment applied
  to problematic Chinese time series

11
Main results 2/2

• Useful discussion on problematic aspects of
  seasonal adjustment direct vs. indirect
  approach, nominal vs. volume SA data
• Knowledge of general framework and main tools for
  the short-term analysis of the economy
• Guidelines for selecting models and indicators
  for short-term forecasts of main aggegates
  (possible use for validating preliminary QNA
  estimates)

12
Open issues for future work 1/3Concerning the
methods

• Careful review of pitfalls and advantages of the
  benchmarking method chosen by the NBS (enhanced
  Denton PFD in a shortcut version)
• Validation of seasonal adjustment results
  according to well-established diagnostic tools
  (statistical tests, revision and stability
  analysis)

13
Open issues for future work 2/3Concerning the
data

• Keeping investing on better quality of raw data,
  also as a key requirement for a sound seasonal
  adjustment (room for more extensive use of sample
  survey?)
• Clear commitment on the dissemination strategy of
  QNA data, with particular concern to the release
  calendar and procedures
• Establishing a coherent, transparent, officially
  published revisions policy of both SA and
  unadjusted data

14
Open issues for future work 3/3Concerning the
organization

• Establishing a group of experts in SA techniques
  in order to coordinate the activity in the NBS as
  a whole in the very early stage of production and
  dissemination of data
• Enhancing training activities to make expertise
  more widespread across different departments of
  the NBS, in view of a progressively
  de-centralization of the SA responsability
• Starting a network with preferred users (such as
  PBC, University and main analysts) to collect
  feedback about the quality of data and future
  statistical needs

15
Thank you!