1. Session 9: Data Dissemination Principles
TRAINING COUSE ON EXTERNAL TRADE STATISTICS
Luxembourg– 17/19 April 2012

2. Outline International recommendations on data dissemination
SDDS (IMF)
GDDS (IMF)
UNSD Recommendations
Practises in the EU
5 .Confidentiality
5.1 What is active/passive confidentiality?
5.2 EU legal basis
5.3 Good practices
6. Data revision

3. 1.SDDS SDDS: Special Data Dissemination Standard
Proposed by the IMF following International Financial Crises
Role of information deficiencies and importance of transparency in the dissemination of economic statistics
Subscription to the SDDS opened to countries in 1996

4. 1.SDDS
SDDS focuses on the most important data for assessing macroeconomic policy:
Real sector : national accounts, production index, employment, wages, CPI, PPI,…
Fiscal sector : public sector operations, central government operations and debt,…
Financial sector : interest rates, share price index,…
External sector : Balance of payments, merchandise trade,…

5. 1.SDDS SDDS covers 4 dimensions:
Data: coverage, periodicity, timeliness
Access by the public
Integrity of the disseminated data
Quality of the disseminated data

6. 1.SDDS
“Monitorable elements” of the four dimensions that subscribing countries have to follow:
Data:
Disseminate the prescribed categories of data with the specified periodicity and timeliness
Access by the public
Disseminate advance release calendars for the data
Release the data to all interested parties simultaneously
Integrity of the disseminated data
Disseminate the terms and conditions under which official statistics are produced, including rules of confidentiality
Identify internal government access to data before release
Identify ministerial commentary on the occasion of statistical releases
Provide information about revision and advance notice of major changes in methodology
Quality of the disseminated data
Disseminate documentation on methodology and sources used in preparing statistics
Disseminate component detail, reconciliations with related data and statistical frameworks that support statistical cross-checks.

7. 1.SDDS Subscription to SDDS is on a voluntary basis
Subscription means: to observe the various requirements and provide the necessary information to IMF
The following countries have subsribed to SDDS: Egypt, Israel, Jordan, Morocco, Tunisia

8. 1.SDDS SDDS prescriptions for merchandise trade statistics:
Coverage: imports, exports
Periodicity : Monthly
Timeliness: 8 weeks (4-6 weeks encouraged)
Prescriptions respected by most MPCs…

9. 1.SDDS “Other considerations”:
Flow data for a reference period should cover transactions for the period, not cumulative totals showing transactions added from one period to another
 Production of monthly seasonally adjusted time series
Important to allow users to assess the quality of data:
dissemination of documentation on methodology and sources
Provision of statistical cross-checks and assurance of reasonableness

10. 1.SDDS Information to be provided to the IMF by subscribers:
Advanced release calendars
Base page
Summary methodology
Link to the National Summary Data Page (NSDP)

11. 2.GDDS General Data Dissemination System More flexible than SDDS
Takes in account the different levels of statistical systems and resources
Voluntary basis
Participating countries: Algeria, Lebanon, Palestine

12. 2.GDDS General Requirements
Commitment to use the GDDS as a framework for the development of national systems for the compilation and dissemination of data in the economic, financial and socio-demographic sectors
Designation of a country coordinator to work with the IMF
Production of metadata on current practices and action plans for improvement.

13. 2.GDDS Requirements for trade in goods Data to be compiled monthly
Data to be published within 8 to 12 weeks after the reference month
Use of International Classifications is encouraged
Production of metadata:

14. 3.UNSD Recommendations
IMTS 2010 recommendations on data dissemination:
Timetable: publicly announce scheduled release and revision dates
Timeliness: encouragement to issue first release within:
45 days for monthly data (at least for major trading partners and basic commodity breakdown
60 days for quarterly data
90 days for yearly data
Encouragement to explore possibility to publish provisional estimates soon after the end of the reference period
Revision:
Encouragement to disseminate all revised monthly, quarterly and annual data to ensure consistency of data available to users
Encouragement to ensure consistency between provisional data and final data
Encouragement to develop a revision policy which is synchronised with the release calendar, transparent and well coordinated with other areas of statistics.

15. 3.UNSD Recommendations
IMTS 2010 recommendations on data dissemination:
Dissemination strategy:
It is recommended that:
all users are treated equally
countries choose the dissemination format that best suits their user needs (electronic or paper publication)
IMTS be accessed through the electronic dissemination databases maintained by the responsible agency
Other recommendations:
Use of passive confidentiality (see part 5)
Publication of volume and unit value (or price) indices recommended (at least monthly for total imports and exports). Quarterly indices are encouraged for commodity groups.
Publication of seasonally adjusted data encouraged
Countries are encouraged to cooperate with international, supranational and regional organizations to identify and apply the most efficient ways of international dissemination of their trade statistics and related metadata (SDMX exchange format)

16. 3.UNSD Recommendations
IMTS 2010 recommendations on metadata dissemination:
metadata: all information used to describe other data (“data about data”): definition of concepts applied, description of the collection/production/ dissemination process…
Taking into account many types of users and uses for any given set of data, a broad spectrum of metadata requirements must be addressed
It is recommended, as a minimum segmentation, to make metadata available at the following two levels:
(a) Reference metadata should be presented in the form of a detailed explanatory note describing the scope, coverage and quality of a data set. It should be made available electronically alongside the database or in a special publication;
(b) Structural metadata should be presented as an integral part of the international merchandise trade statistics database that can be extracted together with any data item; it may be published as part of a statistical table

17. 3.UNSD Recommendations
IMTS 2010 recommendations on metadata dissemination:
It is recommended to cover at least the following categories of metadata:
a) Underlying concepts and definitions, including trade system and deviations from international standards, if any
b) Legal framework, institutional arrangements and data sources;
c) Data collection and processing procedures;
d) Estimation methods;
e) Data dissemination policy, including release and revision schedules;
f) Description of all data fields / variables:* reference period, trade flow, commodity classification, valuation, currency, quantity (net weight), weight unit, supplementary quantity, supplementary quantity unit, partner country
g) explanations and footnotes concerning the data as required, e.g., informing about revisions, breaks in series, and application of confidentiality;
h) quality reporting.

18. 4.Practices in the EU European Statistics “Code of Practice”
Recommendations in line with IMF and UNSD…
Statistical releases are clearly distinguished and issued separately from political statements
Product quality is regularly monitored according to European Statistical System (ESS) quality components
Processes in place to monitor the quality of dissemination
Statistics are coherent or reconcilable over a reasonable period of time
Cross-national comparability of the data is ensured through periodical exchanges between the ESS (and other statistical systems)
Eurostat practice (often followed by EU Member States…)
Periodic Quality reports
Asymmetry studies
Periodic Users satisfaction surveys
Free dissemination on the website, including detailed trade statistics
Publication of monthly seasonally adjusted time series

19. 5. Confidentiality 5.1 What is Active/Passive confidentiality ?
Active confidentiality : is applied when National Statistical Authorities checks all data before disclosure and take the initiative to suppress data automatically (without informing the exporter or importer concerned) when data allows to identify directly the amount of trade carried out by the exporter or the importer.
General rule: data are disclosed only if there are at least 3 reporting companies, without one dominating company
Passive confidentiality : means that the exporter or the importer concerned has to take the initiative and ask the National Statistical Authorities to keep the data confidential. The combinations of around 10,000 commodity codes broken down by trading partner must entail numerous examples where only one or perhaps two enterprises are involved in the trade. This means that the information provided to users would be significantly reduced under a system of active confidentiality. That is why passive confidentiality is generally applied in the field of external trade statistics
Remark: Data with direct identification of importing/exporting companies are always confidential

20. 5. Confidentiality 5.1 What is Active/Passive confidentiality ?
Active confidentiality is generally the principal referred to in basic national statistical laws for most statistical domains (households and business statistics)
The application of active confidentiality to external trade statistics would have severe consequences on the quality of statistics: in many countries more than half of detailed data (product X partner country level) would need to be hidden
This is why passive confidentiality is applied in the field of external trade statistics in almost all countries.
It is recommended to state clearly the use of passive confidentiality in national statistical laws.

21. 5. Confidentiality 5.1 What is Active/Passive confidentiality ?
UNSD recommendations (IMTS 2010; par. 10.3)
For trade statistics it is common practice that trade data are made confidential only when the trader requests so and the statistical authority finds the request justified based on the confidentiality rules. This approach to confidentiality is called “passive confidentiality” as opposed to “active confidentiality” which aims to systematically prevent the identification of any statistical unit in the disseminated data, thereby disclosing individual information. In view of the high demand for detailed trade statistics and the burden and potential difficulties of applying active confidentiality to trade statistics, it is recommended that passive confidentiality is applied as much as possible, unless the use of active confidentiality is already the established, desired and accepted practice.
It is further recommended that in suppressing data due to confidentiality, any information deemed confidential (suppressed) be reported in full detail at the next higher level of commodity and/ or partner aggregation that adequately protects confidentiality. Countries might develop other measures which can serve the same purpose of protecting confidentiality while making information available to the best possible extent

22. 5. Confidentiality 5.2 EU legal basis
Extrastat basic legislation (2009/471: Article 10, par. 1):
Only where an importer or exporter so requests shall the national authorities of a given Member State decide whether the external trade statistics of that Member State which may make it possible to identify that importer or exporter are to be disseminated or are to be amended in such a way that their dissemination does not prejudice statistical confidentiality.
Extrastat Commission Implementation regulation (92/2010: Article 2, par. 5)
The statistics shall contain data affected by confidentiality provisions in the compiling Member State. National statistical authorities shall flag the data to be considered as confidential in such a way that as much information as possible can be disseminated, at least at chapter level of the Combined Nomenclature, provided confidentiality is thereby ensured.
 Extrastat/Intrastat legislation makes the use of passive confidentiality methodology compulsory for EU Member States

23. 5. Confidentiality 5.3 Good practices…
Use passive confidentiality in external trade statistics
Include a clear statement in national Statistical Law about the use of passive confidentiality for external trade statistics
Inform clearly private companies about their possibilities to request confidentiality for their trade
When data are made confidential, keep data available at the most detailed possible level of classification (keep at least the real Chapter HS2 to hide data)
Assess periodically the need to confidentialise
Assess periodically the impact of confidentiality on the data accuracy

24. 6.Data revision IMTS 2010: Main causes of revision:
Revisions are an essential part of country practices
Consequence of the trade-off between timeliness and accuracy
Compilation of provisional data is encouraged
For normal statistical data revisions, countries are encouraged to develop a revision policy which is synchronized with the release calendar, and coordinated with other areas of statistics…
Main causes of revision:
Usual revisions:
Late submission of data by declarants
Correction from traders or from the administration
Unusual revisions:
Changes concerning methodology (trade system, coverage…)

25. 6.Data revision Practical recommendations:
Include the national practice on revisions in metadata
Flag the provisional data in the press releases
Release revised data as soon as possible
Monitor the change of total trade value between the first results and the last available results
If the changes are significant and permanent, try to estimate future revisions in the the preliminary data

26. Thank you for your attention
6.Data revision
Thank you for your attention
Questions and remarks ?