UNITED NATIONS STATISTICS DIVISION Trade Statistics Branch Distributive Trade Statistics Section

1. DisseminationWorkshop for African countries on the Implementation of International Recommendations for Distributive Trade Statistics27-30 May 2008, Addis Ababa, Ethiopia UNITED NATIONS STATISTICS DIVISION Trade Statistics Branch Distributive Trade Statistics Section

2. Outline of the presentation • National reporting • Statistical confidentiality • Data dissemination timetable • Data revisions • International reporting • List of data items on DTS recommended for international reporting

3. National reporting • Data dissemination • Key activity of NSOs • Responsibilities of NSOs • Provide policy makers, business community and other users with high quality statistical information • Motivate respondents to participate in statistical surveys • Protect confidentiality of respondents

4. Statistical confidentiality • Most of the information about individual statistical units classified in section G of ISIC, Rev.4 is considered to be confidential • Statistical confidentiality is necessary in order to gain and keep the trust of: • Those required to respond to statistical surveys • Those using the statistical information • UN Fundamental Principles of Official Statistics (the sixth principle) • National Statistical Laws or other supplementary governmental regulations • National definitions of confidentiality and rules for microdata access may differ, but they should be consistent with this fundamental principle

5. DTS and statistical confidentiality • NSOs usually disseminate DTS in the form of various statistical tables • The statistical confidentiality is protected if tabulated data do not allow statistical units to be identified either directly or indirectly, thereby disclosing individual information • Direct identification - if data of only one statistical unit are reported in a cell • Indirect identification - if individual data can be derived from disseminated data (e.g. because there are too few units in a cell, or because of the dominance of one or two units in a cell)

6. Statistical disclosure control • Set of methods to reduce the risk of disclosing information on individual units • Manifest themselves at the dissemination stage, however • Pertinent to all stages of statistical production process • Disclosure control methods - attempt to find an optimal balance between the improvement in confidentiality protection and the reduction in data quality • Identification of sensitive cells - first step in statistical disclosure control • Sensitive cells – cells that tend to reveal directly or indirectly information about individual statistical units

7. Methods for protecting confidentiality • Aggregation • A confidential cell in a table is aggregated with another cell and then the information is disseminated for the aggregate and not for the two individual cells • Suppression • Removing records from a database or a table that contains confidential data • Primary suppression - values in sensitive cells are not published while the original values of the others are • Secondary suppression -some other cells are also suppressed to guarantee the protection of the values under the primary suppression • Other methods • Control rounding – the original value of each cell is modified by rounding it up or down to a near multiple of a base number • Perturbation – linear programming variant of the controlled rounding technique

8. Confidentiality rules for DTS • Rules for protecting confidentiality of DTS - should be in line with the provisions of countries national legislation and practice • Confidentiality criteria • Number of units in a tabulation cell • Dominance of a unit or units contribution over the total value of a tabulation cell • Decision in respect to the exact definition of the confidentiality criteria in terms of number of units per cell and percentage of dominance - left to the NSOs • Confidentiality rules may be relaxed in individual cases by requesting the permission of the dominating respondent (s) to disclose their data

9. Recommendations • Countries are encouraged to develop their own statistical disclosure methods suiting best their specific circumstances • If suppression is preferred as a method for protecting confidentiality of DTS • Any data deemed confidential should be reported in full detail at the next higher level of classification that adequately protect confidentiality • The confidentiality protection must result in a minimum loss of information

10. Data dissemination timetable (1) • Advance release calendar • Guarantees well established relations between NSOs as producers of DTS and the user community • Countries should develop and announce in advance the precise dates at which DTS statistics will be released • To be posted in the beginning of each year on the country’s statistical office website • Timelines of DTS data • Monthly – 45 days after the reference month • Quarterly – 3 months after the end of the quarter • Annual – 18 months after the end of the year

11. Data dissemination timetable (2) • Issues of importance for the development of advance release calendar • Timing of collection of initial data from major DTS surveys • To what extent data derived from the major data sources are subject to revisions • Timing of preparation of important national economic policy documents that need DTS as inputs • Modes of data dissemination

12. Data revisions • Essential part of countries practices on compilation of DTS • Occur as a consequence from the trade-off between the timeliness of published data and their reliability, accuracy and comprehensiveness • Conducted for the benefit of users - to provide them with data that are as timely and accurate as possible • Affect both annual and short-term DTS but are more pronounced for the short-term data

13. Reasons for revision • Revisions due to “normal” statistical procedures (current revisions) - conducted at regular intervals of time • New information becomes available • Change in the methodology • Change in data source • Change of base year • Revisions due to the correction of errors that may occur in source data or in processing • Special revisions - carried out at longer, irregular intervals of time • Re-assessing the data • Investigating in depth some new economic structures

14. Revision policy • Countries are encouraged to develop a well-designed, carefully managed and coordinated with other areas of statistics revision policy for DTS • Essential features of a well-established revision policy • Predetermine schedule • Reasonable stability from year to year • Openness • Advance notice of reasons and effects • Easy access of users to sufficiently long time series of revised data • Adequate documentation of revisions included in the statistical publications and databases

15. Dissemination of metadata • Provision of an adequate metadata and quality assessment of DTS is as important to users as provision of data • Countries are encouraged to develop and disseminate metadata on DTS • Metadata components • Data coverage, periodicity and timeliness • Access by the public • Integrity of disseminated data • Data quality • Summary methodology • Dissemination formats

16. International reporting – Annual data

17. International reporting – Quarterly data

18. International reporting – Monthly data

19. Thank You