Fifth Review of the Fund's Data Standards Initiatives

This paper reviews developments under the Fund's Data Standards Initiativ--the Special Data Dissemination Standard (SDDS) and the General Data Dissemination System (GDDS)--and their integration under the Data Quality Program. The paper discusses proposals for updating the SDDS and the GDDS to maintain their relevance and reflect evolving international best practice, and for follow-up on the data module of the Report on the Observance of Standards and Codes (ROSC). The paper also proposes the development of a Compendium on Good Statistical Practices.

Abstract

This paper reviews developments under the Fund's Data Standards Initiativ--the Special Data Dissemination Standard (SDDS) and the General Data Dissemination System (GDDS)--and their integration under the Data Quality Program. The paper discusses proposals for updating the SDDS and the GDDS to maintain their relevance and reflect evolving international best practice, and for follow-up on the data module of the Report on the Observance of Standards and Codes (ROSC). The paper also proposes the development of a Compendium on Good Statistical Practices.

I. Introduction

1. In the wake of the 1994 crisis in Mexico, the international community recognized the essential role of data transparency for meeting the challenges and risks of globalization and reducing the likelihood of financial crises. It called for timely dissemination of reliable macroeconomic and financial data and an improved early warning system that would permit a swifter response to financial shocks.1 With the report of the Executive Board to the Interim Committee of the Board of Governors of the IMF on Strengthening of Fund Surveillance serving as background,2 the 1995 Halifax Summit urged the IMF to take the following steps:

  • establish benchmarks for the timely publication of key economic and financial data;

  • establish a procedure for the regular public identification of countries that comply with these benchmarks; and

  • insist on full and timely reporting by members of standard sets of data, provide sharper policy advice to all governments, and deliver franker messages to countries that appear to be avoiding necessary actions.

2. Taking account of the diversity of the Fund’s membership and differing sensitivities to international capital market movements, the Interim Committee in October 1995 approved the recommendations of the Executive Board for a two-tier approach to standards for the dissemination of data.3 The Special Data Dissemination Standard (SDDS) was established in March 1996 to guide market-access countries on the dissemination of economic and financial data, and the General Data Dissemination System (GDDS) was established in December 1997 to help participating countries develop sound statistical systems as a basis for dissemination of economic, financial, and socio-demographic data to the public. Overall, by end-May 2003, 111 Fund members, or more than 60 percent of the membership, were either SDDS subscribers or GDDS participants (see Chart 1). For background information on the Fund’s data standards, see Box 1.

Chart 1.
Chart 1.

Number of SDDS Subscribers and GDDS Participants

Citation: Policy Papers 2003, 026; 10.5089/9781498329347.007.A001

Source: IMF—Dissemination Standards Bulletin Board (DSBB)

3. In responding to the Asian financial crisis of 1997 and the needs that emerged subsequently, the data template on international reserves and foreign currency liquidity4 and the external debt data category were added to the SDDS. Also, a fixed transition period for the implementation of the International Investment Position (IIP)—an original data category under the SDDS—was established. Appropriate adaptations were made in the GDDS.

4. In the Third Review of the Fund’s Data Standards Initiatives in March 2000 and again in June of that year during the discussion on Data Provision to the Fund for Surveillance Purposes, the Executive Board addressed data quality more comprehensively. The Board welcomed the staff’s efforts to develop a framework that would allow users and compilers to make their own data quality assessments. The extensive research and consultation process that followed included launching the IMF Data Quality Reference Site on the Internet to foster on a worldwide basis a common understanding of data quality. The Data Quality Assessment Framework (DQAF) that resulted from these efforts provides an integrated and flexible methodology to assess data quality.

5. Under the Fourth Review in July 2001, the Executive Board supported the integration of the DQAF into the data module of the Report on the Observance of Standards and Codes (ROSC). At the same time, it also endorsed the integration of the Fund’s various applications of the DQAF in an overall Data Quality Assessment Program. The DQAF has been proven to be robust in a wide range of activities. The staff recommends that henceforth the overarching program be referred to as the Data Quality Program (DQP), as it encompasses initiatives to promote good statistical practices as well as assessment.

6. This paper reviews developments in the Fund’s Data Standards Initiatives since the Fourth Review. Section II covers the SDDS and Section III the GDDS. Section IV reviews and updates the DQP. Section V addresses resource implications. Section VI contains issues for discussion. Appendix I presents the main conclusions of previous reviews, Appendix II lists SDDS subscribers, and Appendix III lists GDDS participants. Appendix IV lists members having undertaken the data module of the ROSC.

Establishment of the Fund’s Data Standards

In March 1996, to guide countries that have, or might seek, access to international capital markets, the Executive Board established the SDDS, which provides guidance in the dissemination of economic and financial data (see SM/96/60 and SM/96/83). It carries a commitment to observe the standard and to provide certain information about practices in disseminating economic and financial data. Subscribers must agree (1) to post information about their data dissemination practices on the IMF’s external website on an electronic bulletin board known as the DSBB, and (2) to establish an Internet site containing the actual data, called a National Summary Data Page (NSDP), which is linked to the DSBB.

The GDDS was established in 1997 to develop sound statistical systems as the basis for dissemination of data to the public (see SM/97/275). The purposes of the GDDS are (1) to encourage member countries to improve data quality, (2) to provide a framework for evaluating needs for data improvement, setting priorities in this respect, and mobilizing technical assistance, and (3) to guide member countries in the dissemination to the public of comprehensive, timely, accessible, and reliable economic, financial, and socio-demographic statistics.

The GDDS is focused on improving statistical systems, whereas the SDDS is focused on dissemination of data used by financial markets. Both are voluntary, but once a country subscribes to the SDDS, observance of the standard is an obligation.

In establishing the SDDS, Executive Directors emphasized that it should be implemented flexibly to adapt to changing circumstances and provided for periodic reviews to make needed adjustments. This approach also has been applied to the GDDS, and both have been strengthened to meet the need for more comprehensive information on international reserves and external debt.

There have been four reviews so far of the Data Standards Initiatives; the conclusions are summarized in Appendix I, along with conclusions from three reviews of Data Provision to the Fund for Surveillance Purposes.

II. The Special Data Dissemination Standard (SDDS)

A. Overall Trends and Developments

7. Since its inception in 1996, the SDDS has had a significant impact on subscribing countries’ data dissemination practices and has served as a stimulus for the IMF’s work in promoting data quality, leading to the establishment of the DQP under the Fourth Review.

8. Initially, subscribing countries were not expected immediately to meet all the SDDS requirements and, therefore, a transition period until the end of 1998 was in place. During this period, subscribers could adopt so-called “transition plans” to upgrade their statistical systems, such as moving from annual to quarterly national accounts statistics. Meeting all the requirements proved a daunting task for many subscribers, notwithstanding the substantial technical assistance effort undertaken by the Fund, in particular to assist countries moving from annual to higher frequency data. The number of outstanding transition plans is an overall indicator of work to be done to meet SDDS requirements. The sharp decline from the high level at end-1998 points to a significant improvement in subscribers’ dissemination practices (see Chart 2). Since the end of the transition period, new subscribers meet all SDDS requirements before subscription so that any improvements made to meet the standard are not captured in the chart. In addition to meeting the initial requirements, SDDS subscribers now disseminate the reserves template, the IIP, and are beginning to disseminate the external debt data category (see Table 1).

Chart 2.
Chart 2.

Number of Outstanding SDDS Transition Plans for Coverage, Periodicity, and Timeliness 1/

Citation: Policy Papers 2003, 026; 10.5089/9781498329347.007.A001

Source: IMF—DSBB1/ Excluding the three new external sector data categories, i.e., the reserves template, the IIP, and the external debt.
Table 1.

SDDS Indicators

article image
Source: IMF

As of end-May 2003.

9. There are currently 53 SDDS subscribers compared with 49 at the time of the Fourth Review. On March 24, 2003, the Republic of Kazakhstan became the fifty-third SDDS subscriber, making it the first GDDS participant to become an SDDS subscriber. For a brief case history of Kazakhstan and other recent SDDS subscribers, including, inter alia, the role of technical assistance and of the data module of the ROSCs in their statistical development, see Box 2.

New SDDS Subscribers Since June 2001

Tunisia on June 20, 2001: Following participation in the SDDS regional seminar in June 1996, Tunisia modernized its statistical system to bring it into line with international standards to help raise its profile in global capital markets. In order to meet SDDS requirements, Tunisia adopted a statistical development plan supported by IMF technical assistance, in particular for compiling quarterly national accounts. Other measures included the adoption of modern statistical legislation in 1999, guaranteeing the transparency of statistics and their harmonization with international methods and concepts; the introduction of annual surveys of households to compile data on unemployment using International Labour Organization guidelines; and the widespread use of the Internet for disseminating SDDS data.

Costa Rica on November 28, 2001: Marked improvements in statistics followed the recommendations of the data ROSC mission in July 2001. In particular, Costa Rica compiled and disseminated quarterly national accounts and, for the first time, began disseminating information on statistical methodologies to the public via the DSBB.

Greece on November 8, 2002: Greece’s SDDS subscription marked a major step forward in the development of the country’s statistical system. The Greek authorities view subscription as the culmination of many initiatives on the statistical front aimed at improving statistical compilation and dissemination practices in Greece as a member of the European Union and the Euro zone. Subscription follows an initiative launched by the Ministry of National Economy in May 2001 to subscribe to the SDDS and an IMF mission in June 2002, which assisted the authorities in formulating a plan of action and a time frame for formal subscription. Also, subscription coincided with Greece’s participation in the preparation of a data ROSC.

Ukraine on January 10, 2003: The first country of the Commonwealth of Independent States to subscribe to the SDDS. Since its independence in 1991, Ukraine has made a significant effort to bring its official statistics into line with international standards. This effort has been supported by technical assistance from the Statistical Office of the European Commission-Technical Assistance to the Commonwealth of Independent States (Eurostat-TACIS), International Labour Organization, IMF, Organisation for Economic Co-operation and Development, World Bank, various United Nations organizations, and development cooperation agencies of a number of countries. Subscription was also the result of the authorities implementing the recommendations of the data ROSC mission in April 2002.

Kazakhstan on March 24, 2003: Kazakhstan has demonstrated a strong commitment to bringing its official statistics into line with international standards. Since independence in 1991, this effort has been supported by technical assistance from the IMF and other agencies, including the Asian Development Bank, World Bank, and development cooperation agencies of a number of countries. The authorities had made substantial progress in implementing technical assistance recommendations, and in 1998 Kazakhstan was a pilot country using the GDDS as a framework for further improvement of the national statistical system. Kazakhstan also undertook a data ROSC in April 2002. The authorities focused their efforts on implementing the mission’s recommendations and as a result, Kazakhstan became the second country of the Commonwealth of Independent States to subscribe to the SDDS and the first GDDS participant to become an SDDS subscriber.

10. Along with improvements in the coverage, periodicity, and timeliness of data disseminated to the public, the SDDS has given rise to increasing certainty and predictability in the release of these data.

11. In the beginning, very few SDDS subscribers published advance release calendars showing three months ahead the dates on which data would be disseminated to the public. By the end of 1998, there were over 500 transition plans for the introduction of advance release calendars. Since that time, the number has declined dramatically, and at present, there are 50 transition plans for the introduction of these calendars.

12. Staff monitoring, initiated in July 2000 following the Third Review, provides measures of the extent to which subscribers meet their commitments to disseminate timely data.5 There has been a rising trend in the compliance rate for both high and low frequency data categories. Compliance rates increased from around 70 percent in late 2000 to over 90 percent by the first quarter of 2003.6 Reports on subscribers’ record in meeting their dissemination dates, by data category, are included in the Quarterly Updates of the SDDS, which are published on the Internet at http://www.imf.org/cgi-shl/create_x.pl?sdds.

B. Update Since the Fourth Review

Experience with Implementing New Data Categories

13. Following the Asian financial crisis, the importance of a country’s dissemination of comprehensive information on its external position was widely recognized. The IMF responded by establishing a transition period for the dissemination of the IIP data category and expanding the SDDS to include the template on International Reserves and Foreign Currency Liquidity (reserves template) and a new external debt data category. The implementation of the reserves template, which carried a transition period of one year, was discussed under the Fourth Review. At that time, most Executive Directors supported the retention of the current prescriptions of monthly periodicity and monthly timeliness. At the time of the Review of Data Provision to the Fund for Surveillance Purposes in May 2002, most Executive Directors considered that increasing the frequency and timeliness for the dissemination of reserves template data under the SDDS was not necessary at that time. The section below focuses on recent developments under the IIP and external debt.

IIP and External Debt

14. The three-year transition period for the dissemination of the IIP ended on December 31, 2001. By end-March 2003, all subscribers were meeting the SDDS requirements for the IIP (see Table 2).7 The great majority of subscribers disseminate more than is required, as follows:

  • 81 percent (43 subscribers) disseminate data according to the detailed instrument breakdowns of the Balance of Payments Manual, fifth edition (BPM5), showing components of direct investments, bonds and notes, money market instruments, trade credits, loans, currency and deposits, and other assets;

  • 83 percent (44 subscribers) disseminate data by economic sector, showing monetary authorities, general government, banks, and other sectors;

  • 68 percent (36 subscribers) provide maturity breakdowns for assets and liabilities associated with trade credits, loans, and other assets and liabilities; and

  • 32 percent (17 subscribers) disseminate IIP statistics with quarterly periodicity, which is encouraged (but not required) by the SDDS.

Moreover, including SDDS subscribers, a total of 83 economies published IIP statistics in the International Financial Statistics (IFS) in May 2003, 8 compared with 37 economies in May 1998.

Table 2.

Implementation of the IIP Data Category

article image
Source: IMF

For any two prescribed data categories (except national accounts, balance of payments, international reserves, and external debt), a subscriber may disseminate data with a lower frequency and/or with a greater lag than prescribed.

In the event that IIP data were based on a fiscal year that was not the same as the calendar year, the first reference period was the fiscal year ending after June 30, 2001, and the dissemination deadline was six months (or nine months) after the end of that reference period (for example, for IIP data based on the fiscal year ending on September 30, 2001, the dissemination deadline was March 31, 2002 or June 30, 2002).

15. Although the first required dissemination of the new quarterly external debt data category is not until end-September 2003, 12 SDDS subscribers, including ten emerging market economies, already have begun disseminating external debt data in advance of the deadline.9 The dissemination of external debt statistics has been facilitated by the Fund’s technical assistance and outreach efforts. In June 2003, the Fund published External Debt Statistics: Guide for Compilers and Users (Debt Guide),10 which was prepared by the Inter-Agency Task Force on Finance Statistics.11 The Debt Guide provides a comprehensive conceptual framework that links the external debt statistics to the IIP. Also, in early 2002, the IMF Statistics Department (STA) completed the series of seven seminars targeted at managers and senior external debt compilers in the SDDS subscribing countries.12 The seminars raised awareness of the importance of compiling external debt data that meets the SDDS requirements. Finally, in May 2002, the first of a new series of IMF regional training seminars directed at mid-level compilers was offered at the Joint Vienna Institute. The second seminar was offered at the Joint Regional Training Center for Latin America in Brazil in May 2003. The seminar program will continue through 2004.

Impact of the SDDS Subscription on Capital Market Access

16. There is increasing evidence that adherence to international transparency standards generally, and the SDDS in particular, has a positive impact on a country’s ability to access international capital markets. The evidence indicates that subscription to the SDDS reduces the cost of international borrowing and contributes positively to a country’s rating by major credit rating agencies. An early empirical study by the Institute for International Finance indicated that SDDS subscription substantially reduced credit spreads for emerging market economies. A recent study entitled Transparency and International Investor Behavior (National Bureau of Economic Research Working Paper No. 9260) determined that international investment funds preferred to hold more assets in more transparent markets, with transparency being gauged by a country’s undertaking of ROSCs, including those on data transparency. Also, a more recent study entitled The Link Between Adherence to International Standards of Good Practice, Foreign Exchange Spreads, and Ratings published in the IMF Working Paper Series (WP/03/74) incorporated a wider set of variables—transparency standards—than the Institute for International Finance’s study and used a larger and higher frequency dataset. The results again confirmed that subscription to the SDDS had a significant impact in reducing borrowing costs.

DSBB Enhancements

17. The IMF’s DSBB (http://dsbb.imf.org/Applications/web/dsbbhome/) provides information about data compilation and dissemination practices of SDDS subscribers and GDDS participants. The DSBB also provides direct links to SDDS subscribers’ National Summary Data Pages (NSDPs), which contain the actual data described on the DSBB. In addition, the DSBB includes the Data Quality Reference Site, which contains articles and other information to promote a common understanding of data quality.

18. On March 10, 2003, STA launched an enhanced DSBB website13 that provides improved access to metadata and the associated datasets on the NSDPs. DSBB users can now access a query facility that provides several “views” of metadata that allow users to quickly find an answer about a particular country’s practices or to compare statistical practices across countries and data categories (see Box 3), as follows:

  • information on elements such as the coverage, periodicity, and timeliness of data categories for either SDDS subscribers or GDDS participants;

  • more detailed information for SDDS subscribers on topics such as the base/weight reference periods for a price index;

  • advance release calendars for any combination of SDDS subscribers and data categories on a quarter-ahead basis; and

  • information on how SDDS subscribers are meeting the requirements of the standard, i.e., use of flexibility options, etc.

19. The enhanced DSBB also provides staff with more efficient and reliable data management (storage/retrieval) functions, and supports improved information processing that will lead to automated SDDS observance monitoring on the web. Future enhancements will include the ability to build on existing queries and consult a library of pre-defined queries.

New Ways to Look at DSBB Content

The enhanced DSBB website permits users to conduct comparative analysis of statistical practices among all 53 SDDS subscribers. For example, a user interested in comparing the institutional coverage (“queried concept”) of the data on central government operations could submit a query and get the information displayed below.

More than 250 analytically useful key concepts may be queried within the SDDS subscribers’ metadata. These key concepts are being further refined and expanded within the context of STA’s collaboration with Statistical Data and Metadata Exchange (SDMX) partners as outlined in paragraph 20.

20. At the Fourth Review, the Executive Board supported efforts to implement an open exchange system for the distribution and exchange of statistical information on the Internet. Fund staff, under the Statistical Data and Metadata Exchange (SDMX) initiative, is developing a set of tools to extend the functionality of the DSBB to a broader range of statistical information.14 The staff is working to develop a broad-based model for statistical metadata in conjunction with work under way by other SDMX partners to develop a common vocabulary for metadata terms. This open exchange model would enable users of the DSBB and/or other websites to search and query statistical information across multiple websites.

21. In line with the guidance of the Fourth Review, STA is conducting an Internet-based users’ survey to assess the impact of recent enhancements on the usefulness of the DSBB. The short survey is targeted at those who have had experience with the website both before and after the introduction of the enhancements on March 10, 2003. The survey was initiated on May 30 to allow sufficient time for users to become fully familiar with the new features of the website. A comprehensive report on survey findings will be presented in an upcoming Quarterly Report on the SDDS published on the Fund external website.

Expanding SDDS Subscription: The 25/50 Program

22. Under the Fourth Review, the Executive Board underscored the desirability of expanding SDDS subscription to promote greater access to international capital markets, noting that the Fund is encouraging members to prepare to be able to access capital markets. In response, the staff has implemented an aggressive SDDS outreach and technical assistance effort called the “25/50 program.” This program identifies countries considered capable of meeting the SDDS within the next two to three years— approximately 25 countries—and a larger group of about 50 countries that could meet the requirements within about five years. The staff works intensively with as many countries as resources permit within the group of 25 and, as countries move on to subscription—e.g., Ukraine, Greece, and Kazakhstan—the smaller group is replenished from the larger pool.

23. The program is conducted, in part, through outreach seminars among countries and, in part, through close collaboration of staff at headquarters with designated SDDS coordinators. These efforts are closely integrated with STA’s technical assistance program.

24. The first seminar was held in Athens in October 2002 and was hosted by the Bank of Greece; the second seminar was held in Mexico City in April 2003 and hosted by the Center for Latin American Monetary Studies. A future seminar is planned for Africa. These outreach seminars help countries to assess their data dissemination practices against the SDDS and provide guidance on how countries can meet the standard, including on the use of flexibility options. The seminar also provides information on the day-to-day operations of the SDDS including the role of the country coordinator and the system of staff monitoring of compliance with scheduled release dates.

C. Proposed Modifications to the SDDS

25. As noted by the Executive Board at the inception of the SDDS, the standard must evolve to meet new challenges and changing circumstances. In this light, the staff has kept under review a number of issues for which adjustments, modifications, or additions to the SDDS might be desirable. These issues are discussed below.

Fiscal Data

26. At the Third Review of the Fund’s Data Standards Initiatives in March 2000, Executive Directors stressed that the SDDS should continue to encourage the adoption of data quality improvements and agreed with the proposal that, for SDDS subscribers implementing accrual-based reporting systems, the periodicity and timeliness of fiscal data be on a “best efforts” basis until end-June 2002 (see BUFF/00/43). In the run-up to that date, staff consultation with several Fund members showed that the number of countries implementing accrual reporting was still limited, and that existing experience did not provide an adequate basis for making specific recommendations for modifications to the SDDS at that time. The staff then proposed the extension of these special arrangements until the Fifth Review.

27. The current SDDS prescriptions for the timeliness (monthly) of data on central government operations (CGO) allow these data to serve as a timely lead indicator of more comprehensive fiscal developments captured under general government operations (GGO). However, the timeliness prescription for CGO data serves as an impediment to data quality improvements as subscribers move towards the compilation of government finance statistics according to international best practice as set forth in the Government Finance Statistics Manual 2001 (GFSM 2001) or equivalent guidelines. Specifically, the SDDS requirement that monthly CGO data be disseminated with a one-month lag creates strains at the turn of the fiscal year when the resource pressures associated with the compilation of accrual-based data are most intense.

28. For fiscal policy formulation and analysis, data on GGO are preferable to data on CGO because they are more comprehensive. However, the SDDS prescribes GGO data on an annual basis with six-months timeliness, a prescription that does not support timely analysis. The compilation and dissemination of quarterly GGO data, following GFSM 2001 or equivalent guidelines, therefore would be desirable. Several SDDS subscribers already have committed to the compilation and dissemination of these data on a quarterly basis, and it is anticipated that quarterly GGO data will increasingly become standard practice.

29. In formulating a forward-looking recommendation for SDDS prescriptions for fiscal data, the staff seeks to promote data quality improvements in the form of more timely accrual-based comprehensive fiscal data. Therefore, if a subscriber disseminates, with a one-quarter lag, quarterly GGO data in line with GFSM 2001 or an equivalent standard, the staff proposes that a targeted timeliness flexibility option be available for monthly CGO data. This targeted flexibility option would be allowed for the last month of the fiscal year (up to three months lag) and the first month of the new fiscal year (up to two months lag), the period when resource pressures arising from the compilation of accrual data are most intense. In order to make use of this flexibility option, a subscriber would need to begin disseminating quarterly GGO data for at least the last quarter of the fiscal year in which the option is exercised. (For example, when the fiscal year is the same as the calendar year, to use the option for monthly CGO data for December 2003 a subscriber would need to disseminate quarterly GGO data for the last quarter of 2003.) The staff recognizes the costs involved in the dissemination of less timely CGO data for two months of the year, but believes the benefits gained in promoting the dissemination of higher quality data outweigh the costs.

30. This proposal replaces the special transitional arrangements for subscribers implementing accrual-based reporting systems for fiscal data approved by the Executive Board under the Third Review in March 2000.

Financial Soundness Indicators (FSIs)

31. The inclusion of FSIs in the SDDS would complement the macroeconomic data now covered by the standard, given the underlying role of the SDDS as one of the IMF’s core standards to reduce economic vulnerability. At the Executive Board discussion on FSIs on June 2, 2003, Executive Directors reaffirmed the importance of ongoing staff work on FSIs and endorsed staff proposals for a coordinated exercise involving about 60 members for the compilation of FSIs over the next few years. The Executive Board did not, at this stage, agree on a target date for the inclusion of the core FSIs, or a subset of the core, into the SDDS. The Executive Board will return to this issue in future reviews of the FSIs, following completion of the Compilation Guide on Financial Soundness Indicators, and as work on the compilation exercise proceeds and further experience is gained with making FSIs operational.

Inflation Targeting Indicators

32. A two-day seminar in 2002, hosted by STA, addressed the role of the IMF’s work on data standards in supporting the statistical aspects of an inflation targeting regime.15 Two IMF initiatives—the SDDS and the data module of the ROSC—were highlighted as particularly relevant for inflation targeting countries. Countries that agree to undertake the data module of the ROSC and publish the report send clear signals about their openness and their commitment to addressing any statistical issues, and thereby foster better public understanding of the data used for inflation forecasting and policymaking.16 Moreover, subscription to the SDDS carries with it a commitment to transparency in the compilation and dissemination of macroeconomic statistics, identified as a key element in supporting the credibility of an inflation-targeting regime.

33. While recognizing that the SDDS is a general framework applicable to countries regardless of policy regime, the seminar considered whether the SDDS could be adapted to accommodate any special data needs of countries with inflation-targeting regimes.17 Information provided by SDDS subscribers who have inflation targeting regimes could be enhanced, in particular by explaining how such countries derive their measures of “core” inflation that determine the policy target, and their use of forward-looking indicators and interest rates as operating targets.18 Moreover, similar information would be useful for other SDDS subscribers who do not officially have inflation-targeting regimes. Therefore, to make the DSBB more relevant, the staff proposes that all SDDS subscribers be encouraged to provide additional information on topics such as their practice with respect to measuring core inflation, forward-looking indicators, and interest rates used as operating targets for posting on the website.

III. The General Data Dissemination System (GDDS)

A. Impact of the GDDS

34. Since the Fourth Review, the number of GDDS participants has increased steadily from 30 in July 2001 to 58 at present. The staff expects that about eleven additional countries will become participants by the end of 2003.19

35. The usefulness of the GDDS in identifying countries’ statistical strengths and weaknesses and in directing their efforts to develop their statistical systems has become increasingly recognized. The GDDS has proven valuable in helping countries to harmonize efforts across agencies, to improve data transparency, and to mobilize technical assistance resources. The GDDS serves as a catalyst in the mobilization and coordination of technical assistance by the Fund, the World Bank, and other donors. In addition to technical assistance provided by the Fund for economic and financial data, the World Bank provides technical assistance on socio-demographic indicators. Box 4 provides a sample of brief case histories of countries that have used the GDDS successfully as a framework for statistical development.

GDDS—Making a Difference

The IMF’s promotion of statistical capacity-building in member countries has undergone substantial change in the last decade. One component of this change, the GDDS, provides a framework for countries to set objectives and to map a path to achieve them. Moreover, the GDDS provides a mechanism for the mobilization of technical assistance resources—both from the IMF and other donors—by disseminating countries’ plans for improvement together with a statement of the resources required for their implementation.

There are now 58 GDDS participants. Many had received technical assistance before participating in the GDDS. However, the decision to become a GDDS participant requires a public commitment by each country to use the GDDS as a framework for statistical development and a public statement, via the IMF’s website, of the path to be taken to develop the statistical system. The IMF, in managing the GDDS program, developed a flexible program of technical assistance to aid the process, as can be seen in the following examples.

China became a GDDS participant in April 2002. This milestone was welcomed in the national press as a “major step forward” in the direction of a transparent statistical system. In addition to publishing metadata on the agencies’ own websites, China also produced a book of GDDS metadata for widespread reference. China has made progress over a number of years in strengthening its statistics, and the IMF continues to work closely with Chinese agencies, in collaboration with other providers of technical assistance in statistics.

Namibia, with 13 other African countries in the Anglophone Africa Project, participates in the GDDS. 1/ Namibia hosted a workshop in February 2002, where IMF staff and experts from all project countries participated. Plans for improvement were drafted and technical assistance requirements identified. The workshop was followed by a program of technical assistance for participating countries, to assist in the implementation of those plans. Namibia received assistance in all four macroeconomic sectors. By December 2002, Namibia’s metadata and plans for improvement were posted on the DSBB. At a recent national workshop (under the Anglophone Africa project), disseminating agencies shared these plans with high-profile data users from government and the private sector and with donors. This coordinated approach bodes well for significant improvements in Namibia’s national statistical system.

Fiji was an inaugural GDDS participant in 2000, and its initial metadata served as a model for other Pacific Island countries. The Pacific Financial Technical Assistance Center (PFTAC) has devised an assessment instrument for countries to develop work plans based on the GDDS. These countries vary widely in their current institutional capacity and their ability to implement plans for improvement. Nevertheless, the GDDS has proved sufficiently flexible in providing a framework for development and for prioritizing technical assistance. The current GDDS project at the Pacific Financial Technical Assistance Center has reached the stage where a number of members are expected to become participants in the GDDS.

The GDDS may also be viewed as a stepping stone towards subscription to the more demanding SDDS. Technical assistance provided to GDDS countries can lay the foundation. At the opening of the African Technical Assistance Center (East) in October 2002, both Uganda and Tanzania, current GDDS participants, indicated that they saw SDDS subscription as medium-term objectives.

1/ This project is managed by the IMF, with funding by the United Kingdom Department for International Development. See paragraph 35.

36. A regional approach has leveraged the Fund’s scarce technical assistance resources to assist members who wish to participate in the GDDS. These regional projects, with financial support mainly from Japan and the United Kingdom Department for International Development, assist country representatives in understanding the GDDS and preparing plans to improve their statistical systems. These initial steps are followed up by technical assistance—a resident technical advisor and a group of short-term experts—to assist countries to implement their plans for improvement. Since the Fourth Review, GDDS projects have been launched for member countries of the West African Economic and Monetary Union (WAEMU), for Anglophone African countries, Lusophone African countries, and for Pacific Island countries and territories. Additional projects are being planned for Central American and Middle Eastern countries.

37. The statistical work at the regional technical assistance centers20 also focus on GDDS. In some cases, the regional technical centers play the coordinating role in the regional GDDS project (see Box 4).

B. Millennium Development Goals (MDGs) in the GDDS

38. From the outset, the GDDS was conceived to cover socio-demographic data as well as economic and financial data categories. In light of the growing recognition that poverty reduction strategies require a substantial amount of data to establish baselines, provide a foundation for policymaking, and monitor outcomes, the Executive Board under the Fourth Review endorsed a fully articulated socio-demographic component of the GDDS. Since then, statistical capacity building has found entry into many countries’ poverty reduction strategies, as documented in their Poverty Reduction Strategy Papers (PRSPs).

39. The socio-demographic component of the GDDS does not currently make explicit reference to the MDGs and associated indicators. The now well-established importance of the MDGs, and the role the GDDS can play in promoting appropriate monitoring systems, require that this be remedied.

40. There is already substantial coverage by the GDDS of the basic data components needed to compile MDG indicators. A Supplementary paper prepared jointly by World Bank and IMF staff sets out the context and the relationship between the GDDS and the body of statistical indicators that countries are committed to support in monitoring progress towards meeting the MDGs.

41. The joint paper notes possible amendments to the GDDS to improve the alignment of the GDDS with MDG indicators and to provide guidance to countries in building the statistical platform needed for this purpose. It also encourages participants in the compilation of MDG indicators, and in the incorporation in GDDS metadata of information that assists data users in understanding the methodologies and other practices used in the process.

C. Future Directions in Managing the GDDS

42. The expanding number of participants in the GDDS and continued interest by nonparticipating countries strain the staff and financial resources available to manage the system. In order to be able to attract and serve new GDDS participants, the staff intends to concentrate its work with participating countries on those that make a strong effort to improve their statistical systems. Countries exerting a strong effort to implement their improvement plans should find most of their technical assistance requirements in key areas met with little delay. Other countries, progressing more slowly, would continue to receive significant technical assistance from the Fund, guided by their highest-priority plans for improvement. Meanwhile, countries that show little effort in the GDDS would receive a reduced share of resources, but staff would remain engaged. These measures are in line with the priorities already set out by the Executive Board on its policy discussions on technical assistance.

43. In working closely with the most committed participants, the staff intends where appropriate to advise countries that have met the recommendations of the GDDS to turn to the SDDS as a more demanding standard for data dissemination while continuing to improve other aspects of their statistics.

44. Internet-based dissemination of statistical data is growing year by year, and is attractive to the broad spectrum of data users. GDDS metadata provide an opportunity to alert data users to the growing number of websites of national disseminating agencies. At present, some GDDS metadata pages for specific data categories on the DSBB already provide links to many country websites currently providing statistics. As countries add this facility to their dissemination strategies, their metadata will be updated with an active link to those websites.

IV. The Data Quality Program (DQP)

45. The DQP is built around the SDDS, the GDDS, the Data Quality Assessment Framework (DQAF), and the applications of all these in the ROSC. The implementation of the DQP has evolved to encompass efforts to promote the adoption of good statistical practices by national authorities. The DQAF is at the heart of this program by providing a methodology that covers every aspect of the data production cycle. The DQAF, therefore, addresses issues under five dimensions of quality—namely assurances of integrity, methodological soundness, accuracy and reliability, serviceability, and accessibility—as well as the pre-requisites of quality.

46. The key role of the DQAF in the program underscores the importance of keeping that methodology up-to-date. Since the Fourth Review, the DQAF has been further sharpened, in large part from the experience gained from the DQAF-based data modules of the ROSC. The fine-tuning of the DQAF since the Fourth Review is presented in the accompanying background paper.

47. This section of the paper reviews selected components of the DQP. The focus is first on the standards-related initiatives where the key strands that link the SDDS, GDDS and the DQAF are pulled together. This is followed by an overview of lessons learned from the ROSC as a major DQAF application. Other DQAF-based initiatives are presented to highlight some of the work carried out under the DQP.

A. The Data Standards and the DQAF

48. The primary focus of the SDDS is on the dissemination of statistics, but it recognizes the multidimensionality of quality in statistics by covering integrity, accessibility, and other aspects of quality. The GDDS uses the same multidimensional approach to quality as the SDDS, but emphasizes improving the compilation process to provide for the dissemination of both economic and socio-demographic data.

49. The DQAF evolved from the approach taken in the SDDS and GDDS. Three of the four SDDS/GDDS elements, i.e., the data (or coverage, periodicity, and timeliness), access, and integrity correspond directly with the DQAF. The SDDS/GDDS quality element corresponds to three dimensions in the DQAF, namely the prerequisites of quality, methodological soundness, and accuracy and reliability. Given the rising profile of data standards and data quality more generally, some consolidation of the data standards and the DQAF may be desirable to make the coherence of the approach more transparent. However, developing the modalities for such a consolidation would require further consultation within the Fund and with subscribers/participants of the SDDS/GDDS as well as data users. The staff will keep the Executive Board informed of progress in this area.

B. ROSCs

50. Since the Fourth Review, STA has undertaken about 15 ROSCs per year using the DQAF and completed 23 by June 3, 2003. The results of these ROSCs are summarized in Box 5 of this paper (see also Supplement 3 of SM/03/86 on Review of the Fund’s Experience with the Data Module ROSCs). Present plans call for STA to complete data module ROSCs for almost 80 percent of SDDS subscribers—including the Group of Eight countries, which agreed to set an example for other countries—by the end of FY 2005.21 In addition, for SDDS subscribers/GDDS participants, direct electronic links will be established from their metadata describing data dissemination practices and posted on the DSBB and the data ROSC on the Fund’s external website, which assesses those practices.

51. Since the inclusion of the DQAF, ROSCs have been at the forefront of the DQP. The ROSC assesses not only data dissemination practices against the SDDS/GDDS, but also the full range of statistical practices governing the production of specific macroeconomic datasets. As such, it provides valuable information to market participants on how national statistical practices compare to internationally accepted practices. In addition, the prioritized recommendations in these reports provide a clear link with the IMF’s technical assistance. Publication of ROSCs is voluntary; all finalized data modules using the DQAF have been published.

52. At the time of the Executive Board’s discussion on the international standards on March 19, 2003, Executive Directors agreed to adjust the intensity and frequency of follow-up work to keep ROSCs current. Resource constraints mandate that a careful prioritization of staff resources be established to ensure timely updates within the overall resource envelope earmarked for standards work. The staff, therefore, proposes the following program of updates:

  • All data ROSCs will continue to have factual updates within the context of the Article IV consultation discussions. For about six Article IV discussions each year, STA will conduct more substantive updates by assigning a staff member to the area department team.

  • Within the overall resource envelope of doing 15 data ROSCs per year, STA will progressively switch part of its resources from “first” ROSCs to follow-up ROSCs. In this connection, as of FY 2005, STA will begin to invite those countries that have undertaken a ROSC before the incorporation of the DQAF to undertake a DQAF-based ROSC. 22

Experience with the Data Module of the ROSC 1/

Data modules of the ROSC are promoting improvements in the quality of macroeconomic data. As of June 3, 2003, data ROSCs with assessments based on the DQAF for 23 countries had been published and 12 are in progress. Reports include prioritized recommendations intended to guide country efforts and technical assistance.

The main findings of the data modules are as follows:

  • Countries with robust legal and institutional frameworks for statistical production performed generally better in terms of overall data quality.

  • Although most countries adhered broadly to internationally accepted methodologies, shortcomings were most evident in coverage, classification, and sectorization.

  • Production of comprehensive source data was a major challenge in most countries.

  • Fiscal statistics were the most seriously affected by weaknesses, often owing to incomplete coverage of operations beyond the budgetary central government.

  • National statistical institutes tended to focus most on dissemination to the public. Both central banks and finance ministries maintained a close link between statistical production and policy formulation. Finance ministries tended to focus on reporting to the executive and legislative branches in the context of budget preparation.

  • Revision studies were undertaken routinely only in a few countries. This finding has spurred staff work to identify best practices.

  • Even in most advanced statistical systems, access to data and metadata could be improved.

1/ Based largely on “The Fund’s Experience with Data Module ROSCs” (SM/03/86 Supp. 3).

C. Promotion of Good Statistical Practices

53. The DQP aims not only to assess data quality, but also to identify good statistical practices at the national level and promote their widespread adoption. The data ROSCs so far have drawn attention to areas in which countries follow good statistical practices and other areas in which these practices are not well-understood. Concerning the latter, revision policies are of particular note. Although knowledge of procedures governing revisions can be critically important to data users, statistical agencies often are less than transparent in providing such information. Thus, during the May 2002 Executive Board discussion on Data Provision to the Fund for Surveillance, Executive Directors encouraged national authorities to articulate their policies on data revisions, which would enhance the transparency of the data provided to the Fund. STA staff has under way a paper that, drawing on country practices (including those identified in data ROSCs), proposes a set of good practices. STA will invite international comment, including at the September 2003 meeting of the heads of statistical units of international organizations, as a step toward an internationally agreed set of practices to guide countries.

54. The staff proposes that good statistical practices be brought together in a “Compendium of Good Statistical Practices,” along the same line as the Manual on Fiscal Transparency. This compendium would draw extensively from the ROSC experience and would serve as a guide to countries by providing concrete examples where good statistical practices were successfully implemented.

D. Collaboration With Other Organizations

55. Recognizing that a common view on data quality among all interested parties is important for the success of quality initiatives, the staff has actively collaborated with other international agencies. A major part of this work to date consists in extending DQAF applications such as for statistical capacity building.

56. A DQAF application has been developed for statistical capacity-building under the aegis of Partnership in Statistics for Development in the 21st Century (PARIS21)23 that resulted in the PARIS21 Statistical Capacity Building Indicators. This work was chaired by the Fund and involved the World Bank, the United Nations, United Nations Economic Commission for Europe (UNECE), United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP), Direction générale de l’Observatoire économique et statistique d’Afrique subsaharienne (AFRISTAT), and close consultations with developing countries. By focusing on selected aspects of the statistical production, the statistical capacity building indicators provide for a diagnostic self-assessment or a peer review of statistical capacity.

V. Resource Issues

57. The work program associated with all the components of the DQP (SDDS, GDDS, DQAF, and ROSCs) has entailed a substantial and, particularly in relation to ROSCs, rising commitment of resources. The Board-mandated objective of conducting around 15 data ROSCs during each of the past two fiscal years was fulfilled despite the significantly greater resource intensity of ROSC work than initially expected. In line with the Board decision in the context of the Fourth Review, the DQAF is now fully integrated into the ROSC process. In order to generate the resources needed to conduct this work, STA has relied on a rigorous prioritization process and achieved efficiency gains in other areas of the work program. Although data ROSC missions will continue to be resource-intensive, considerable savings have been realized by combining the ROSC data modules for all four macroeconomic sectors—usually involving six separate dataset-specific assessment frameworks. This practice maximizes efficiency and helps to ensure that the assessment of statistical practices is as consistent and internationally comparable as possible. The consolidation of resources into one integrated data ROSC mission means that resource costs of each mission are higher compared with ROSCs that focus on a single topic and, therefore, cost comparisons can be misleading.

58. Resource pressures are likely to remain at a high level for several reasons. First, the expected increase in SDDS subscription and GDDS participation will generate significant additional operational work. Furthermore, the plans for improvement that accompany GDDS participation will require an increase in technical assistance for their implementation. The regional GDDS projects will require additional backstopping support and increased resources devoted to technical assistance, particularly in Africa. Second, the rapid increase in the number of countries that have completed a data module ROSC also will lead to a rise in demand for technical assistance. Third, updating ROSCs for an increasing number of countries that have participated in the ROSC process will require additional resources if such reports are to remain useful to the Fund and the financial markets.

59. In order to meet the increased demands on staff resources cited above, STA will monitor the situation carefully, further prioritizing services, including through rescheduling some of its mission activities. In addition, STA will endeavor to realize even more efficiency gains. Our plans will not require additional resources either for the current year or the medium-term estimates.

VI. Issues for Discussion

A. Concerning the Overall Strategy for Promoting Data Transparency

60. The paper lays out an overall strategy for promoting data transparency by increasing Fund members’ subscription to/participation in the Data Standards Initiatives, aligning the SDDS and the GDDS with international best practice and maintaining their relevance, and integrating the Data Standards Initiatives (the SDDS, the GDDS, the DQAF, and the ROSC) in the DQP to sharpen the focus on both quality assessments and the promotion of good statistical practices.

Do Executive Directors agree with the overall strategy for promoting data transparency laid out in the paper?

B. Concerning the Modification of the SDDS for Fiscal Data

Under the SDDS, monthly CGO data—with monthly timeliness—would continue to be used as lead indicators for fiscal developments. In order to encourage data quality improvements, the staff proposes that a targeted timeliness flexibility option be introduced in the SDDS for monthly CGO data if subscribers disseminate quarterly GGO data according to GFSM 2001 or an equivalent guideline with a one-quarter lag. The targeted flexibility option would be allowed for the last month of the fiscal year (up to three months) and the first month of the new fiscal year (up to two months). This proposal replaces the special transitional arrangements for subscribers implementing accrual-based reporting systems for fiscal data approved by the Executive Board under the Third Review in March 2000.

Do Executive Directors agree with the proposed flexibility for the timeliness of CGO data for the last month of the fiscal year (up to three months) and the first month of the new fiscal year (up to two months) if subscribers disseminate quarterly GGO data according to GFSM 2001 or an equivalent guideline with a one-quarter lag?

C. Concerning Inflation Targeting Indicators

61. While recognizing that the SDDS is a general framework applicable to countries regardless of policy regime, SDDS metadata could be augmented on a voluntary basis to accommodate subscribers with inflation-targeting regimes and other subscribers for whom information about their practice with respect to measuring core inflation, forward-looking indicators, and interest rates used as operating targets would be useful. Such additional information would enhance the relevance of the DSBB. The staff would welcome directors’ views on the following:

Do Executive Directors agree that subscribers should be encouraged to provide additional metadata for the DSBB to promote public knowledge and understanding of their practice with respect to measuring core inflation, forward-looking indicators, and interest rates used as operating targets?

D. Concerning the Enhancement of the GDDS to Promote the Availability of Statistical Data to Compile MDG Indicators

62. The GDDS has a role to play in encouraging countries to build a statistical base that allows for more complete compilation of the MDG indicators. Staff proposes therefore that the current suite of basic components and encouraged extensions with relevance to MDGs be revised and expanded where appropriate to support compilation of MDG indicators, and that the GDDS be amended accordingly.

Do Executive Directors agree with the proposed approach to giving more explicit recognition to MDG indicators in the GDDS?

E. Concerning Prioritization for ROSC Updates and Undertaking New ROSCs

63. Careful prioritization of IMF staff resources will be needed to ensure that the information contained in data ROSCs remains as timely as possible. The staff proposes, therefore, that all data ROSCs have factual updates in the context of Article IV Consultation discussions; for about six Article IV discussions per year, an STA staff will be available to join the Article IV team to conduct a more substantive update. Over time, within the overall constraint of undertaking fifteen data ROSCs per year, STA will move to balancing “first” ROSCs for countries with undertaking follow-up ROSCs. In this connection, the staff proposes that, as of FY 2005, those countries for which a ROSC was undertaken before the incorporation of the DQAF be invited to undertake a DQAF-based ROSC.

Do Executive Directors agree with the above prioritization of staff resources for updating existing ROSCs and undertaking new ROSCs, including follow-up ROSCs for countries whose initial ROSC did not include the DQAF?

F. Concerning a Compendium on Good Statistical Practices

64. The ROSC experience has provided examples of good statistical practices in a number of areas. The staff believes that such information would prove highly valuable to all members as they set about to improve their statistical systems. The staff proposes that good statistical practices, including concrete examples of how countries implemented such practices, be brought together in a Compendium of Good Statistical Practice.

Do Executive Directors endorse the staff’s efforts to develop a Compendium of Good Statistical Practice?

G. Concerning the Timing of the Next Review of the Fund’s Data Standards Initiatives

65. The staff believes that the next review of the Data Standards Initiatives should take place in about two years, in the second half of 2005. The staff would continue to report to the Board, as needed, on critical observance issues and, as necessary, the Board could address any further strengthening that may be necessary owing to unforeseen circumstances.

Do Executive Directors agree with the proposed timing of the second half of 2005 for the next review of the Fund’s Data Standards Initiatives?

66. Based on the outcome of Executive Directors’ discussion of the above issues, amendments to the SDDS Annex and GDDS Document, if required, will be prepared and circulated to the Board for approval.

Fifth Review of the Fund's Data Standards Initiatives
Author: International Monetary Fund