Chapter

3 The General Data Dissemination System: What Has Been Accomplished After 10 Years and Where Do We Go from Here?

Author(s):
William Alexander, John Cady, and Jesus Gonzalez-Garcia
Published Date:
March 2008
Share
  • ShareShare
Show Summary Details
Author(s)

Although the name of the General Data Dissemination System (GDDS) infers that its central focus is dissemination, in its initial stages the GDDS emphasized the development of national systems in an explicit medium-term framework. Attention to data dissemination came only at a later stage. Indeed, participating countries are not required to make any formal commitments regarding data dissemination. The main premise underlying the GDDS is to give high priority to improvements in data quality, which may need to precede improvement in dissemination practice.

IMF member countries have enthusiastically adopted the GDDS: to date, 95 countries have participated in it, including six that have progressed to the SDDS. As of November 2007, approximately 83 percent of the IMF’s membership participated in the Data Dissemination Initiative—the GDDS and the Special Data Dissemination Standard (SDDS). Judging by this metric, the data initiative has been highly effective. Financial market participants also clearly appreciate the merits of the initiative, as confirmed by evidence that countries can reduce borrowing costs significantly by subscribing to or participating in either the SDDS or GDDS. This chapter shows that countries that have participated in the GDDS have met many of its developmental objectives, resulting in a material improvement in the comprehensiveness and quality of their statistical systems.

While these achievements are significant, in some other respects the impact of the GDDS has been more modest. For instance, only six participants have progressed from GDDS to SDDS status, and participants often have lagged behind the timetables they established for meeting particular developmental objectives. Moreover, after 10 years of experience with the GDDS, a marked improvement in data dissemination could have been expected. But as this chapter will show, data dissemination remains weak, particularly in terms of the periodicity and timeliness of data.

More emphasis on putting data into the public domain might well have helped countries progress more rapidly. Earlier experiences with the SDDS show that supply creates its own demand. By publishing data, even with some flaws, statistical agencies benefit from the input of users, including other government agencies. In addition, user input constitutes an important vehicle for data quality improvements. A strong case exists, therefore, for adjusting the GDDS to place substantially more emphasis on data dissemination. This could be achieved in part by importing key dissemination elements of the SDDS and by bringing the data dimension of the GDDS into closer conformity with that of the SDDS. In an important sense, the SDDS would become a special case of the GDDS.

Going forward, this chapter proposes recasting the GDDS by incorporating elements of the SDDS, especially the national summary data page and the advance release calendar. Also, given that several GDDS countries are now borrowing in international capital markets and are subject to sovereign ratings, a revamped GDDS would incorporate the relevant data categories specifically developed to better serve capital market needs, such as the international reserves and foreign currency liquidity template. At the same time, the data dimension of the GDDS would be simplified.

The reformed GDDS would thus be more truly a general case of the SDDS. It would include a larger number of data categories, owing to sociodemographic data categories, with recommended ranges of timeliness and periodicity rather than the prescriptiveness of the SDDS. While participating countries would have more options and guidance on how to move to SDDS subscription, they would still choose their own pace of development.

This chapter provides background to the GDDS, describes its current membership in terms of regional patterns of participation and GDP per capita, and addresses GDDS performance to date. The chapter analyzes GDDS data and metadata compilation, discussing methodologies for compiling the GDDS, countries’ plans for improvement, and sociodemographic data. It then examines GDDS data dissemination, reviewing participants’ statistical practices against the GDDS, comparing data dissemination practices of GDDS participants with dissemination goals set out by the GDDS and SDDS, and highlighting key areas of weakness as perceived by GDDS countries.2 The capital market access of GDDS countries is discussed along with the extent to which the GDDS has been successful in guiding countries to progress to the SDDS.

Background and Membership

The IMF initiated its work on the Data Dissemination Initiative in 1995 in the aftermath of the 1994–95 international financial crisis. The IMF Executive Board approved the SDDS in March 1996 and the GDDS in December 1997. The intention was to establish a basis to guide members in disseminating their economic and financial data to the public. The initiative comprises two tiers: the GDDS, open to all IMF member countries, and the SDDS, which applies to those member countries having or seeking access to international capital markets. The ultimate objective of the two tiers of the Data Dissemination Initiative was to enhance the availability of timely and comprehensive statistics, thereby contributing to the formulation and conduct of sound macroeconomic polices, as well as the improved functioning of financial markets. Countries elect to join the initiative on a voluntary basis.3 They can participate in one of the initiatives but not in both.

The IMF designed the GDDS as a general framework to guide countries in developing sound statistical systems as the basis for dissemination of data to the public. Participation requires that countries appoint a national coordinator, prepare metadata, describe their current practices on data production and dissemination, develop plans for improvement in the short and medium term, and identify associated needs for assistance in implementing these plans. Participating countries also voluntarily commit to revising their metadata at least annually to accurately reflect their data compilation and dissemination activities. The GDDS contains a data dimension identifying periodicity and timeliness goals for key datasets (to a degree paralleling the data dissemination standards in the SDDS). Also, an overarching goal of the GDDS is to focus on developing and disseminating a full range of economic and financial data. However, there is no requirement for GDDS participants to actually disseminate data; nor does the IMF monitor participants’ data dissemination practices (as in the case of the SDDS).

The design and implementation of the GDDS has benefited from close collaboration with member countries and other international organizations, notably the World Bank with regard to sociodemographic data. The GDDS has been implemented in two phases. The first phase focused on education and training through regional seminars for country officials and preparation of pilot metadata for several countries. The second phase started in May 2000 when the first metadata for countries participating in the GDDS were posted on the IMF’s Dissemination Standards Bulletin Board (DSBB).

Participation

The first visible success of the initiative was the rapid increase in participation. As shown in Table 3.1, 39 participants joined the GDDS during its first two years, followed by a steady expansion in the following three years. As of November 2007, the GDDS had 95 participants, six of which subscribed to the SDDS, resulting in 89 current GDDS participants.

Table 3.1.New General Data Dissemination System Participants by Year and Region, 2000–07(Numbers of countries)
Regions120002001200220032004200520062007Total
African Department57108621039
Asia & Pacific Department3221301113
Middle East & Central
Asia Department3413213017
European Department310110006
Western Hemisphere
Department8302322020
Total221713151557195
Graduated to Special
Data Dissemination
Standard311106
Cumulative total (net)2239526478828889
Source: IMF Statistics Department.

Regional classification follows the structure of IMF area departments.

Source: IMF Statistics Department.

Regional classification follows the structure of IMF area departments.

Figure 3.1 shows the combined membership in the GDDS and SDDS and how close each region is to universal membership. Overall, in November 2007, participation was approximately 83 percent of the IMF membership, but participation is not evenly distributed across regions. Africa, Europe, and the Western Hemisphere have achieved about 90 percent representation in the two initiatives. The representation of African countries increased sharply during 2000–04, as compared with other regions, but only one country (South Africa) subscribes to the SDDS.

Figure 3.1.Regional Representation in the GDDS and SDDS Combined

(Percentage of IMF member countries)

Source: IMF Statistics Department.

Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard. AFR = African Department; APD = Asia & Pacific Department; MCD = Middle East & Central Asia Department; EUR = European Department; WHD = Western Hemisphere Department.

About 65 percent of countries in the Asia and Pacific region participate in the data standards initiative (the participation rate is 82 percent, excluding seven small island economies not participating). The Middle East and Central Asia is represented by only 63 percent of its countries, making it the least represented region. Within this region, participation is lowest among Middle East countries, while Central Asian countries are well represented.

As already noted, participation in the GDDS requires appointing a national coordinator responsible for coordinating metadata and plans for improvement among statistical agencies and for communicating with the IMF. GDDS countries often assign this task to senior-level managers from the national statistics office or from central banks (see Tables 3.A3 and 3.A4 in Appendix 3.2). Senior-level managers at the rank of head or deputy head of an agency account for 56 percent of total coordinators, and coordinators from the national statistics offices account for 52 percent of total GDDS coordinators. Thus, the GDDS is given a relatively high priority and visibility in participating countries—indicative of the importance attached to the Data Dissemination Initiative.

Per Capita Income of GDDS Participants and SDDS Subscribers

The choice between GDDS participation versus SDDS subscription may reflect, in part, countries’ relative income levels. Clearly, SDDS subscription requires more resources and an ongoing commitment to disseminate data to the public, while GDDS participation carries relatively few day-to-day responsibilities, the data dimension is less demanding and less comprehensive than in the SDDS, and data dissemination is neither required nor monitored as in the case of the SDDS.

Because national income places some practical limits on the countries’ ability to spend more resources on improving the coverage, periodicity, and timeliness of their data, GDP per capita may be used as the proxy indicator of resources that can reasonably be expected to be available for statistical capacity-building. Table 3.2 illustrates this point. Generally, SDDS subscribers have higher per capita incomes than GDDS participants, with a dividing line at a level of GDP per capita of about $2,000 to $5,000. Some 36 percent of SDDS subscribers, in contrast to about 82 percent of GDDS countries, have a GDP per capita of less than $5,000.

Table 3.2.GDDS Participation and SDDS Subscription and GDP Per Capita
GDDSSDDS
GDP Per CapitaTotal number

(cumulative

percentage)
Total number

(cumulative

percentage)
Total
All CountriesUp to $50026(30)026
$501–$2,00028(61)8(12)36
$2,001–$5,00018(82)15(34)33
$5,001–$10,0006(88)11(53)17
$10,001 +10(100)30(100)40
African DepartmentUp to $5002121
$501–$2,0001111
$2,001 –$5,00033
$5,001–$10,000415
$10,001-0
Asia & Pacific DepartmentUp to $50033
$501–$2,000639
$2,001–$5,000213
$5,001–$10,00011
$10,001 +156
European DepartmentUp to $5000
$501 –$2,00022
$2,001–$5,000246
$5,001–$10,00077
$10,001 +12223
Middle East & Central AsiaUp to $50022
Department$501–$2,0007310
$2,001–$5,000123
$5,001–$10,00011
$10,001 +314
Western Hemisphere DepartmentUp to $5000
$501–$2,00044
$2,001 –$5,00010818
$5,001–$10,000123
$10,001 +527
Total GDDS participants/SDDS
subscribers8864
Sources: IMF Statistics Department; and United Nations Statistics Division (2007) for 2005 GDP per capita data.Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard.
Sources: IMF Statistics Department; and United Nations Statistics Division (2007) for 2005 GDP per capita data.Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard.

These observations suggest that, as a practical matter, SDDS subscription may not be a reasonable or realistic goal for many GDDS countries in the foreseeable future. At the same time, they suggest certain important threshold values for identifying future SDDS subscribers. On the basis of affordability, GDDS participants with a per capita income near or above $5,000 should be prime candidates for SDDS subscription, while countries with a per capita income of $2,000 to $5,000 could reasonably establish a goal of moving to the SDDS over a period of years. Below $2,000, the resource requirements probably preclude most countries from subscribing to the SDDS. Although eight subscribers to the SDDS are below the threshold, four are transition economies (indeed, all graduates from GDDS to SDDS have been transition economies) that seem to have inherited an established statistical tradition and infrastructure, and competent statistical staff.

Table 3.2 also shows some regional differences in the distribution of SDDS and GDDS countries. Africa and Europe are two opposite extremes, with only one SDDS subscriber and 39 GDDS participants in Africa—about 44 percent of total GDDS membership. Europe comprises about 50 percent of SDDS participants and has only three GDDS countries. The distribution of countries between the two standards is more balanced in other regions, including the Middle East and Central Asia region.

Box 3.1.The 25/50 Program

At the time of the Fourth Review of the IMF Data Dissemination Standards Initiatives in 2001, the Executive Board encouraged an expansion of subscriptions to the Special Data Dissemination Standard (SDDS) in order to promote greater access to international capital markets. In response, the IMF Statistics Department implemented an aggressive SDDS outreach and technical assistance effort called the “25/50 Program.” This program identified 25 countries considered capable of meeting the SDDS within the following two to three years. It also identified a further group of about 50 countries that could meet the SDDS within about five years. As countries subscribed to the SDDS, the first group was replenished from the larger pool.

Over the next few years, staff worked intensively with as many countries as resources permitted. The program was conducted, in part, through outreach seminars in Greece in 2002, Mexico in 2003, Uruguay in 2004, and South Africa and Thailand in 2005. It also involved close collaboration of staff at headquarters with the designated SDDS coordinators. The seminars helped countries assess their data dissemination practices relative to the SDDS and provided guidance to the countries on how they could meet the SDDS requirements. They also provided information on the roles and responsibilities of country coordinators for the SDDS and other operational aspects of the standard.

Since the start of the 25/50 Program in 2002, 14 countries have subscribed to the SDDS (see table). Of these, six countries graduated from the General Data Dissemination System (GDDS). This increase in the number of countries subscribing to the SDDS or graduating from the GDDS to the SDDS is significant considering that only eight new subscribers joined the SDDS in the five years from 1997–2001.

It should be noted, however, that the pace of subscription is also determined by several exogenous factors, such as countries’ motivation to invest in statistical capacity and the importance that they give to transparency. Experience indicates that the most significant factor hindering subscription to the SDDS among countries with the requisite statistical capacity often is their reluctance to disclose data on the international reserves and foreign currency liquidity template or other SDDS requirements that are considered sensitive.

Countries Joining the Special Data Dissemination Standard Following the 25/50 Program
CountryDate of Subscription
LuxembourgMay 12, 2006
Moldova, Republic ofMay 2, 2006
MoroccoDec. 15, 2005
RomaniaMay 4, 2005
Russian FederationJan. 31, 2005
Egypt, Arab Republic ofJan. 31, 2005
Belarus, Republic ofDec. 22, 2004
Kyrgyz RepublicFeb. 26, 2004
UruguayFeb. 12, 2004
BulgariaDec. 1, 2003
ArmeniaNov. 7, 2003
KazakhstanMar. 24, 2003
UkraineJan. 10, 2003
GreeceNov. 8,2002
Source: IMF Statistics Department.
Source: IMF Statistics Department.

Owing to current resource constraints and a smaller pool of countries in a position to subscribe to the SDDS in the next few years, the Statistics Department has deemphasized the 25/50 Program by targeting a smaller group of countries since the beginning of 2006. The department assists these identified countries in developing work plans targeting SDDS subscription. It also provides additional technical assistance through financial resources provided by a project funded by the UK Department for International Development and by Japanese authorities.

Box 3.1 elaborates on the IMF’s initiatives to increase the number of SDDS subscribers, which includes measures to encourage selected GDDS participants to progress to the SDDS.

Performance

This section assesses the performance of the GDDS from the perspective of commitments of GDDS members. The results illuminate why some GDDS countries do not meet SDDS requirements of timeliness and periodicity of data and also identify existing gaps in the system.

Previous assessments of the performance of the GDDS by the IMF’s Executive Board have been relatively broad. Recognizing the develop-mental nature of the GDDS, regular reviews of the data standards by the board have focused on the growing participation of countries and the fact that countries and the donor community “broadly recognize the GDDS as the core framework for statistical capacity building” (Sixth Review, paragraph 26) as measures of the extent to which the GDDS has met its objectives (IMF, 2005). The Fifth Review of Data Standards took a similar tack (IMF, 2003).

Looking more specifically at the commitments of GDDS members, however, this chapter assesses experience with the GDDS by analyzing metadata and methodologies, plans for improvement against developmental needs, and sociodemographic data. Ten years after the inception of the GDDS, it is also appropriate to assess data dissemination practices—identifying areas of weaknesses in meeting timeliness and periodicity requirements.4

In assessing the GDDS, we consider practices of compilation separately from dissemination. We use the methodologies currently used by GDDS participants as the key criterion for analyzing the compilation of macroeconomic statistics. For example, metadata may specify that a country prepare government finance statistics using the IMF’s Manual on Government Finance Statistics (GFSM 1986) methodology or adopt the current best practice manual (GFSM 2001). The countries’ GDDS plans for improvement are also examined. For sociodemographic data, which are recommended by the GDDS but not included in the SDDS, we analyze time-series data on statistical capacity indicators prepared by the World Bank.

Thereafter, more emphasis is placed on data dimension, including the dissemination aspects of the system. We assess progress directly against the data dissemination “targets” contained in the GDDS by looking both at the observance of periodicity and timeliness of data dissemination.

This step is then carried further by comparing dissemination by countries against the tougher standards of the SDDS. In doing so, we recognize that graduation to SDDS was never established as a goal for GDDS participants and cannot, therefore, strictly be used as a test of success or failure of the GDDS. The SDDS standard, however, is useful as a benchmark that can emphasize the distance that countries must yet travel before they can be presumed to meet international best practice, the importance of data dissemination, and the areas where technical assistance is most needed.

Finally, the results of this analysis of progress are used to assess why some GDDS countries do not meet SDDS requirements of timeliness and periodicity of data and to identify areas in most need of improvement.

Analyzing GDDS Metadata and Data Compilation

In analyzing GDDS metadata and data compilation, this section considers the methodologies that GDDS participants use for compiling comprehensive frameworks, countries’ GDDS plans for improvement, and sociodemographic data. Participants have significantly progressed in adopting methodologies, although in some countries, progress has been slow; and participants assign a relatively low priority to data dissemination issues and to the update of sociodemographic data.

Methodologies for Compiling QDDS Comprehensive Frameworks

The macroeconomic data that member countries compile are broadly based on internationlly accepted methodologies that the IMF and other international organizations have developed. For real sector statistics, the methodology used is contained in the System of National Accounts (SNA) that was produced collaboratively by the IMF, World Bank, Commission of the European Communities, Organization of Economic Cooperation and Development, and the United Nations. The first version of the SNA dates to 1953 and has been updated twice since then, in 1968 and 1993 (1993 SNA). For fiscal sector statistics, the IMF has prepared the Government Finance Statistics Manual (GMSM), the first edition of which was published in 1986 and the second and latest one in 2001. With respect to monetary and financial statistics, the IMF published the Monetary and Financial Statistics Manual (MFSM) in 2000 to guide compilers of official statistics in this sector. It replaces the 1984 Guide to Money and Banking Statistics.

The extent to which GDDS participants have adopted these methodologies could be viewed as a measure of the extent to which GDDS participation has led to quality improvements in participants’ statistical systems. It should be noted that the extent of adoption of new methodologies varies from one country to another, and several countries have not fully adopted most methodologies. What is important is that a country broadly follow the recommendations of the latest methodology or follow a path toward its implementation. Table 3.3 shows the adoption of new methodologies by region as of November 2007, judged by countries’ metadata and IMF technical assistance mission reports. According to this table, the rate of adoption of the Balance of Payments Manual, fifth edition (BPM5) at 91 percent was the highest for all regions, followed by that of the 1993 SNA (64 percent), the MFSM (56 percent), and the GFSM 2001 (13 percent).

Table 3.3.Adoption of New Methodologies by Countries Participating in the General Data Dissemination System, by Region1
Region1993 SNAGFSM 2001MFSMBPM5
Number of Countries
African Department1331120
Asia & Pacific Department5028
Middle East & Central Asia Department92711
Western Hemisphere Department821111
Total for all regions3573150
Percent of All Countries in Sample
African Department59145091
Asia & Pacific Department63025100
Middle East & Central Asia Department75175892
Western Hemisphere Department158585
All regions64135691
Source: IMF Statistics Department

For a sample of 55 GDDS participants.

Source: IMF Statistics Department

For a sample of 55 GDDS participants.

A regional analysis of Table 3.3 shows that excluding GFSM 2001 and MFSM in the countries covered by the African Department and the Asia and Pacific Department, the adoption rates of new methodologies are well above 50 percent in all regions, with the highest rate of adoption for the BPM5. For this methodology, the Asia and Pacific Department had the highest adoption rate (100 percent) followed by the Middle East and Central Asia Department (92 percent), African Department (91 percent), and Western Hemisphere Department (85 percent). The SNA has the next highest adoption rate, at 75 percent in the Middle East and Central Asia Department, 63 and 62 percent for the Asia and Pacific and Western Hemisphere Departments, respectively, and 59 percent for Africa. As mentioned earlier, the African and Asia and Pacific Departments still lag behind in the adoption of the MFSM, with just 50 percent for the former and only 25 percent for the latter. The Western Hemisphere Department (85 percent) and Middle East and Central Asia Department (58 percent) are the two regions where this methodology has been widely adopted.

The adoption of the GFSM 2001 is the lowest of the four methodologies in all regions. The countries covered by the Middle East and Central Asia Department lead all regions with 17 percent of countries adopting this methodology, followed by the Western Hemisphere and African Departments, with 15 and 14 percent, respectively. None of the Asia and Pacific countries in the sample has adopted this methodology to date.

The availability of resources also plays an important role in adopting and implementing new methodologies. Financial resources are needed to conduct more demanding surveys in terms of expanded coverage, improvements in periodicity and timeliness, and compensation for additional staff. Technical expertise frequently is needed to guide the countries in implementing these methodologies.

Overall, it may be concluded that GDDS participants have made significant progress in adopting and implementing current best-practice statistical methodologies. At the same time, it must be acknowledged that, for many reasons, progress in some countries has been slow, and some distance remains to be traveled.

Plans for Improvement

Plans for improvement are central to the GDDS. Initial plans are developed during technical assistance missions in collaboration with country authorities and are expected to be updated once a year. These plans reflect the actions that the country needs to take to at least meet the GDDS recommendations. Countries are encouraged to determine a time frame for implementation of the plans, as well as for the financing and technical assistance needed for implementation.

This section analyzes plans for improvement by dataset for each sector and by region. To facilitate the analysis, plans were categorized using the IMF’s Data Quality Assessment Framework (DQAF). As shown in Tables 3.4, 3.5, and 3.6, the analysis concludes that data dissemination issues are assigned a relatively low priority.

Table 3.4.Regional Differences in Major Issues Identified in General Data Dissemination System Improvement Plans(In percent of major issues)
Data Quality Assessment

Framework Element
African

Department
Asia & Pacific

Department
Middle East &

Central Asia

Department
Western

Hemisphere

Department
3.1 Source data25282321
2.2 Scope11141322
0.2 Resources16953
3.3 Statistical techniques9101411
5.1 Data accessibility78912
2.1 Concepts and definitions79128
4.1 Periodicity and timeliness5557
2.3 Classification/sectorization5654
5.2 Metadata accessibility4332
0.1 Legal and institutional environment4212
4.2 Consistency2132
2.4 Basis for recording1223
Other5243
Total (percent)100100100100
Total (number of issues)989354348427
Source: IMF Statistics Department.
Source: IMF Statistics Department.
Table 3.5.Major Issues by Data Category(Numbers of major issues)
Data Quality Assessment Framework

Dimensions and Elements
National

Accounts
PricesGovernment

Operations
Government

Debt
Debt Clearing

System
Balance of

Payments
Total
0. Prerequisites of quality691836111214
1. Integrity001000
2. Methodological soundness28254313392629
3. Accuracy and reliability48541715274736
4. Serviceability1011010998
4.1 Periodicity and timeliness5177766
4.2 Consistency4023122
4.3 Revision policy and practice1010011
5. Accessibility811132514512
Total (percent)100100100100100100100
Total(number of issues)5292673332872944082,118
Source: IMF Statistics Department.
Source: IMF Statistics Department.
Table 3.6.Number of Times an Issue Related to the Data Quality Assessment Framework (DQAF) Is Mentioned in Improvement Plans
RankDQAF Dimensions and ElementsAll RegionsPercent
13.1 Source data51224
22.2 Scope30214
30.2 Resources22010
43.3 Statistical techniques21810
55.1 Data accessibility1788
62.1 Concepts and definitions1768
74.1 Periodicity and timeliness1186
82.3 Classification/sectorization1055
95.2 Metadata accessibility703
100.1 Legal and institutional environment573
114.2 Consistency462
122.4 Basis for recording372
3.4 Assessment and validation of intermediate
13data and statistical outputs201
140.4 Other quality management191
154.3 Revision policy and practice151
16All other251
Total2,118100
Source: IMF Statistics Department.
Source: IMF Statistics Department.

Table 3.4 provides a regional analysis of GDDS plans for improvement. In Africa, major constraints to data dissemination comprise source data, inadequate resources, scope, and statistical techniques. In the Asia and Pacific region, source data, scope, and statistical techniques are mentioned most often as a major issue. In the Middle East and Central Asia region, source data are most often mentioned, followed by statistical techniques, and scope. In the Western Hemisphere region, the major issues are scope and source data, followed by data accessibility and statistical techniques.

Table 3.5 shows a breakdown of major issues by data categories evaluated in IMF reports that assess a country’s adherence to good statistical practices (Reports on the Observance of Standards and Codes, or ROSCs). This table, which shows the share of issues (DQAF dimensions and elements) for each category, finds that the accuracy and reliability category, of which an element is source data, is the most important issue the countries list in their plans for improvement with respect to national accounts, prices, and the balance of payments. Methodological soundness, comprising concepts and definitions, scope, classification, and sectorization, is a major issue for the government operations and the depository corporations survey. The prerequisites for quality, comprising resources, are a major concern in compiling government debt data. Independently conducted ROSC assessments therefore tend to confirm the pattern of weakness that countries identify in their metadata.

Table 3.6 summarizes the results for any issues mentioned by the 55 countries in the sample. Common issues reflected in the plans for improvement for all countries are ranked. The countries’ plans for improvement are not always fully comprehensive. For example, in one section of the metadata, a country may refer to the need to improve timeliness and periodicity, while the plans of improvement do not take up this point. Thus, the findings of this analysis may not represent the full extent of existing weaknesses. As well, plans for improvement may reflect the authorities’ own assessments of weaknesses and priorities.

Nevertheless, source data, scope, resources, statistical techniques, and concepts and definitions are the major issues facing countries in all regions. Significantly, data dissemination issues consistently are assigned a relatively low priority. This reflects the current orientation and emphasis in the GDDS. Going forward, if the IMF were to focus more on the dissemination aspect of the GDDS, periodicity and timeliness of data would become more prominent in the plans for improvement.

Sociodemographic Data

This section examines developments in the area of sociodemographic data over the last seven years and concludes that participants assign the update of sociodemographic data a relatively low priority. The section uses the statistical capacity indicator (SCI) of the World Bank as a proxy. Although this indicator measures overall statistical capacity, its measurement is largely influenced by sociodemographic data, with more than 70 percent of the criteria included in the dimensions being sociodemographic data.

From Table 3.7, it is clear that the SCI has increased substantially since 1999 in all regions (see also Appendix 3.3). For the GDDS countries as a group, the SCI increased by 24 percent over the seven years, with the highest increase of 34.9 percent in the Asia and Pacific and the lowest increase of 13.7 percent in Africa. The increase in the SCI of the GDDS countries did, however, slow down significantly over the last two years. The SCI increased by only 4.3 percent for all the countries as a group, with the highest increase in the Middle East and Central Asia countries and the lowest in Western Hemisphere countries.

Table 3.7.Statistical Capacity Indicator
Number of

Countries

that Included

Millennium

Development

Goals in

Metadata
Percent Change
Average Score1
1999–

2004
2004–

06
1999–

2006
1999200420052006
African Department47.652.552.154.110.33.113.73
Asia & Pacific Department54.168.972.373.027.46.034.90
Middle East & Central Asia
Department53.165.570.170.523.47.632.81
Western Hemisphere
Department56.571.471.572.826.41.928.81
Total51.361.162.363.719.14.324.25
Source: World Bank (2007).

Scale of 0–100. A score of 100 indicates that a region meets all the criteria.

Source: World Bank (2007).

Scale of 0–100. A score of 100 indicates that a region meets all the criteria.

The strong improvement in the SCI between 1999 and 2004 is likely most attributable to the commitment made by countries in 2000—the United Nations Millennium Declaration—and the subsequent development of statistics to track the eight Millennium Development Goals (MDGs).

Although the GDDS was amended to explicitly recognize the United Nations’ MDG indicators and the development of appropriate statistical monitoring systems in late 2003, only 9.1 percent of the GDDS participants have adjusted their sociodemographic metadata to include the MDGs. This may indicate that the GDDS participants do not regard updating the GDDS sociodemographic metadata to include data on the MDGs a priority, because extensive data on the sociodemographic data categories and the MDGs are available on the websites of the World Bank and the United Nations.

Analyzing GDDS Data Dissemination: Timeliness and Periodicity of Data

How well have GDDS countries managed to achieve the data dissemination goals set out in the GDDS? And to what extent have countries been able to move beyond the GDDS and achieve the more stringent requirements for the SDDS? The sections that follow address these questions first by comparing the dissemination practices of timeliness and periodicity of data of GDDS participants with those recommended as good practice by the GDDS. Second, dissemination is compared against the tougher standards of the SDDS. Finally, we summarize reasons why many GDDS countries do not meet SDDS requirements.

Comparing Participants’ Statistical Practices Against the QDDS

As shown in Table 3.8a, GDDS countries on average are able to meet some of the periodicity and timeliness recommendations for the comprehensive framework of the GDDS.5 For example, close to 73 percent of GDDS countries meet the GDDS recommendation for periodicity and timeliness for the comprehensive framework for national accounts and 89 percent for the depository corporations survey, but only 64 percent meet the recommendation for government operations, and only about 69 percent meet the balance of payments recommendation. For the GDDS, the periodicity of the comprehensive frameworks is annual, except for the depository corporations survey, for which monthly periodicity is recommended.

Table 3.8a.Participants’ Compliance with General Data Dissemination System Recommendations for Timeliness and Periodicity, by Data Category(In percent)
Data CategorySectorAllAfrican

Department
Asia &

Pacific

Department
Middle

East &

Central Asia

Department
Western

Hemisphere

Department
National accountsReal sector72.763.677.872.784.6
Production indexReal sector30.99.155.645.538.5
UnemploymentReal sector40.04.544.472.769.2
Wages/earningsReal sector45.531.844.463.653.8
EmploymentReal sector54.531.855.681.869.2
Producer price indexReal sector20.04.533.345.515.4
Consumer price indexReal sector92.7100.088.990.984.6
Government operationsFiscal sector63.663.644.472.769.2
Central government debtFiscal sector60.059.133.354.584.6
Central bankFinancial sector83.681.877.881.892.3
Banking surveyFinancial sector89.181.888.990.9100.0
Official reservesExternal sector34.531.855.627.330.8
Balance of paymentsExternal sector69.159.177.881.869.2
Merchandise tradeExternal sector58.240.966.772.769.2
Simple averageAll sectors58.247.460.368.266.5
PopulationSociodemographic sector43.613.644.463.676.9
Source: IMF Statistics Department.
Source: IMF Statistics Department.
Table 3.8b.Participants’ Compliance with General Data Dissemination System Recommendations for Periodicity, by Data Category(In percent)
Data CategorySectorAllAfrican

Department
Asia &

Pacific

Department
Middle

East &

Central Asia

Department
Western

Hemisphere

Department
National accountsReal sector98.2100.088.9100.0100.0
Production indexReal sector38.218.255.654.546.2
UnemploymentReal sector45.59.155.681.869.2
Wages/earningsReal sector54.536.455.681.861.5
EmploymentReal sector69.154.566.790.976.9
Producer price indexReal sector25.59.133.354.523.1
Consumer price indexReal sector94.5100.088.990.992.3
Government operationsFiscal sector69.172.755.672.769.2
Central government debtFiscal sector92.7100.077.881.8100.0
Central bankFinancial sector94.5100.077.8100.092.3
Banking surveyFinancial sector94.590.988.9100.0100.0
Official reservesExternal sector74.581.877.881.853.8
Balance of paymentsExternal sector100.0100.0100.0100.0100.0
Merchandise tradeExternal sector63.650.066.772.776.9
Simple averageAll sectors72.565.970.683.175.8
PopulationSociodemographic sector47.318.255.663.676.9
Source: IMF Statistics Department.
Source: IMF Statistics Department.

It would be useful to further investigate the reasons for this—whether it is due to a lack of policy focus on timely dissemination or whether there are other hurdles, perhaps related to resource constraints or for other reasons (e.g., hard-copy publications, which take time to produce). Timeliness is also a factor for some of the other datasets such as employment data, where close to 70 percent of countries meet periodicity, but only about 55 percent are able to meet the timeliness recommendation as well. These results are relevant for technical assistance priorities, both in terms of subject areas and the focus within them.

Because the purpose of the GDDS is to help countries develop their statistical systems and to move at least some countries beyond the level of GDDS recommendations, the next section compares country practices against the more stringent SDDS periodicity and timeliness standards, particularly relevant for the 24 countries borrowing in private capital markets.

Comparing Data Dissemination Practices Against the SDDS “Benchmark”

This section compares the current dissemination practices of GDDS participants with those required by the SDDS and analyzes existing gaps on a regional basis. It concludes that it could be desirable to consider expanding the GDDS to include all SDDS data categories.

As shown in Tables 3.9 and 3.10, most GDDS countries are not able to meet the SDDS requirements. Table 3.9 shows the extent to which GDDS participants achieve both periodicity and timeliness for the various data categories. Setting the bar relatively high, the SDDS requires, for example, quarterly GDP with timeliness of one quarter, while the GDDS requires annual data for GDP with generously defined timeliness of six to nine months. Table 3.10 considers just periodicity requirements, which are less difficult to meet.

Table 3.9.Measuring How Well GDDS Participants Meet SDDS Requirements for Periodicity and Timeliness, by Data Category(In percent)
Data CategorySectorAllAfrican

Department
Asia &

Pacific

Department
Middle

East &

Central Asia

Department
Western

Hemisphere

Department
National accountsReal sector10.90.033.318.27.7
Production indexReal sector25.54.555.636.430.8
UnemploymentReal sector16.40.033.345.57.7
Wages/earningsReal sector18.24.533.336.415.4
EmploymentReal sector20.00.044.445.515.4
Producer price indexReal sector16.44.522.245.57.7
Consumer price indexReal sector89.1100.088.981.876.9
Government operationsFiscal sector21.818.222.227.323.1
Central government debtFiscal sector43.654.522.245.538.5
Central bankFinancial sector23.618.233.318.230.8
Banking surveyFinancial sector36.431.833.345.538.5
Official reservesExternal sector34.531.855.627.330.8
Balance of paymentsExternal sector40.027.355.663.630.8
Merchandise tradeExternal sector47.336.466.763.638.5
Simple averageAll sectors31.723.742.942.928.0
PopulationSociodemographic sector43.613.644.463.676.9
Source: IMF Statistics Department.Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard.
Source: IMF Statistics Department.Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard.
Table 3.10.Percentage of GDDS Participants Meeting the SDDS Periodicity Requirements, by Data Category
Data CategorySectorPercent
National accountsReal20.0
Production indexReal38.2
UnemploymentReal21.8
Wages/earningsReal23.6
EmploymentReal27.3
Producer price indexReal25.5
Consumer price indexReal94.5
Government operationsFiscal47.3
Central government debtFiscal65.5
Central bankFinancial94.5
Banking surveyFinancial94.5
Official reservesExternal74.5
Balance of paymentsExternal52.7
Merchandise tradeExternal63.3
Simple average53.1
PopulationSociodemographic47.3
Source: IMF Statistics Department.Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard.
Source: IMF Statistics Department.Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard.

About 32 percent of the GDDS countries included in the sample are able to meet both periodicity and timeliness requirements (Table 3.9), but if we consider only the periodicity requirements, 53 percent of these countries meet the requirements. This suggests that an important constraint is the ability to disseminate data in a timely manner. (Some reasons for this are further discussed later in this chapter.)

Countries experience the most serious problems in the real sector, excluding the CPI, and the least serious problems in the financial sector, where the major problems are countries’ ability to comply with the SDDS timeliness requirement. For GDDS countries as a group, less than 26 percent met the SDDS requirements for any real sector data category.

In addition, many countries experience problems meeting the dissemination recommendations with respect to short-term or tracking indicators, especially for those data categories recommended in the real sector, with the least of the problems experienced in the financial sector, followed by the fiscal sector. Except for the consumer price index (CPI), less than 50 percent of the countries in the Africa and Asia and Pacific regions compile and disseminate these real sector data categories meeting both the GDDS periodicity and timeliness recommendations. Averaged over all datasets, about 62 percent of requirements are met by GDDS participants.

Table 3.8b shows performance relative to just the periodicity indicator. The overall averages are higher by more than 10 percentage points. Almost universal observance exists for national accounts, CPI, and both the financial and external sectors. This suggests that an important constraint for countries trying to carry out the GDDS recommendations is timeliness. However, this percentage increased to 40 percent when compared only with the SDDS periodicity requirement. In the case of the financial sector, less than 37 percent of the group met the SDDS requirements for both data categories, but close to 95 percent met the periodicity requirements of both data categories.

To meet SDDS data requirements, it would be necessary for countries to compile and disseminate data on the international reserves template, general government sector, external debt, and the international investment position (IIP). The reserves template is not included in the GDDS, and the latter three data categories are included as encouraged extensions; therefore, no comprehensive data are available for analysis. One could consider expanding the GDDS to include all SDDS data categories. This would help define the path for countries to graduate to the SDDS.

Looking at the overall results across regions suggests that the Asia and Pacific and Middle East and Central Asia regions are able to meet about 43 percent of the requirements, followed by Western Hemisphere countries at about 28 percent, and Africa, accomplishing just 24 percent of the periodicity and timeliness requirements.

Why GDDS Countries Do Not Meet SDDS Requirements

From the above tables, it is clear that the real sector, except for the CPI, is the main area in which countries are experiencing the most problems. Countries experience problems not only in meeting the SDDS timeliness requirements for real sector data but also in compiling data to meet SDDS periodicity requirements. From the regional analysis, it appears that the problems are more pronounced in Africa and the Western Hemisphere. These two regions are the farthest from the SDDS requirements (timeliness and periodicity) in the areas of real and external sector data.

Possible reasons why countries find it particularly difficult to meet timeliness requirements may be because:

  • SDDS timeliness requirements, stringent to meet capital market needs, are considerably tougher than those of the GDDS recommendations, particularly for GDP, the labor market, central government operations, and balance of payments statistics. In this respect, GDDS participation does not prepare countries to move to the SDDS standards.

  • GDDS countries may not be giving a high priority to improving timeliness of dissemination. Reasons could be the absence of demand for such data (although, as noted above, about 60 percent of the countries studied have sovereign credit ratings and would face demand for timely data from the rating agencies). Another reason may simply be that timeliness is not emphasized in technical assistance programs, including in the GDDS. Yet another explanation could be that data are mainly prepared for internal government access and for interested parties and the rating agencies, while dissemination to the general public is a lower priority.

  • The first data release is a hard-copy publication, which takes more time (even if it is subsequently posted on the Internet).

The relatively large lags for real sector and external sector data may simply reflect the fact that the work on these sectors involves expensive and resource-intensive source data. The regional deviation in compliance with the SDDS requirements could be the result of the differences in the availability of resources to absorb and retain capacity-building technical assistance.

Capital Market Access

The GDDS was developed for a broad group of countries that do not necessarily lack ambition to access capital markets, but that are more likely to be recipients of official development financing and technical assistance. The SDDS was developed against the backdrop of informational failures affecting capital markets. Virtually all SDDS subscribers are active borrowers in capital markets. However, Tables 3.11 and 3.12 show why the assumption about GDDS countries is no longer fully justified.

Table 3.11.General Data Dissemination System Countries with Sovereign Credit Ratings, by Region and GDP Per Capita

(as of April 2007)1

Rating AgencyGDP Per

Capita (US$)
Region/CountryS&PMoody’sFitch
African Department
1. Seychelles8,668
2. Mauritius5,052
3. Botswana5,014
4. Namibia3,018
5. Nigeria863
6. Senegal710
7. Kenya560
8. Mozambique338
9. The Gambia316
10. Uganda316
11. Madagascar266
12. Malawi166
Asia & Pacific Department
1. China1,533
2. Sri Lanka1,154
3. Mongolia706
4. Vietnam631
Middle East & Central Asia Department
1. Qatar51,809
2. Kuwait27,621
3. Oman11,792
4. Macedonia2,778
5. Jordan2,198
6. Azerbaijan1,493
7. Georgia1,450
8. Pakistan697
Western Hemisphere Department
1. Trinidad and Tobago11,311
2. Venezuela4,949
3. Panama4,716
4. Grenada4,415
5. Belize4,097
6. Guatemala2,534
7. Honduras1,162
8. Bolivia1,059
9. Nicaragua895
Sources: IMF Statistics Department; United Nations Statistics Division (2007) for 2005 GDP per capita data; and Fitch Ratings, Standard and Poor’s, and Moody’s Investors Service.

Countries in the sample of 55 with sovereign ratings.

Sources: IMF Statistics Department; United Nations Statistics Division (2007) for 2005 GDP per capita data; and Fitch Ratings, Standard and Poor’s, and Moody’s Investors Service.

Countries in the sample of 55 with sovereign ratings.

Table 3.12.Data Categories Used by Moody’s for Sovereign Bond Ratings That Are Not Fully Covered by the GDDS and SDDS
Indicators Not CoveredGDDS1SDDS
I. Economic structure and performanceReal sectorReal sector
GDP per capita (purchasing power parity basis)NoNo
Gross investment/GDP (%)PartiallyPartially
Nominal and real exports and imports of goods and services (% change)PartiallyPartially
Net exports of goods and services/GDP (%)PartiallyPartially
Openness of the economy (exports + imports of goods and services/GDP) (%)PartiallyPartially
II. Government financeFiscal sectorFiscal sector
General government revenue/GDP, expenditure/GDP, and financial balance/GDP (%)EncouragedYes
General government primary balance/GDPNoEncouraged
General government debt/GDP and general government debt/general government revenue (%)NoPartially
General government interest payments/general government revenue (%)NoEncouraged
III.External payments and debtExternal sectorExternal sector
Real effective exchange rate (% change)NoNo
Relative unit labor costs (index)PartiallyPartially
External debt (U.S. dollars) and external debt/GDP (%)PartiallyYes
External debt/current account receipts (%)PartiallyYes
Net foreign direct investment/GDP (%)PartiallyYes
Net international investment position/GDP (%)EncouragedYes
IV.Monetary, vulnerability, and liquidity indicatorsFinancial sectorFinancial sector
Debt service ratio (interest + current year principal/current account receipts) (%)Partially (external sector)Encouraged (external sector)
Dollarization ratio (total foreign currency deposits in domestic banks/total deposits in domestic banks) (%)PartiallyPartially
Dollarization vulnerability indicator (foreign currency deposits in domestic banks/official

foreign exchange reserves + foreign assets in domestic banks) (%)
PartiallyPartially
V. Financial soundness indicators
External vulnerability indicatorNoPartially (reserves template)
Liquidity ratio (liabilities to BIS banks within one year/total assets held in BIS banks)NoNo
Number/percent of data categories covered (fully, partially, or encouraged) from the selected indicators (30 in total)22 (73.3%)26 (86.7%)
Source: Moody’s Investors Service (2004).Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard. BIS = Bank for International Settlements.

The comparison is mainly based on data categories included in part B of Table 1 of the GDDS Guide (updated in October 2004). If the comprehensive frameworks in part A of the same table were accounted for, the GDDS would virtually cover all indicators for four sectors, as the comprehensive frameworks are too broad and are rather targets for developing statistical systems than indicators that countries practically disseminate under the GDDS.

Source: Moody’s Investors Service (2004).Note: GDDS = General Data Dissemination System; SDDS = Special Data Dissemination Standard. BIS = Bank for International Settlements.

The comparison is mainly based on data categories included in part B of Table 1 of the GDDS Guide (updated in October 2004). If the comprehensive frameworks in part A of the same table were accounted for, the GDDS would virtually cover all indicators for four sectors, as the comprehensive frameworks are too broad and are rather targets for developing statistical systems than indicators that countries practically disseminate under the GDDS.

Table 3.11 lists GDDS countries that have received sovereign credit ratings and have borrowed in international capital markets. The table shows that 33 GDDS countries included in the sample (60 percent) have sovereign credit ratings issued by international rating agencies. Obviously, the rating agencies have not been deterred by possible data shortcomings in GDDS countries, and at the same time have had sufficient data at their disposal when assigning sovereign ratings. This means that GDDS countries in many cases are incurring the costs of compiling the information that credit rating agencies require, but are not reaping the benefit of public dissemination. Moreover, these GDDS countries have also forgone a benefit that could be reaped with SDDS subscription, given the empirical evidence that SDDS subscription lowers borrowing costs for its subscribers.6 With access to capital markets, the SDDS would be the relevant standard of data dissemination for these countries, and GDDS membership should therefore spell out the transition to the SDDS.

This chapter already discussed this point in the analysis of how close GDDS countries are to meeting GDDS and SDDS requirements. Of these countries, 64 percent have income levels (GDP per capita) of above $1,000, and 85 percent have levels above $500. Considering these resource constraints, the SDDS may not be a realistic goal for all these countries. Of 33 GDDS countries with sovereign ratings, 15 meet the $2,000 threshold for future SDDS subscription and should therefore aim to subscribe to the SDDS, while for the others a good performance at the GDDS level might be a realistic goal. Perhaps a somewhat modified GDDS approach, with more emphasis on disseminating the data relevant for capital markets, could be considered for these countries.

Table 3.12 addresses a related question as to whether and to what extent the GDDS datasets are relevant for the analysis performed by rating agencies. The table lists those data considered by a major rating agency for sovereign ratings that are not fully covered in the data categories required for the GDDS. The table also makes the same comparison for the SDDS to confirm that the datasets are in line with the requirements of capital market analysts. The table shows that both the GDDS and SDDS broadly cover Moody’s data requirements, although the SDDS is a closer match for some fiscal and external sector statistics (covering 86.7 percent of data categories). Also, the GDDS does not require dissemination of these data categories but recommends that countries develop these datasets.

Neither the GDDS nor SDDS covers the financial soundness indicators required by Moody’s. Given that the 33 countries are accessing capital markets, they would benefit from aligning their data dissemination programs closely with the requirements of the SDDS and including some financial soundness indicators. It should be noted that the data requirements as expressed in Moody’s reports are not explicit on timeliness and periodicity requirements (for instance, they do not mention the need for quarterly national accounts, as required by the SDDS, or whether annual national accounts are sufficient as recommended by the GDDS).

How Successful Was the GDDS in Guiding Countries to the SDDS?

What progress have GDDS participants made and how successful has the GDDS been in guiding countries to meet the SDDS requirements? To answer these questions, we assessed a sample of five GDDS participants from different regions. The sample includes Botswana (African Department), Cambodia (Asia and Pacific), Jordan (Middle East and Central Asia), Mauritius (Africa), and Panama (Western Hemisphere). Some countries made more progress than others; more focused plans for improvement and data dissemination aspects would have accelerated progress.

The assessment here is based on the following four aspects of data compilation and dissemination practices: (1) new data categories compiled/disseminated (from the list of GDDS and SDDS macroeconomic data categories); (2) improvements in coverage, (3) improvements in periodicity; and (4) improvements in timeliness. We first compared the current compilation and dissemination practices of the five GDDS participants with those at the time of GDDS participation (based on available data ROSCs, SDDS assessment mission reports, or metadata) and highlighted improvements. Second, we compared the current compilation and dissemination practices with SDDS requirements and identified shortcomings in the above-mentioned four aspects. The average assessment time frame is about five years.

The assessment of the statistical compilation and dissemination practices reveals that all countries in the sample made some progress in developing statistical compilation and dissemination practices in about five years, especially with regard to timeliness. As shown in Table 3.13, Cambodia, Jordan, and Mauritius made significant improvements, while Botswana and Panama made relatively less progress.

Table 3.13.Progress Made by Selected Countries Under the General Data Dissemination System1
Improvements Needed

to Meet the Special Data

Dissemination Requirements

(Number of requirements)
Improvements in Statistical Practices

(Number of data categories improved)
CountryGDP Per

Capita

(US$)
Assessment

Time Frame

(years)
New

data

compiled
Improvements

in coverage
Improvements

in periodicity
Improvements

in timeliness
Data categories

improved (total)
New

data
CoveragePeriodicity

and/or

timeliness
Botswana5,0145None11243111
Mauritius5,0525.5None291114139
Cambodia3844.5252510832
Jordan2,1984.523None913135
Panama4,7165223282110
Total/average4.9613152949151137
Source: IMF Statistics Department.

Botswana as of November 2006, compared with October 2001; Mauritius as of November 2006, compared with July 2001; Cambodia as of March 2002, compared with June-July 2006; Jordan as of November 2006, compared with February 2002; and Panama as of February 2006, compared with December 2000.

Source: IMF Statistics Department.

Botswana as of November 2006, compared with October 2001; Mauritius as of November 2006, compared with July 2001; Cambodia as of March 2002, compared with June-July 2006; Jordan as of November 2006, compared with February 2002; and Panama as of February 2006, compared with December 2000.

Despite the significant progress in three countries in the sample, the pace has been slow. For example, according to the 2002 data ROSC mission, Jordan should have been able to meet all the SDDS requirements in February 2005. According to the 2001 data ROSC mission, Mauritius should have been able to subscribe to the SDDS by July 2004. Of course, a user needs to interpret the ROSC missions’ assessments with caution, because the possible SDDS subscription time frames are obviously attached to a number of prerequisites, chief among them resources devoted and the commitment of the authorities.

It is clear that the progress made under GDDS participation in any given country greatly depends on the authorities’ commitment to data dissemination standards and statistical development in general, as well as on resources made available for both sustaining and developing statistical practices. While the level of commitment and available resources for statistics in the sampled countries may vary to a certain extent, the overall assessment of progress points to the conclusion that more focused plans for improvements on meeting SDDS requirements and more focus on data dissemination aspects would have accelerated progress significantly.

Overall Assessment and Recommendations

In view of the medium-term developmental framework of the GDDS and the specific commitments that participants were asked to make, the effectiveness of the GDDS can be judged most directly by observing the extent of countries’ participation in the system and examining their metadata, including plans for improvement, against developmental needs. At the same time, following 10 years of experience in using the GDDS, it would seem essential to judge effectiveness against the ultimate objective of improving data dissemination. The extent to which the GDDS has supported data improvements needed for progressing toward the SDDS for countries interested in doing so would also be germane.

The analysis presented in this chapter provides a somewhat mixed assessment of the experience with the GDDS. On the positive side of the ledger, participation has continued to grow—the combined participation/subscription in the GDDS/SDDS now covers five-sixths of the IMF membership—and the developmental aspects of the GDDS have been widely recognized. On the negative side, overall progress has been slower than envisaged. In part, this reflects a scarcity of resources and an often low priority assigned to statistics in national development plans. However, slow progress appears also to suggest a need to reconsider basic elements of the GDDS design.

How Well Has the GDDS Performed?

Promoting the production and dissemination of economic and financial data are the ultimate objectives of the GDDS. However, unlike the special standard, which commits its subscribing countries to observe a specific list of statistical practices, the general system commits its participating countries only to making more qualitative improvements in their statistical systems. The initial focus of the GDDS was on developing national systems in an explicit medium-term framework, with attention to producing and disseminating economic and financial data coming at a later stage. Reflecting these priorities, participating countries commit only to use the GDDS as a developmental framework, designate a country coordinator, and provide metadata that describe current statistical practices and plans for improvement. There is no commitment to data dissemination per se. Indeed, a premise that underlies the GDDS is that improvements in data quality need to be given a high priority and may need to precede improvement in dissemination practice.

As expected, metadata and plans for improvement generally confirm the existence of the sorts of weaknesses in statistical frameworks that justified the emphasis in the initial GDDS design on implementing comprehensive statistical frameworks. The most common problems, classified according to DQAF categories, comprise source data, scope, resources, statistical techniques, and concepts and definitions. Plans for improving data dissemination do not figure very prominently, possibly reflecting the initial focus of the GDDS, even though serious deficiencies exist in data dissemination and some “low-hanging fruit” remain to be picked.

In hindsight, success in adopting comprehensive statistical frameworks has been quite mixed, ranging from 91 percent of GDDS participants having adopted BPM5, to 13 percent having adopted GFSM 2001. Resource constraints in particular countries and the availability of technical assistance (or lack thereof) appear to have been important factors in determining the speed of adoption. Also, pronounced regional differences indicate the extent to which particular methodologies have been implemented.

Further, overall progress in strengthening statistical systems most likely has been slower than envisaged when the GDDS was established. The case studies point to the fact that countries with an interest in progressing toward the SDDS consistently failed to do so within the time frames judged feasible by IMF staff. It is disappointing, too, that only six countries have managed to progress from the GDDS to subscription to the more demanding SDDS (all of them transition countries mostly with a strong tradition and infrastructure in statistics as a remnant of the command economy). Although a substantial improvement was observed in sociodemographic data after 1999, GDDS participants did not respond by updating their sociodemographic metadata to include the MDGs. This may indicate that the GDDS participants do not regard updating the GDDS sociodemographic metadata to include data on the MDGs a priority, given that extensive data on the sociodemographic data categories and the MDGs are available on the websites of the World Bank and the United Nations.

Moreover, GDDS participants generally still have a long way to go in meeting the ultimate objective of strengthening data dissemination. About 60 percent of participants meet both the periodicity and timeliness objectives of the GDDS comprehensive frameworks. Whereas almost three-quarters of participants meet the periodicity goals, fewer than half meet the timeliness objective. When judged against the more demanding standards of the SDDS, slightly more than half of GDDS participants meet periodicity standards and about a third meet timeliness standards. Certain GDDS participants have obtained a sovereign credit rating and have become market borrowers but are still some distance from meeting SDDS subscription requirements—the standard for countries with market access. This distance may be related in part to weaknesses in data dissemination. Meeting the SDDS subscription requirements would, however, result in a significant reduction in borrowing costs for these countries.

Finally, important changes in the world since the inception of the GDDS have yet to be reflected in the system’s design and implementation. Most notable among these changes are globalization that increasingly relies on open capital and product markets, heightened emphasis on transparency and good governance, and technical developments associated with the spread of the Internet and growing reliance on electronic forms of disseminating information.

The next section provides a basis to reconsider and fine-tune certain aspects of the GDDS in order to improve its performance and relevance.

Strengthening the Emphasis on Data Dissemination

Five actions would bolster the performance of the GDDS in terms of strengthening the emphasis on data dissemination: rebalancing the GDDS, using per capita income to identify candidates for the SDDS, simplifying the data dimension, retaining plans for improvement as an important feature, and more explicitly recognizing regional differences.

Rebalancing the GDDS

A rebalanced formulation of the GDDS could give more emphasis to disseminating data to the public and less to generating and updating metadata descriptions of existing statistical practice. The original formulation gave more emphasis to developmental processes than to data dissemination because quality deficiencies could have undermined the usefulness of any data that might be disseminated, and because data users’ needs in nonmarket borrowing countries were less time-sensitive than in SDDS countries. A consequence of this emphasis is that GDDS participants have generally disseminated fewer data series, and in a less timely way than they might otherwise have done.

Moreover, views on the importance of data dissemination have changed since the inception of the GDDS. It is now widely recognized that the dissemination of data creates its own demand for better quality information and more extensive coverage of indicators. Data dissemination is likely to raise the profile and visibility of the statistical agencies and, by creating a demand for more and better statistics, may lead to a higher priority being placed on statistics in a country’s developmental plan and to more resources being allocated to the statistical agencies. The time sensitivity of users in many GDDS countries has increased, as evidenced by the significant number of market borrowers among them. Rising standards of governance and accountability have further increased the demand for timely data. The spread of the Internet and the increasing reliance on electronic publication as the best-practice first channel of dissemination has both raised the bar and reduced the cost of data dissemination, but has yet to be reflected in the GDDS.

These developments could be reflected in a revamped GDDS. Specifically, a new formulation could place greater emphasis on data dissemination to the public and less emphasis on updating the description of the existing system in the metadata. Plans for improvement could assign higher priority to the periodicity and timeliness objectives in the GDDS. Cost reductions and technological change make it feasible to adopt key aspects of the SDDS, notably the publication of the advance release calendar and the national summary data page. As part of their GDDS obligations, countries could be asked to make a good faith commitment to achieving the dissemination objectives, although it is not proposed that these objectives become a monitorable standard, as in the SDDS. For GDDS countries above a certain income threshold, progressing toward the SDDS could be an explicit goal.

Using Per Capita Income to Identify Candidates for the SDDS

The analysis of per capita income suggests it is a useful way to identify GDDS candidates that should be encouraged to establish SDDS subscription as an important goal. Analysts should presume that countries at or above $5,000 per capita should adopt an accelerated timetable for SDDS subscription, while those in the $2,000 to $5,000 range should adopt SDDS subscription as a priority medium-term goal. For most countries below $2,000, the focus should be on meeting the dissemination targets in the GDDS, while adopting SDDS practices on a selective basis as cost and technology permit.

Realigning Data Categories

A GDDS recasting could simplify and reformulate the data dimension to more closely approximate that of the SDDS. The difficulties that many countries (e.g., Botswana, Jordan, and Mauritius) have had in making the jump from GDDS to SDDS underscores the value of beginning to compile certain data categories as part of the GDDS, such as the reserves template, general government statistics, the IIP, and external debt. Necessarily, periodicity and timeliness requirements would be at a lower frequency than for the SDDS.

The experience of many GDDS countries as market borrowers, along with their need to provide many of these data to credit rating agencies, further reinforces the desirability of realigning the GDDS data dimension with that of the SDDS (because it meets most requirements of the credit rating agencies). Countries in many cases are incurring the costs of compiling the information that credit rating agencies require, but are not necessarily reaping the full benefit of public dissemination.

As part of a realignment of the data dimension to more closely approximate the SDDS, somewhat more attention could be given to indicator series, and somewhat less attention to full-blown comprehensive frameworks. When combined with an explicit SDDS end goal, these changes could be characterized as providing a “capital market track” for the GDDS.

Retaining Plans for Improvement

The developmental aspect of the GDDS should remain a priority. While analysts may reasonably argue that, after 10 years of experience, it is time to move beyond the process of merely describing existing practices, plans for improvement should continue to be an important feature of the GDDS. The system would maintain the existing obligation that these plans be updated regularly and be comprehensive. The plans for improvement are extremely important as a basis for interagency coordination of technical assistance and training.

The GDDS should explicitly recognize that IMF staff do not have a comparative advantage in providing certain types of technical support, for example, in helping countries develop household, government, and enterprise source data. Nor do IMF staff provide technical assistance for developing certain key data series such as labor market statistics. The GDDS could provide an explicit basis for interagency (both bilateral and multilateral) coordination. It also could acknowledge the desirability of linking statistical plans for improvement to medium-term national public expenditure frameworks and access to donor resources.

More Explicitly Recognizing Regional and Sectoral Differences

The GDDS should remain as a source of guidance for all IMF member countries. In doing so, however, the system could more explicitly recognize the important regional differences as regards the extent to which countries have bought into the data standards, adopted comprehensive statistical frameworks, and faced common statistical problems. Experience shows a pervasive need for intensive technical assistance in the area of real sector statistics.

Appendix 3.1. Selection of Sample Countries

This chapter has analyzed GDDS participants’ performance and progress in data dissemination practices on the basis of a representative sample designed to have adequate regional and economic representation and to focus on countries that actually used the GDDS as a development tool. We grouped the 88 GDDS participants (as of February 2007) by region following the IMF’s area departments: Africa, Asia and Pacific, Europe, and the Western Hemisphere. The two European countries in the sample—Albania and Macedonia, FYR—were included in Middle East and Central Asia Department, the closest region to Europe.

From these 88 countries, 12 countries were excluded that have not updated their metadata in the last four years (48 months) and can thus be considered as not actively participating in the GDDS. This also ensures that the information used for the analysis is current, since the analysis is done entirely based on countries’ metadata. About 70 percent of the frame (76 countries) was chosen in the sample, which therefore includes 55 countries.

The 55 countries in the sample were distributed by region, proportionally to the share of each of the four regions in the group of 76 countries in the sampling frame. The resulting breakdown of GDDS participants in the sample is as follows: Africa—22 participants; Asia and Pacific—8; Middle East and Central Asia—12; and Western Hemisphere—13 (see Table 3.A1). Finally, the list of 76 participants was grouped by region and ordered alphabetically. We then randomly selected GDDS participants from each region to be included in the sample (see the complete sample list in Table 3.A2), using a modified systematic random sampling method.

Appendix 3.2. GDDS and SDDS Coordinators Appointed by Countries

Both GDDS and SDDS participation require that a country appoint a national coordinator. Participation in the GDDS also constitutes a commitment by country authorities to update metadata and plans for improvement at least once a year. The countries appoint GDDS coordinators as part of their participation, and coordinators are responsible for updating the metadata. GDDS coordinators thus can play an important role in moving the reform agenda ahead. SDDS coordinators have considerably more day-to-day responsibilities because data are posted on the website on an ongoing basis, and metadata are certified every quarter. On the other hand, SDDS coordinators generally do not have a role to play in their country’s reform agenda. The SDDS coordinator role is thus more technical and less strategic than that of the GDDS coordinator.

Table 3.A1.Breakdown of Total and Sample General Data Dissemination System Participants by Region1
RegionsAfrican

Department
Asia &

Pacific

Department
Middle

East &

Central Asia

Department/

European

Department
Western

Hemisphere

Department
Total
Total GDDS participants3912172088
Percent in total44.313.619.322.7100
Metadata updated within 48 months2912171876
Percent in total38.215.822.423.7100
GDDS participants in the sample2281221355
Percent in total40.014.521.823.6100
Source: IMF Statistics Department.

The sum of percent shares may not exactly equal to 100 percent due to rounding.

Includes Albania and Macedonia.

Source: IMF Statistics Department.

The sum of percent shares may not exactly equal to 100 percent due to rounding.

Includes Albania and Macedonia.

Table 3.A2.General Data Dissemination System Participants Included in the Sample by Region
African DepartmentAsia & Pacific

Department
Middle East & Central

Asia Department
Western Hemisphere

Department
1. Angola1. Bangladesh1. Albania1. Antigua and Barbuda
2. Botswana2. Cambodia2. Afghanistan2. Belize
3. Central African Republic3. China3. Azerbaijan3. Bolivia
4. Congo, Dem Rep. of4. Kiribati4. Georgia4. Dominica
5. Congo, Rep. of5. Mongolia5. Jordan5. Grenada
6. Ethiopia6. Nepal6. Kuwait6. Guatemala
7. Gambia, The7. Sri Lanka7. Macedonia, FYR7. Honduras
8. Kenya8. Vietnam8. Oman8. Nicaragua
9. Liberia9. Pakistan9. Panama
10. Madagascar10. Qatar10. St. Kitts and Nevis
11. Malawi11. Tajikistan11. St. Lucia
12. Mauritius12. West Bank and

Gaza
12. Trinidad and Tobago
13. Mozambique13. Venezuela
14. Namibia
15. Nigeria
16. Rwanda
17. Senegal
18. Seychelles
19. Sierra Leone
20. Tanzania
21. Uganda
22. Zambia
Source: IMF Statistics Department.
Source: IMF Statistics Department.
Table 3.A3.Ranks of General Data Dissemination System Country Coordinators(Percent of total number of coordinators as of February 2007)
TotalSenior1Middle2Other
Total number of country coordinators100562519
By department:
African443186
Asia & Pacific14555
European3120
Middle East & Central Asia16961
Western Hemisphere231058
By institution:
Central bank238132
Ministry of finance251358
Statistics office523589
Sources: Country authorities; and IMF staff.

Senior management refers to head or deputy head of agency.

Middle management refers to head and deputy head of departments or divisions.

Sources: Country authorities; and IMF staff.

Senior management refers to head or deputy head of agency.

Middle management refers to head and deputy head of departments or divisions.

An analysis of the GDDS coordinators by rank and agency suggests that countries most often assign this task to senior-level managers from the national statistics office or central banks (Table 3.A3). Senior-level managers (defined as a head or deputy head of an agency) account for 56 percent of total GDDS coordinators, and coordinators from the statistical office account for 52 percent of total coordinators. This distribution is the same for countries in the Africa, Middle East and Central Asia, and Western Hemisphere regions.

This contrasts somewhat with the practice of SDDS countries, where advanced economies appoint mainly mid-level managers (defined as a head or deputy head of a department or division) in central banks and statistics office to coordinate the dissemination of data and other levels. The distribution of rank in all the regions is the same (Table 3.A4).

The different practices likely reflect the perception that the GDDS coordinator is mainly a strategic planner, while SDDS coordinators are responsible for the day-to-day operation of providing data and metadata updates for their countries.

Table 3.A4.Ranks of Special Data Dissemination Standard Country Coordinators(Percent of total number of coordinators as of February 2007)
TotalSenior1Middle2Other
Total number of country coordinators100283438
By department:
African2002
Asia & Pacific16583
European55161722
Middle East & Central Asia9352
Western Hemisphere19559
By institution:
Central bank4481620
Ministry of finance16655
Statistics office41141413
Sources: Country authorities; and IMF staff.

Senior management refers to head or deputy head of agency.

Middle management refers to head and deputy head of departments or divisions.

Sources: Country authorities; and IMF staff.

Senior management refers to head or deputy head of agency.

Middle management refers to head and deputy head of departments or divisions.

Appendix 3.3. Statistical Capacity Indicators
Score1Millennium

Development

Goals Included

in Metadata
1999200420052006PercentChange
African Department
Senegal707575757.10.0
Uganda5260677340.421.7
Mozambique626368689.77.9
Madagascar626253631.61.6X
Malawi5267606321.2–6.0
Mauritius606363635.00.0
Kenya65655362–4.6–4.6
Tanzania65656562–4.6–4.6
Ethiopia586363603.4–4.8
Rwanda4353536039.513.2
Gambia, The3860535339.5–11.7
Nigeria53405252–1.930.0
Republic of Congo25404050100.025.0
Namibia505352500.0–5.7
Botswana53655847–11.3–27.7X
Sierra Leone22273747113.674.1
Democratic Republic of
Congo423838432.413.2
Central African Republic40403838–5.0–5.0
Angola2733373529.66.1X
Liberia1317171838.55.9
Average47.652.552.154.113.73.1
Asia & Pacific Department
Bangladesh6073788033.39.6
Mongolia6070808033.314.3
Nepal5765737735.118.5
Vietnam5075757550.00.0
Sri Lanka5578727230.9–7.7
Cambodia32586365103.112.1
China65636562–4.6–1.6
Average54.168.972.37334.86.0
Middle East & Central Asia
Department
Albania6380808331.73.8
Pakistan6373808027.09.6X
Azerbaijan5075777754.02.7
Macedonia, FYR6773777511.92.7
Tajikistan4563727566.719.0
Georgia5072737346.01.4
Jordan77737773–5.20.0
Afghanistan10152528180.086.7
Average53.165.570.170.532.77.6
Western Hemisphere
Department
Guatemala4383808086.0–3.6
Bolivia6368707722.213.2
Venezuela5875777732.82.7
Nicaragua5282787544.2–8.5
Panama5875757529.30.0
Trinidad5858677020.720.7X
Honduras606255658.34.8
Dominica606870635.0–7.4
Average56.571.471.572.828.81.9
6 of 55
Total51.361.162.363.724.04.310.9%
Source: World Bank (2007).

Scale of 0–100. A score of 100 indicates that a country meets all the criteria.

Source: World Bank (2007).

Scale of 0–100. A score of 100 indicates that a country meets all the criteria.

References

A version of this chapter was discussed in an Informal Executive Board Seminar on February 1, 2008.

Most of the analysis is conducted for a sample of 55 GDDS participants. Appendix 3.1 describes the sample selection. Tables 3.A1 and 3.A2 provide breakdowns of the sample by region.

Although subscription to the SDDS is voluntary, observance of the standard by subscribers is mandatory, and the IMF monitors observance.

The IMF (1997, p. 3) paper that led to the adoption of the standard by the Executive Board placed some emphasis on the importance of data dissemination and on supporting “data improvements needed by users, including investors, and for progressing toward the SDDS for countries interested in doing so.”

GDDS and SDDS datasets are organized into “comprehensive frameworks” (e.g., national accounts) and “tracking indicators” (e.g., production index).

See Cady (2005) and Chapter 4 of this volume.

    Other Resources Citing This Publication