The Special Data Dissemination Standard: From Concept to Tool

The IMF’s work on standards and codes began in the wake of the 1994–95 international financial crisis, which underscored the role that information deficiencies play in contributing to market turmoil. In the mid-1990s, many countries had regulations detailing the financial information that enterprises must regularly disclose to inform shareholders and the public. However, no counterpart existed for countries’ disclosure of economic and financial data. As a result, financial markets, for example, relied on information that too often was incomplete and out of date and thus could adversely affect resource allocation and the pricing of country risks. In response to these circumstances, the international community asked the IMF—in line with its role in the international financial system—to set standards in the provision of economic and financial statistics to the public.³ In response to this request, the IMF established the SDDS in 1996 as the first of its core standards.

The SDDS is a voluntary disclosure standard designed to guide IMF members that have, or that might seek, access to international capital markets in the provision of their economic and financial data to the public. It prescribes that countries disseminate key macroeconomic data categories covering the real, fiscal, financial, and external sectors. (See Table 1.1 for a list of these data categories.) It also calls for descriptions of these statistics and of statistical practices with respect to access—including preannounced time schedules for data releases—integrity, and quality. The dimensions and elements of the SDDS are identified in Box 1.1.

Table 1.1.

Central Banks as Disseminators of Special Data Dissemination Standard (SDDS) Categories

(Percent of total SDDS countries)

¹Central banks are identified as disseminators when they are responsible for the compilation and dissemination of components of the data categories.

²Does not yet include external debt because as of January 2003 a transition period ending March 2003 was in effect for this new data category.

Table 1.1.

Central Banks as Disseminators of Special Data Dissemination Standard (SDDS) Categories

(Percent of total SDDS countries)

SDDS Data Categories		Percent
Real sector
	National accounts	14
	Production indices	10
	Labor market indices: employment, unemployment, and wages and earnings	4
	Consumer and producer price indices	6
Fiscal sector1
	General government operations	20
	Central government operations	20
	Central government debt	24
Financial sector
	Analytical accounts of the banking sector	100
	Analytical accounts of the central bank	100
	Interest rates	100
	Share price index	71
External sector2
	Balance of payments	71
	International reserves and foreign currency liquidity	90
	Merchandise trade	17
	International investment position	77
	Exchange rates	98
	Population	4

¹Central banks are identified as disseminators when they are responsible for the compilation and dissemination of components of the data categories.

²Does not yet include external debt because as of January 2003 a transition period ending March 2003 was in effect for this new data category.

View Table

Table 1.1.

Central Banks as Disseminators of Special Data Dissemination Standard (SDDS) Categories

(Percent of total SDDS countries)

SDDS Data Categories		Percent
Real sector
	National accounts	14
	Production indices	10
	Labor market indices: employment, unemployment, and wages and earnings	4
	Consumer and producer price indices	6
Fiscal sector1
	General government operations	20
	Central government operations	20
	Central government debt	24
Financial sector
	Analytical accounts of the banking sector	100
	Analytical accounts of the central bank	100
	Interest rates	100
	Share price index	71
External sector2
	Balance of payments	71
	International reserves and foreign currency liquidity	90
	Merchandise trade	17
	International investment position	77
	Exchange rates	98
	Population	4

¹Central banks are identified as disseminators when they are responsible for the compilation and dissemination of components of the data categories.

²Does not yet include external debt because as of January 2003 a transition period ending March 2003 was in effect for this new data category.

The other tier of the IMF’s data standards initiative—the General Data Dissemination System (GDDS)—followed in 1997. The GDDS is aimed at assisting countries to develop sound statistical systems as the basis for timely dissemination of data to the public. The purposes of the GDDS are to encourage member countries to improve data quality; to provide a framework for evaluating needs for data improvement and setting priorities in this respect; and to guide member countries in disseminating comprehensive, timely, accessible, and reliable economic, financial, and sociodemographic statistics to the public. The GDDS, while maintaining key features of the SDDS, was designed to assist IMF member countries that are not yet in a position to subscribe to the SDDS.⁴ By January 2003 the GDDS had more than 50 participants and, over the longer term, is on course to be an important catalyst in upgrading statistical capacity around the developing world.

The Dissemination Standards Bulletin Board (DSBB)—maintained by the IMF on the Internet—is the public face of the SDDS. From the time it was launched in September 1996, the DSBB grew to provide easily accessible information about the statistics and statistical practices (metadata) of the 53 SDDS subscribers as of January 2003.⁵ Further, it hosts the GDDS webpage and the Data Quality Reference Site, and serves as a gateway to more comprehensive national data sites as well as related sites maintained by regional and other international organizations.

A member country’s presence on the DSBB indicates that it subscribes to and observes certain tenets of good statistical citizenship. The metadata are useful in their own right, especially because they are presented in a common format: as key information about data quality, as background to help assess comparability across countries, and as a guide to the appropriateness of the data to the user’s intended application. Similarly, the data disseminated on the national summary data page have a common structure that facilitates access by financial markets and other data users. A national summary data page is an electronic webpage that disseminates the subscriber’s data described in the SDDS metadata and contains an electronic link to the DSBB.

Standards, by definition, must set some level of minimally accepted practice. But in recognition of differences in economic structures and institutional arrangements across countries, the SDDS has some flexibility that allows for adapting statistical best practice to local conditions. First, the SDDS marks certain categories for dissemination on an “as relevant” basis. For example, in an agricultural economy, an economy-wide measure of wages and earnings may not be a useful labor market indicator. Second, the SDDS identifies some data categories or components of data categories as “encouraged” rather than “prescribed.” These are typically data items that are of analytical value but may require a more extensive statistical system to produce the level of detail. Periodicity and/or timeliness exceeding the SDDS requirements are also encouraged for a number of data categories, including those of the financial and external sector. Finally, with respect to periodicity and timeliness, a subscribing country may exercise additional flexibility in two data categories (excluding international reserves and foreign currency liquidity, and external debt) while being deemed in observance of the SDDS.

A formal transition began with the opening of subscription in early April 1996 and ended on December 31, 1998. During this period, an IMF member could subscribe to the SDDS even if its dissemination practices were not yet fully in line with the SDDS at that time. This period gave subscribers time to adjust their practices according to a plan (referred to as a transition plan) to bring them into line with the SDDS. However, the time frame proved unattainable for most subscribers; at the end of 1999, only 13 of the 47 subscribers were in observance of the SDDS. A number of factors—including competing demands to meet the Y2K challenge and, for European subscribers, to launch the European Economic and Monetary Union—may have slowed progress. In the ensuing two years, the number of countries working off their transition plans—in other words, improving their statistics—increased. Improvements included meeting the SDDS requirements for coverage, periodicity, and timeliness, and providing advance release calendars. By the end of 2001, all but one of these subscribers were in observance of the standard.

The Role of Central Banks

By the mid-1990s, increasing financial liberalization and the internationalization of capital markets spurred many central banks to repurpose their communications and statistical policies to meet the needs of a diverse audience. In some central banks, this change has been explicit. For example, former Deputy Governor Y. V. Reddy of the Reserve Bank of India used several public occasions to explain the Reserve Bank’s communication policy and highlight the sources of information available from the Reserve Bank. He explains the rationale as being that “wider dissemination of information by all economic agents and transparency of policy and operations on part of the government and other regulatory authorities contribute significantly to efficient markets” (Reddy, 2001, p. 6).⁶

Inflation-targeting regimes, in particular, led to new ways of thinking about the information that should be provided to the public. An IMF seminar entitled the “Statistical Implications of Inflation Targeting” concluded that, although transparency itself is not an end, it is an important means to foster the credibility on which such regimes depend.⁷ The SDDS entered on this evolving scene and provided a stimulus for central banks to take on a broader role within their national statistical systems.

Central banks have, indeed, moved front and center on the SDDS stage. Of the 53 subscribers to the SDDS as of January 2003, central bank staffs had assumed the role of national SDDS coordinator in 30. These coordinators work across national institutions to facilitate integrating SDDS requirements into national statistical practices and compiling metadata during the subscription process. They actively promote continued observance of the standard through timely updates of metadata, dissemination of advance release calendars, and posting of statistical data on national summary data pages. For 22 subscribers, central bank websites host the country’s national summary data pages.

Central banks are key disseminators of macroeconomic data. As shown in Table 1.1, central banks are solely responsible for disseminating data for the SDDS financial sector, namely the analytical accounts of the banking sector and of the central bank, and interest rates. They also redisseminate data on share price indices. The dissemination of external sector data categories, with the exception of the merchandise trade category, also resides largely with central banks.

For the fiscal sector, a notable number of central banks have assumed the role of compilers and disseminators of the financing components of transactions of general and central government operations, as is typically the case in countries that do not compile an integrated set of government finance statistics. A smaller number of central banks also undertake to disseminate data on output (quarterly GDP and monthly production indices), prices, and the labor market. The emerging economies raise these percentages in categories that are not traditional central bank territory. At least vis-à-vis national statistical offices, relative resource availability appears to be an important reason central banks take on these additional responsibilities to satisfy increasing demands for these economic indicators, emanating both from within central banks and from policy planning agencies and financial markets.

For most central banks, SDDS subscription has introduced new aspects of transparency. For example, disclosure on the DSBB of the terms and conditions under which the data are produced and identification of internal government access to data before public release shed new light on institutions often known for their “veil of secrecy.” The need to provide summary methodology statements has stimulated many central banks to document and publicize information on their data compilation practices, including deviations from international methodological guidelines. In order to disclose data revision practices, central banks have also developed structured and transparent revisions policies. Many central banks introduced preannounced schedules for the release of data—a practice better known among national statistical offices. These schedules level the playing field for access to data by private markets and reinforce the professionalism of central banks’ statistical work.

SDDS subscription also provided strong incentives to move beyond the confines of the print media in the interest of enhancing data availability. Work on electronic dissemination included establishing central bank websites, redesigning websites to improve visibility of statistics, and increasing free access to some data series.

Impact of the SDDS

The SDDS has been in place long enough that it is reasonable to ask, “Has the SDDS made a difference to national statistical systems, and to central banks in particular?” The answer is clearly yes. Some of the evidence is quantitative in terms of the number of improvements introduced and the improved record of on-time performance in data dissemination. Other evidence comes from the expanded availability of external sector data that the international financial system has found it needs in the last few years. Additional proof lies in how financial market participants bring SDDS subscription into their judgments.

Balance sheet data are, of course, intrinsic to central banks’ operations. Therefore, it is not surprising that the frequency for disseminating the analytical accounts of the banking sector and the central bank were already in line with SDDS requirements—monthly dissemination for both categories⁸—when the SDDS was established. As shown in Table 1.2, subscribers needed only a few transition plans to move into observance of this requirement by 2000. However, they found it more challenging to meet the SDDS requirements for coverage and timeliness. Their success—that is, the improvements they made with respect to coverage and timeliness—can be gauged by the reduction over time in the number of transition plans. In 1998, six central banks had begun work to meet the coverage specifications for these two data categories. For example, several needed to distinguish between the private and public components of domestic claims and expand the institutional coverage of the data on the banking sector. Within two years, they had completed these improvements. They recorded more numerous improvements in the timeliness of the two categories; the number of transition plans declined from 36 in 1997 to three at the end of 2000.

Table 1.2.

Special Data Dissemination Standard Transition Plans for Major Financial Sector Data Categories, 1997–2000

Table 1.2.

Special Data Dissemination Standard Transition Plans for Major Financial Sector Data Categories, 1997–2000

Number of Transition Plans		1997	1998	1999	2000
Coverage
	Analytical accounts of the banking sector	4	6	2	0
	Analytical accounts of the central bank	3	6	1	0
Periodicity
	Analytical accounts of the banking sector	1	1	2	1
	Analytical accounts of the central bank	2	3	0	0
Timeliness
	Analytical accounts of the banking sector	21	21	6	2
	Analytical accounts of the central bank	15	14	5	1
	Number of subscribers (end of period)	37	46	47	47

View Table

Table 1.2.

Special Data Dissemination Standard Transition Plans for Major Financial Sector Data Categories, 1997–2000

Number of Transition Plans		1997	1998	1999	2000
Coverage
	Analytical accounts of the banking sector	4	6	2	0
	Analytical accounts of the central bank	3	6	1	0
Periodicity
	Analytical accounts of the banking sector	1	1	2	1
	Analytical accounts of the central bank	2	3	0	0
Timeliness
	Analytical accounts of the banking sector	21	21	6	2
	Analytical accounts of the central bank	15	14	5	1
	Number of subscribers (end of period)	37	46	47	47

The improvements in timeliness in these data categories, as well as others, illustrate how the SDDS focused attention and stimulated creative solutions. In the mid-1990s, central banks found that dissemination of their data, including monthly central bank balance sheets, was often hostage to the hard-copy publication cycle. In many cases, the monthly central bank bulletin was the major vehicle for dissemination. The banks improved timeliness by using electronic media and/or press releases, which permitted them to disseminate data the same day that the figures were ready to start down the path toward eventual publication in the monthly bulletin. The case of South Africa is probably typical. The South African Reserve Bank supplemented its Monthly Release of Selected Data by releasing daily current market rates on the bank’s Internet website, enabling the bank to disseminate interest and exchange rates and share price indices daily.⁹

This chapter noted earlier that many central banks began publishing schedules for their upcoming data releases. The obvious next question is whether they would actually release according to that schedule. Table 1.3 presents the track record for selected data categories in terms of the percentage of quarterly and monthly releases that are on time, using a comparison of the third quarters of 2000 and 2002.¹⁰ As seen in the table, the ability to release on time, according to the preannounced release schedule, improved dramatically. For all the monthly and quarterly data categories for which the central bank is responsible, at least 94 percent were released on time in the third quarter of 2002, up from 64 to 76 percent two years earlier. Also, the percentage of releases of the monthly reserves template and quarterly balance of payments statistics consistently exceeded the all-categories average for monthly and quarterly data, respectively.

Table 1.3.

Percentage of Monthly and Quarterly Data Disseminated on Time

(Quarterly averages)

Table 1.3.

Percentage of Monthly and Quarterly Data Disseminated on Time

(Quarterly averages)

Selected Data Categories		2000/Q3	2001/Q3	2002/Q3
All monthly data categories		71.7	93.9	94.7
	Analytical accounts of the banking sector	72.7	91.6	95.9
	Analytical accounts of the central bank	64.0	94.5	94.0
	Gross reserves	75.9	95.9	94.0
	Reserves template	73.8	96.6	98.7
All quarterly data categories		65.2	91.0	95.5
	Balance of payments	66.6	91.9	98.8

View Table

Table 1.3.

Percentage of Monthly and Quarterly Data Disseminated on Time

(Quarterly averages)

Selected Data Categories		2000/Q3	2001/Q3	2002/Q3
All monthly data categories		71.7	93.9	94.7
	Analytical accounts of the banking sector	72.7	91.6	95.9
	Analytical accounts of the central bank	64.0	94.5	94.0
	Gross reserves	75.9	95.9	94.0
	Reserves template	73.8	96.6	98.7
All quarterly data categories		65.2	91.0	95.5
	Balance of payments	66.6	91.9	98.8

At the inception of the data standards initiative, the IMF recognized that the SDDS would have to evolve with changing circumstances. As it turned out, the transition period accorded to early subscribers was still in place when the need arose to supplement the data categories of the original SDDS. The financial crises of 1997–98 focused attention on the need for more comprehensive data on the external sector.¹¹ To encourage countries to meet that need, the IMF introduced three sets of enhancements to the SDDS: in December 1998, a time frame was established for disseminating annual international investment position (IIP) statistics;¹² in March 1999, the international reserves and foreign currency liquidity (reserves template) was added to provide more detailed and expanded specifications; and in March 2000, external debt was included as a new data category. In expanding the scope of the SDDS, the IMF recognized that subscribers would need time to build up their capacity to meet the data coverage, periodicity, and timeliness for the new or expanded data categories. Accordingly, the IMF provided a transition period for disseminating data on the reserves template, IIP, and external debt ending in March 2000, December 2001, and March 2003, respectively.

International reserves—long a key variable in central bankers’ databases—is a good case study. The financial crises of 1997–98 precipitated increased pressure on central banks to improve the accountability of their reserves management and the transparency of their reserves statistics.¹³ In response, the IMF strengthened the SDDS by enhancing the international reserves data category to include dissemination of monthly data on both gross international reserves and foreign currency liquidity in a single framework. Developed in consultation with central banks, as well as with SDDS subscribers and data users, the reserves template requires a substantially expanded set of information. It covers (1) the amount and composition of official reserve assets; (2) other foreign currency assets held by the monetary authorities and the central government; (3) short-term foreign currency obligations; and (4) related activities (such as financial derivatives positions and guarantees extended by the government for quasiofficial and private sector borrowing) of the monetary authorities and the central government that can lead to drains on reserves and other foreign currency assets. The IMF issued operational guidelines designed to assist countries in preparing reserves template data in October 1999. By the end of the transition period, most central banks were disseminating the reserves template in accordance with the requirements of the standard.¹⁴ Within 18 months after the decision to include the reserves template in the SDDS, 42 subscribers had begun disseminating the reserves template.

Similarly, central banks led national efforts to meet the SDDS requirements to disseminate annual data on the international investment position. This balance sheet of the stock of external financial assets and liabilities of an economy is a valuable statistical product for financial markets.¹⁵ As noted by Price (2000, p. 7), “… in our own sovereign ratings analysis at Fitch, we have often struggled to derive comprehensive up-to-date debt numbers from information national authorities were able to give us, but that task has become less difficult as two-thirds of the countries we rate now publish an international investment position.”

The external debt data category that was added calls for disseminating external debt of the general government, the monetary authorities, the banking sector, and all other sectors. Data should also be broken down by maturity (short- and long-term), on an original maturity basis, and by instrument. The SDDS prescribes quarterly periodicity and timeliness. Subscribers were expected to commence actual dissemination in the third quarter of 2003—for the first reference quarter after the end of the transition period—but a number of countries were ahead of the schedules. As of January 2003, six subscribers were already disseminating the data in accordance with the SDDS requirements. Following the introduction of the category, an interagency Task Force on Finance Statistics, chaired by the IMF, completed the External Debt Statistics: Guide for Compilers and Users (External Debt Guide).¹⁶ The guide, which has been used as the basis for extensive training for national compilers, provides a scheme for classifying external debt by instruments and sectors. This scheme has evolved into a presentation table for the gross external debt position. Data disseminated using this presentation table, and employing the concepts outlined in the External Debt Guide, provide a comprehensive and informed picture of the gross external debt position for the whole economy.¹⁷ In line with their earlier work on the external sector, central banks are expected to collaborate with national debt management agencies in facilitating countries’ observance of the SDDS requirements for this data category.

When the SDDS was emerging as a concept, its architects envisaged that, aside from the intangible benefit of giving a country a reputation for good statistical citizenship, subscription would bring tangible benefits in the form of higher credit ratings, reduced spreads, and increased flows of investment. However, evidence is elusive; the counterfactual situation—or example, what the spread would have been if a subscriber had not subscribed—is not observable. It is worthwhile to note that subscription to the SDDS is a variable in country risk assessments undertaken by private institutions, including JPMorgan. Furthermore, an econometric study of the impact of SDDS subscription on sovereign risk spreads, undertaken by the International Institute of Finance, estimates that SDDS subscription might reduce such spreads by as much as 300 basis points.¹⁸

Standing Still Means Moving Backward

When the IMF established the SDDS in March 1996, the Executive Directors of the IMF emphasized its design and implementation should be both flexible and evolutionary, in order to maintain its standing as an international standard that embodies best practices for data dissemination. Since then, the SDDS has evolved to adapt to changing circumstances by encompassing data quality improvements emanating from the adoption of new and internationally-accepted statistical methodologies and by responding to the data needs associated with assessing external sector vulnerabilities.

The IMF’s work on international standards took a leap forward in 1998 when, in discussion of the reform of the international financial architecture, it became clear that international standards were limited in value unless countries implemented them. In this context, the IMF was asked to assess countries’ observance of international standards and codes in its area of expertise. These assessments appear as Reports on the Observance of Standards and Codes (ROSCs)¹⁹ and are targeted primarily at a private sector audience on the premise that if the findings can be fed into country risk assessments, markets may discipline countries into adhering to standards.

Within the ROSC program, the first data modules assessed data dissemination practices—that is, they determined the extent to which countries disseminated data and information about their statistical practices as called for by the SDDS (or GDDS). Subsequently, the IMF recognized that an assessment that dealt only with dissemination, without asking anything about the quality of what was disseminated, was not adequate for users. Thus, the IMF enhanced the data module to include an assessment of the quality of the data that was being disseminated. This complementary assessment is done using the Data Quality Assessment Framework.²⁰ This framework covers much the same ground as the SDDS (and the GDDS), but introduces a structure for assessing the extent to which countries have the prerequisites of data quality and follow international best practices with respect to integrity, methodological soundness, accuracy and reliability, serviceability, and accessibility (Carson, 2001).

In recognizing the importance of methodological soundness as a dimension of data quality and as an essential complement to and outgrowth of data standards, the IMF has also intensified efforts to assist countries in improving the quality of their data. This effort includes developing internationally-agreed-upon guidelines on statistical methodology. In addition to the guides on external debt and on international reserves and foreign currency liquidity mentioned earlier in this chapter, the Manual on Monetary and Financial Statistics—of particular interest to central banks—was launched in 2000.

Another outgrowth of the SDDS has been the ongoing work on statistical metadata and the development of an open system for disseminating and exchanging statistical information on the Internet.²¹ With their standardized format and comprehensive coverage, SDDS metadata templates have attracted wide usage. The European Central Bank, Eurostat, and central banks of some countries that do not currently subscribe to the SDDS have adopted the format for their metadata presentations on the Internet. These moves bring greater homogeneity and structure to metadata dissemination by central banks and provide a platform for introducing portal and data mining capabilities on the DSBB.

The IMF now encourages a larger SDDS subscription base, setting a goal of including more countries in order to help them position themselves for access to international capital markets. Thus, as more subscriptions are sought, an appropriate balance must be struck. On the one hand, interest will increase in making the SDDS comprehensive and inclusive of newly emerging data sets, such as financial soundness indicators. On the other hand, a need persists to be realistic about demands on statistics-producing agencies—many of which are being stretched by national, regional, and international calls for more, faster, and better data. Various factors, both economic and political, impinge on this balancing act, which may well determine the course of the SDDS in the future.