3. Dimensions of the General Data Dissemination System
- International Monetary Fund
- Published Date:
- July 2007
3.1 This section introduces the GDDS nomenclature, describes the four dimensions of the system, and describes “good practices” for each dimension.
3.2 The GDDS employs the following technical terms.
Dimension refers to one of the four dimensions of the system—data, quality, integrity, and access. Each dimension provides an insight into how effectively a statistical system is achieving its objectives. In the GDDS, the data dimension relates to statistical products, while the other three relate to the conditions under which the products are compiled and disseminated as seen from a user’s perspective. Underlying the concept of dimensions in the GDDS is the identification of practices that are conducive to ensuring that the data produced are of a high standard. For each dimension, two to four such practices are specified (see Box 2.1). Thus, the dimensions serve as vehicles to facilitate the grouping of practices that are especially helpful in promoting data quality, which is the ultimate goal of the GDDS. The particular practices that are associated with each sector were determined with reference to the UN’s Fundamental Principles of Official Statistics.
Comprehensive framework1 is an integrated statistical approach that provides well-recognized measures of developments in a macroeconomic sector. Each of the comprehensive frameworks has a core framework, and two of them include encouraged extensions. The encouraged extensions should be viewed as less pressing priorities when statistical agencies are faced with tight resource constraints.
Core framework is the principal data set of the comprehensive framework for a macroeconomic sector. It reflects a method of organizing data that is considered to be especially useful for economic analysis. Each core framework has distinctive technical features that are relevant to the data requirements of the macroeconomic sector concerned.
Data category refers to data that are considered to be of special importance in their own right. Certain data categories are components of core frameworks, while others exist independently. For the macroeconomic sectors, there are core indicators for each data category and, for most, encouraged categories and/or components. As is the case for the encouraged extensions for the comprehensive frameworks, the encouraged categories should be viewed as less pressing priorities when statistical agencies are faced with tight resource constraints.
Core indicator refers to (i) an indicator for the comprehensive framework; (ii) an additional tracking indicator; or (iii) other indicators relevant to the macroeconomic sector. Each macroeconomic sector has one or more of each of these types of core indicators.
Basic component refers to the data sets that the GDDS recommends be compiled and disseminated for each of the four sociodemographic data categories.
A description follows of the System’s four dimensions—data, quality, integrity, and access.
The Data Dimension
Structure of the Data Dimension
3.3 A hierarchical structure is used in elaborating the data dimension of the GDDS. In the first stage, a distinction is made between the four macroeconomic sectors and the sociodemographic data. Next, the economic and financial data are classified to one of the four macroeconomic sectors (real, fiscal, financial, and external), and the sociodemographic data are broken down into four data categories (population, health, education, and poverty), with basic components identified for each category and encouraged extensions specified for some.
3.4 In the next stage, a distinction is made in the data for each macroeconomic sector between comprehensive frameworks and data categories and indicators. A core framework is then specified for each comprehensive framework and, for two of the macroeconomic sectors, an encouraged extension to the core framework is also specified.
3.5 Next, data categories are specified within each macroeconomic sector and core indicators are identified for each—for some of these data categories, encouraged extensions are also specified. The GDDS also associates directly with 23 out of 35 MDG indicators, mostly in the sociodemographic sector. In the final stage, the data are viewed from three separate perspectives—coverage, periodicity, and timeliness.
3.6 The specifications for the data dimension of the GDDS, shown in Table 3.1 (at the end of this chapter), are neither requirements nor prescriptions that must be observed to participate in the system. Rather, they are good-practice objectives to be approached over time. A participating country should indicate in the metadata its short- and medium-term plans with respect to implementing these good-practice objectives. In developing these plans, core frameworks and core indicators are recommended as the first priorities. In addition, the encouraged extensions noted in Table 3.1 are recommended as second priorities. Further, Table 3.1 designates certain categories to be disseminated on an “as relevant” basis.
|A. Comprehensive Frameworks—Macroeconomic and Financial Sectors|
|B. Data Categories and Core Indicators—Macroeconomic and Financial Sectors|
|C. Sociodemographic Data|
3.7 In its coverage of data sets, the GDDS focuses on the data that are considered most important in evaluating performance and policy in four macroeconomic sectors—real, fiscal, financial, and external—as well as complementary sociodemographic data that shed light on economic development and structural change. The GDDS addresses the development and dissemination of a full range of data by (i) presenting objectives for the development and dissemination of comprehensive frameworks in each of the four macroeconomic sectors and (ii) recommending the development and dissemination of indicators for these four sectors and of basic components in the area of sociodemographic data, with periodicity and timeliness appropriate to the circumstances of each participating country.
3.8 While, for the majority of participating countries, emphasis should first be placed on producing good-quality data in the core areas before substantial resources are devoted to developing the encouraged areas, some countries may wish to place a higher priority on developing certain encouraged areas where these are of particular importance for analysis or policymaking.
3.9 Section A of Table 3.1 presents the broad objectives for long-term development of the comprehensive frameworks for each macroeconomic sector and the recommended periodicity and timeliness for compilation and dissemination of relatively complete sets of information. For each framework, the objective is to achieve the widest coverage possible, using an appropriate statistical framework (sometimes referred to as an “analytical framework”) and classification scheme.
3.10 Section A of Table 3.1 provides illustrations of the aggregates and balances that might be developed within the comprehensive frameworks, but the underlying objective is the development of complete data sets rather than specific indicators. Strong emphasis is placed on the use of international guidelines in specific areas of macroeconomic statistics. A listing of these guidelines is provided in Appendix II.
3.11 In addition, Section A of Table 3.1 presents a set of objectives for national accounts (real sector), central government operations (fiscal sector), the depository corporations survey (financial sector), and balance of payments (external sector). In the fiscal sector, extensions of coverage are encouraged to include general government or public sector operations and, for balance of payments, extensions are encouraged to include the international investment position (IIP). There is no comprehensive framework for sociodemographic data because of the great diversity of areas covered.
3.12 Section B of Table 3.1 specifies the data categories for each of the four macroeconomic sectors and identifies the following core indicators for each sector: (i) indicators for the comprehensive frameworks—nominal and real GDP for national accounts, central government aggregates for central government operations, broad money and credit aggregates for the depository corporations survey, and balance of payments aggregates for the balance of payments; (ii) additional indicators that permit tracking of the core frameworks for each comprehensive framework; and (iii) other indicators relevant to the sector, which will often be in the form of a price variable, such as interest rates and exchange rates. For certain data categories, additional aggregates and/or components are encouraged.
3.13 Section C of Table 3.1 presents the four sociodemographic data categories. Basic components are specified for each category, and each category also has encouraged extensions. Related indicators of MDGs are also specified for each data category.
3.14 While most of the data called for by the GDDS are produced by official national agencies, the system provides for the inclusion of some data categories that are produced by private organizations (at least in some countries) where such data are the subject of official redissemination. The inclusion of privately compiled data is warranted in the interest of obtaining a more complete picture of the economy and more consistent coverage across countries. However, including such data requires some adaptation in the responsibilities that the official (disseminating) agency has with respect to some elements of access by the public, integrity, and data quality.
3.15 Periodicity refers to the frequency of compilation of the data. The periodicity recommended by the GDDS for a particular data set is determined by several factors, including the needs of analysis and the ease of observation or compilation. Although these factors are not the same for specific data sets across countries, in practice there is rather widespread agreement about good practice regarding the highest frequency of compilation for the comprehensive frameworks and many of the indicators in the GDDS.
3.16 In recognition of the longer period required to compile and disseminate complete sets of data, the periodicity recommended for comprehensive frameworks is in many cases longer than that recommended for indicators. In a number of cases, the periodicity for comprehensive frameworks and indicators is the same, but different timeliness is specified.
3.17 Periodicity is usually expressed in terms of divisions of the calendar. For flow data, data compiled at roughly 30-day intervals, for example, usually represent a calendar month. Stock data are compiled as of a point of time, often the end of the week, month, quarter, or year.
3.18 The recommendations for periodicity that are shown in Table 3.1 for flow data are expressed in terms of the longest interval to be represented by a single data observation and, for stock data, in terms of the longest interval between observations. The periodicity recommended for the comprehensive frameworks is annual except in the case of the depository corporations survey where existing good practice for a wide range of countries is monthly periodicity.
3.19 For data categories and indicators, periodicities are recommended to reflect existing good practice across countries. For national accounts aggregates, central government debt and for balance of payments, annual periodicity is recommended, but quarterly periodicity is encouraged.
3.20 There may be some points of interpretation with respect to periodicity. For example, although a quarter is usually viewed as three months, data covering successive intervals of 13 weeks would be considered quarterly. Annual data may be for calendar years or fiscal years, with varying beginning dates.
3.21 Timeliness refers to the lapse of time between the end of a reference period (or a reference date) and dissemination of the data. It reflects many factors, including some that are related to institutional arrangements, such as the preparation of accompanying commentary and printing.
3.22 In recognition of the diversity of countries covered by the system, the objectives for timeliness that are presented in Table 3.1 are set out in terms of ranges. The short end of the timeliness range corresponds generally to the SDDS timeliness requirements for a given data set while the long end of the range reflects good practice across a broad group of countries.
3.23 The short end of the range also reflects current best practices, and these are appropriate targets for countries that wish to use the GDDS as a stepping stone to subscription to the SDDS, but these timeliness criteria may not be appropriate for all countries particularly in the near term.
3.24 In all cases, the principle of the system is that major improvements in timeliness should not be attempted if other aspects of data quality will be significantly compromised. Rather, incremental improvements in timeliness over time, consistent with maintaining and improving other aspects of data quality, should be the objective.
3.25 There may be occasions on which, for certain data sets, timeliness targets may be difficult to meet. For data that are based on the accounting records of government and enterprises, the month or quarter in which the fiscal or financial year is closed may present special difficulties. As a result, the data for those periods may not be available as quickly as for the other periods. Differences in the numbers of days (and business days) in the various months and the incidence of holidays (which differ from year to year in many countries) are among the other factors that may also affect timeliness. Continuing the discussion of the data dimension, the following defines comprehensive frameworks for each macroeconomic sector.
Specifications for Comprehensive Frameworks
Real sector—national accounts
3.26 The comprehensive framework for the real sector comprises the core framework of National Accounts. The objective is the compilation and dissemination of data that cover the widest possible scope of economic activity. Current practices as well as plans for expanding coverage should be presented in the metadata. For example, if certain areas of production (for example, military output, mineral production) are not included in existing estimates, this practice should be indicated and plans to improve coverage should be developed and disseminated. If informal sector production is significant, the present estimation techniques should be described and any plans for improving these techniques noted.
3.27 The use of internationally agreed guidelines for the development of national accounts is recommended. The two most commonly used international guidelines for national accounts are the System of National Accounts 1993 (1993 SNA) and the European System of Accounts 1995 (ESA 1995).2 For the development of quarterly national accounts, countries should also use the IMF’s Quarterly National Accounts Manual.3 It is recognized that each country must design its own path for national accounts development, based on its economic and institutional structure, analytical and policy needs, and human and financial resources.
3.28 Complete sets of national accounts, such as those described in the 1993 SNA, are currently compiled by only a few countries. The complete set of accounts includes not only major aggregates for the total economy (GDP, Gross National Income, Gross Disposable Income, Saving, and Net Lending/Net Borrowing) but also full transaction accounts for institutional sectors and balance sheets for the total economy and institutional sectors. For many countries, such complete accounts exceed current analytical and policy requirements and/or cannot be compiled with available resources.
3.29 The GDDS, therefore, does not recommend that countries develop full sets of national accounts; it recommends, rather, that countries determine their specific medium-term needs for national accounts and develop realistic plans for implementing those parts of the complete set of accounts that are appropriate.
3.30 For many countries at the early stages of national accounts development, priority might be assigned in the medium term to developing major aggregates for the total economy and improving their quality. Longer-term objectives might include the development of accounts for those sectors that are particularly important. Countries in more advanced stages of national accounts development and with more complex data requirements might place the medium-term priority on the development of sectoral accounts and balance sheets.
3.31 The GDDS recommends that annual data on the core framework of national accounts be compiled and disseminated with a timeliness of 10 to 14 months.
Fiscal sector—central government operations
3.32 The comprehensive framework for the fiscal sector comprises a core framework for Central Government Operations and an encouraged extension relating to General Government or Public Sector Operations.
3.33 The GDDS focus for the core framework of central government operations is on the production and dissemination of comprehensive data on transactions and on debt, emphasizing (i) coverage of all central government units; (ii) use of an appropriate analytical framework; and (iii) development of a full range of detailed classifications.
3.34 Separate comprehensive framework tables should be prepared for central government transactions and for central government debt. All units of central government should be included. These units consist of (i) all ministries, agencies, and other units whose transactions are covered by the central government budget; (ii) all units with their own resources or direct sources of financing (extrabudgetary units) whose magnitude is significant relative to the size of the budget; and (iii) social security funds operating at the national level.
3.35 The identification of central government units may be based on definitions of the government sector provided in the 1993 SNA, A Manual on Government Finance Statistics (GFSM 1986), or the Government Finance Statistics Manual 2001 (GFSM 2001).4 The general government sector may be subsectored by treating social security funds as a separate subsector or by consolidating social security funds at the level of government at which they operate. For fiscal analysis, the latter subsectoring is usually preferred, particularly where social security funds have surpluses or are in deficit. The metadata for the comprehensive frameworks should describe the coverage of the widest measures currently disseminated and plans to expand this coverage if necessary.
3.36 The GDDS recommends the use of an appropriate analytical framework to define and present central government transactions data but does not prescribe a particular framework. Any fully identified national definitions of aggregates of revenue and expenditure (expense) and balancing items, such as current account and overall deficit, or operating surplus and net lending/borrowing, may be used. The framework and definitions provided in either the GFSM 1986 or the GFSM 2001 are recommended.
3.37 If a different analytical framework is used, it would be useful to indicate in the metadata how major aggregates and balancing items differ from those presented in those manuals. Further, plans for improving the analytical framework may be based on the recommendations of those manuals. These plans should be described in the metadata.
3.38 The data disseminated should include detailed classifications of the major aggregates in the analytical framework. Specifically recommended are breakdowns of (i) tax and nontax revenue; (ii) expenditure by function (i.e., by purpose); (iii) expenditure by economic type, with a separate identification of the major components of current and capital expenditure; and (iv) financing. Countries implementing the GFSM 2001 should provide breakdowns of revenue, expense, net acquisition of nonfinancial assets, net acquisition of financial assets, and net incurrence of liabilities.
3.39 Financing data should, wherever possible, distinguish domestic from external financing on a residency basis—domestic financing data should distinguish between financing from the banking system and from other sources. Data on financing should be supplemented, as relevant, with breakdowns by financial instrument and/or by currency of denomination.
3.40 Classification by residency is analytically useful and promotes consistency of fiscal and balance of payments data. In countries where government financing operations include the issuing of securities, a residency breakdown may be difficult to compile. In these cases, the primary breakdown of financing may be by currency of denomination or type of instrument, but efforts should still be made to develop a residency breakdown. The metadata should describe current classifications that are produced and disseminated and plans to improve these classifications.
3.41 Data for central government debt should be comprehensive and should include the liabilities of all institutional units that are part of central government. All liabilities in the form of securities, loans, and deposits should be included. Debt of other units that is guaranteed by the central government should also be compiled and disseminated when the amounts involved are significant. The GDDS recommends that countries provide a breakdown of debt by foreign and domestic components according to residence. Other breakdowns recommended by the GDDS are identification of debt by type of holder or type of instrument.
3.42 Classification and definition of debt according to the guidelines of the GFSM1986 and GFSM 2001, External Debt Statistics: Guide for Compilers and Users5 (for external debt), or regional guidelines are encouraged; the use of such guidelines or plans to use them should be noted in the metadata.
3.43 Data for general government or public sector operations is an encouraged extension of the comprehensive framework for the fiscal sector, particularly for countries where state (in countries with a federal structure) and local levels of government and/or public enterprise operations are important for fiscal analysis. In the long run, the GDDS encourages all countries to develop appropriate and comprehensive measures of general government, but for countries where central government data are deficient in terms of coverage, analytical framework, or quality, the priority should normally be placed on improving the data for the central government.
3.44 The general government covers all governmental units (including social security systems) whether operating at the central level, state/province/region level, or local level. The public sector may be defined in a variety of ways. It may include the general government sector plus nonfinancial corporations (the nonfinancial public sector) or it may also include certain public financial institutions (for example, development banks). The nonfinancial public sector is the preferable coverage. In countries where the public sector (however defined) is the main focus of analysis and policy, public sector operations should be the focus of this encouraged extension; in other countries general government may be more appropriate.
3.45 The analytical framework and classifications for general government data should be the same as for central government, although it is recognized that data for state (in federal systems) and local governments may not be available in the same degree of detail as for central government. Data for public enterprises should include overall borrowing and details of financing.
3.46 The GDDS recommends dissemination of annual data on central government operations and general government or public sector operations, with a timeliness of six to nine months.
Financial sector—depository corporations survey
3.47 The comprehensive framework for the financial sector is the core framework relating to the Depository Corporations Survey. The objective for this core framework is the compilation and dissemination of comprehensive data emphasizing (i) coverage of all depository corporations (banking and other deposit-taking institutions); (ii) use of an appropriate analytical framework; and (iii) development of classifications of external assets and liabilities, domestic credit by sector, and components of money (liquidity) and nonmonetary liabilities.
3.48 The appropriate coverage for the GDDS is that group of financial institutions whose liabilities include the money measure(s) most frequently used for monetary analysis. Depository corporations are defined in the 1993 SNA and in the Monetary and Financial Statistics Manual (MFSM)6 to include all financial institutions that incur liabilities in the form of deposits and/or close substitutes for deposits where these deposits or deposit substitutes are included in national measures of broad money. The coverage is, therefore, built around not only a group of institutions but also a country-specific measure of money. This coverage comprises the central bank and all other depository corporations.
3.49 The broad money measure normally includes transferable deposits; time, saving, and fixed-term deposits; participations in money market mutual funds; and short-term securities, such as negotiable certificates of deposit, that have characteristics similar to deposits.
3.50 In countries where several money measures are compiled, the GDDS coverage should include the institutions covered by the broadest measure. In such cases, it would also be useful to describe in the metadata whether other financial institutions exist that incur liabilities in the form of nontransferable deposits or close substitutes for them. Countries that use liquidity measures that include liabilities (for example, government treasury bills, corporate commercial paper) of nonfinancial sectors may describe such measures and the institutional coverage in their metadata.
3.51 The GDDS recommends that the statistical framework contained in the MFSM be used for the depository corporations survey. Stock data (outstanding assets and liabilities) should be disseminated, but transaction data may also be disseminated. The analytical framework should distinguish between external and domestic positions on the basis of the residency criteria as defined in the 1993 SNA and BPM5.
3.52 Gross data on claims on and liabilities to nonresidents should be disseminated. Domestic credit should be classified according to the debtor sector and, ideally, should separately identify claims on government (central, state, and local, as relevant), claims on public nonfinancial corporations, and claims on the private sector. The disseminated data on money measures should include breakdowns by type of monetary instrument, and data on nonmonetary liabilities (for example, long-term securities) should also be disseminated.
3.53 The depository corporations survey should be compiled on a monthly basis because this frequency reflects current good practice across a broad range of countries, and should be disseminated within two or three months of the end of the reference month.
External sector—balance of payments
3.54 The comprehensive framework for the external sector comprises a core framework relating to balance of payments and an encouraged extension for the international investment position (IIP). The objective of the core framework is the compilation and dissemination of comprehensive data on the main aggregates and balancing items of the balance of payments, including, for example, imports and exports of goods and services, trade balance, income and transfers, current account balance, reserves and other financial transactions, and overall balance, with detailed components as relevant.
3.55 Compilation of data according to the BPM5 is recommended.
3.56 The broad objective for compilation and dissemination of balance of payments data is to produce all the standard components of the BPM5 that are relevant to a country’s circumstances. Current, capital, and financial transactions should be distinguished. Components recommended for dissemination within the current account include (i) imports and exports of goods and services; (ii) income transactions—that is, income receipts and payments, both with respect to compensation of employees and investment income, with the latter including income from direct investment, portfolio investment, and other investment; and (iii) receipts and payments with respect to current transfers.
3.57 The capital account should include capital transfers, as relevant, and the financial account should separately identify transactions with respect to direct investment, portfolio investment, financial derivatives, other investment, and reserves.
3.58 Additional breakdowns of debt securities and loans, within portfolio investment and other investment, respectively, by currency of issue and by original maturity (short-term versus medium- and long-term using classifications by instrument), would be highly desirable for an overall view of external debt.
3.59 The metadata for the balance of payments should describe present data compilation and dissemination practices and plans to implement the analytical framework and classifications of the BPM5. In countries where the BPM5 is not used, the metadata should describe the present methodology and plans to move toward conformity with the BPM5.
3.60 The GDDS encourages extensions of external sector data with respect to the IIP, which is increasingly recognized as a useful framework in which to develop an integrated picture of a country’s stock of external financial assets and liabilities.7 The System encourages work toward compiling component detail according to the BPM5—direct investment; portfolio investment, including a breakdown into equity and debt; financial derivatives;8 other investment; and reserves—and disseminating components of the framework as appropriate and feasible.
3.61 Data for the balance of payments core framework, and for the IIP, should be disseminated on an annual basis and within six to nine months of the end of the reference year (fiscal or calendar year). Continuing the discussion on the data dimension, the following paragraphs define data categories and indicators for each macroeconomic sector.
Specifications for Data Categories and Indicators
National accounts aggregates
3.62 The data category corresponding to the comprehensive statistical framework for the real sector is national accounts aggregates. The core indicators for national accounts aggregates are GDP at nominal and real (price-adjusted) levels. The GDDS also encourages the production and dissemination of indicators on gross national income, capital formation, and saving. The system recommends that indicators for national accounts aggregates be produced on an annual basis and disseminated within six to nine months of the end of the calendar or fiscal year, as relevant. The compilation and dissemination of quarterly national accounts aggregates are encouraged.
3.63 The GDDS does not recommend the specific components of the national accounts that should be compiled and disseminated but encourages breakdowns of GDP by major expenditure category, by productive sector (industry), or both. The breakdowns into component items should be those relevant for the country concerned.
3.64 Classification of national accounts aggregates according to the 1993 SNA (or a regional counterpart) is encouraged. The compilation of quarterly GDP estimates is also encouraged. However, if a country with resource constraints and significant weaknesses in its annual national accounts has not yet improved the quality (scope, coverage, and methodology) of its annual national accounts, it would not be advisable for it to develop quarterly estimates until it has implemented a plan for improving its annual national accounts. In this case, priority should be given to developing a program for updating the base or benchmark year of national accounts and implementing the most current international standards.
3.65 The data category that is intended to track productive activity on a more frequent basis is a production index or, if more appropriate, several production indices. The index (and any components of it) or selection of indices that is relevant will depend on a country’s economic structure—industrial production in some countries, commodity production (for example, petroleum) in other countries, or agricultural production in still others.
3.66 The index or selection of indices that is chosen should be the one used within the country as a useful indicator. For instance, it would not be advisable for a mainly agricultural country with resource limitations and shortcomings in its national accounts to invest its scarce resources in developing a manufacturing production index until it improved the quality of its agricultural indicators and, hence, the quality of its national accounts.
3.67 The GDDS does not recommend the dissemination of retail or wholesale sales indices as substitutes for production indices because sales indices may not track GDP developments, owing to the impact of inventory buildups or drawdowns. However, if sales indicators are considered important for analysis, the metadata for the production index/indices may note this.
3.68 The GDDS recommends a monthly measure for the manufacturing or industrial index. The periodicity of other production indices is specified as “as relevant” to a country’s structure and production cycles. For instance, in some countries, such as those where seasonal crop production is important, production may not be well represented by a monthly index. The recommended timeliness of dissemination is 6 to 12 weeks from the end of the reference period for all indices.
3.69 The GDDS recommends that a consumer price index (CPI) be compiled on a monthly basis and disseminated within one to two months of the end of the reference month. The System encourages the compilation and dissemination of a producer price index (PPI) with the same periodicity and timeliness as consumer price indices.
3.70 The use of the international guidelines contained in the Consumer Price Index Manual and the Producer Price Index Manual, or the use of regional guidelines, is encouraged.9
Consumer price index
3.71 The CPI index, sometimes also referred to as a retail price index, indicates the part of the variation over time in a household expenditure aggregate that can be attributed to price movements. The expenditure aggregate comprises expenditures on goods and services that households acquire, pay for, or use for purposes of consumption. The CPI is typically measured by a Laspeyres index of the prices of a fixed set of goods and services items, often referred to as a “fixed basket,” whose weights are the item shares in the expenditure aggregate from a given, generally annual, historical period.
3.72 Countries may prepare several indices, differing, for example, with respect to geographic coverage, reference population, and item coverage. For the GDDS, the metadata should describe the characteristics of the index most widely used in the country but may also note other existing CPIs.
3.73 The GDDS does not recommend the dissemination of any component, or subindex, detail. However, the major items of the United Nations Classification of Individual Consumption by Purpose (COICOP),10 which is also recommended in the 1993 SNA, provide a guide to the types of consumption expenditure aggregates that conform with international guidelines.
Producer price index
3.74 The PPI measures that part of the variation over time in an enterprise output aggregate that can be attributed to the price movements of the goods and services produced. The greater the extent to which the output aggregate includes all production in the economy valued at the prices received by producers, the more closely the PPI approximates the coverage and valuation of an economy’s output price index.
3.75 A wholesale price index (WPI) is generally understood to be that part of the variation over time in the value of transactions through interenterprise or wholesale markets that can be attributed to the price movements of the goods and services bought or sold. The greater the extent to which transactions through wholesale markets constitute all consumption by producers at purchasers’ prices of noncapital goods and services in an economy, the more closely a WPI can be said to approximate the coverage and valuation of a price index for intermediate consumption.
3.76 Indices for productive activities thus may differ across countries—for example, with respect to the nature and degree of coverage of the economy and the valuation used. Like the CPI, the PPI and WPI are generally compiled using a Laspeyres index formula with a fixed set of items (or “basket”) and a corresponding set of weights from a (generally annual) historical period. The GDDS encourages the compilation and dissemination of the PPI but allows for wholesale and other production price indicators as appropriate to the country’s economic structure and statistical requirements.
3.77 The GDDS does not recommend any component, or subindex, detail. However, countries would obviously find it analytically useful to disseminate subindex detail that parallels the breakdowns in the production indices described above. PPIs are commonly used as deflators for compiling national accounts estimates at constant prices (real terms). Taking into account resource constraints, countries are encouraged to compile PPIs to improve the quality of their national accounts estimates. However, the PPI is commonly restricted to manufacturing industries. For predominantly agricultural economies, this index should not have priority.
Purchasing power parity
3.78 The GDDS encourages the provision of detailed price data for the compilation of purchasing power parities (PPPs) in the context of the International Comparison Program (ICP). PPPs compare price levels between different countries or geographical areas for a given accounting period. PPPs are generally used to eliminate the effect of prices in different currency units when comparing the levels of GDP or income between two countries or areas. The price relatives in bilateral PPPs comprise ratios of the local currency prices of identical goods and services between the two countries or areas.
3.79 Using PPPs, the World Bank calculates poverty indicators included in the MDGs, such as the poverty line, the proportion of the population living on less than US$1 a day, and the poverty gap ratio. These indicators are described in the sections of the Guide devoted to sociodemographic data.
3.80 The GDDS recommends the compilation and dissemination of employment, unemployment, and wages/earnings indicators on an “as relevant” basis. It is recognized that labor market data may be most relevant for industrial countries. Nevertheless, some labor market measures are considered important for almost all countries.
3.81 The data on the labor market should be disseminated on an annual basis and within six to nine months of the end of the year. These goals for periodicity and timeliness are recommended after consultation with the Bureau of Statistics of the International Labor Office (ILO). When the three components are based on different basic data, they may well be compiled and therefore disseminated with differing periodicity and timeliness. The components should be described separately in the metadata (i.e., one Table 3.1, Section B for each).
3.82 The GDDS does not recommend specific definitions of labor market indicators or component breakdowns. However, the recommendations of the ILO provide concepts and definitions of the labor force and, within that framework, provide definitions and classifications of employment and unemployment.11
The recommendations of the ILO also contain proposed concepts, definitions, and classifications for an integrated approach to wage statistics. Use of the ILO’s concepts, definitions, and classifications is encouraged. Reference may also be made to the 1993 SNA, which is consistent with the ILO definitions.
3.83 Some countries compile several measures of employment, which may be based on sample surveys of households or individuals, on establishment surveys, or on social insurance records. For the GDDS, the measure identified should be that most widely used in the country.
3.84 The GDDS recommends employment data on an “as relevant” basis. This approach is likely to be of use where, in view of the nature of the economy, there is less than complete coverage of the population by the underlying surveys or administrative records (for example, the basic data are limited to the nonagricultural population). The MDG indicators include the share of women in wage employment in the non-agriculture sector; therefore, the GDDS recommends the collection of employment data by gender.
3.85 Some countries prepare several measures of unemployment—for example, based on sample surveys of households or individuals, social insurance records, or employment office statistics. For the GDDS, the measure identified should be that most widely used in the country. In addition, either the number of unemployed or the unemployed as a percentage of the labor force may be identified.
3.86 The GDDS recommends the compilation and dissemination of unemployment data on an “as relevant” basis. In many countries, the labor market is characterized less by a dichotomy of employment/unemployment than by a continuum, whereby at one extreme people are fully employed while at the other extreme they are not employed at all. Much of the economically active population falls between these extremes. They may depend, for example, on subsistence agriculture, which may be highly seasonal, or on occasional sales of food or other home-produced goods.
3.87 In such situations, unemployment defined on the basis of practices used in industrial countries may not be entirely meaningful. Therefore, a participating country may substitute a more relevant measure of unemployment, using a concept of underemployment or referring only to the portion of the labor force (for example, in manufacturing) in which unemployment has particular significance.
3.88 In the context of labor statistics, wage data comprise direct wages and salaries for time worked or work done. By contrast, earnings data (in cash and in kind) are broader, covering in addition remuneration for time not worked, bonuses, gratuities, and housing and family allowances paid by the employer to the employee. The series identified for the GDDS may be average earnings or time rates of wages (preferably accompanied by consistent data on hours worked). The scope of the series could differ from country to country, and the one identified should be the series most widely used within the country.
3.89 The GDDS recommends the compilation and dissemination of wages/earnings data on an “as relevant” basis. This will be the case where the survey or the administrative records cover less than the full labor force (for example, covering only the nonagricultural population or wage earners in manufacturing).
3.90 In line with the MDG indicators, the GDDS encourages the disaggregation of all labor market indicators by age, gender, employment status, occupation, and industry as appropriate.
Central government aggregates
3.91 For the fiscal sector, the data category corresponding to the comprehensive framework is central government aggregates. The data should be compiled on a quarterly basis and disseminated within one quarter of the end of the reference quarter. The data should cover at least the budgetary accounts. Ideally, the data should also include as wide a range of central government units as feasible—that is, social security and extrabudgetary units and accounts—to track adequately the movements in the whole of central government.
3.92 The GDDS therefore recommends that participating countries establish full coverage of budgetary accounts as a short-term goal. Extensions of coverage to other important units of the central government should be a medium- to long-term goal, and plans for such extensions should be indicated in the metadata.
3.93 Following GFSM 1986, the GDDS recommends as indicators revenue, expenditure, a balance, and financing with breakdowns. Countries following the recommendations of the GFSM 2001 should use revenue, expense, net acquisition of nonfinancial assets, and financing with breakdowns. The balance (surplus/deficit or net lending/borrowing) that is identified for the purposes of the GDDS should be the concept in use in the country. The determinants of the balance will normally include revenues and expenditures and exclude all transactions that increase or decrease central government liabilities.
3.94 Aggregate financing data should be disseminated, and the classification with respect to domestic and foreign components may parallel that described under the comprehensive framework for central government operations; that is, wherever possible, domestic and external financing transactions, determined by residency, should be presented. Domestic financing should be divided between that provided by the banking system and that provided by other domestic sources. Financing transactions may be presented by type of instrument, currency of issue, or other relevant characteristics.
3.95 Classification and definition of the surplus/deficit-determining items according to either the GFSM 1986 or the GFSM 2001 is encouraged, and those manuals or a regional guideline may be used as a point of reference in preparing metadata.
3.96 Dissemination of interest payments, as part of expenditure, is encouraged, particularly in heavily indebted countries.
Central government debt
3.97 The GDDS recommends the compilation of annual data on central government debt, to be disseminated within one to two quarters of the end of the reference fiscal year. These data should reflect the comprehensive debt liabilities of the central government and should include liabilities in the form of securities, loans, and deposits. Government guaranteed debt is an encouraged extension. A note in the metadata should describe any liabilities of central government units that are excluded and/or special characteristics of debt instruments.
3.98 The GDDS recommends that central government debt data be disaggregated as relevant to a country’s circumstances. Breakdowns by maturity are strongly recommended. Where possible, debt should be classified as short-, medium-, and longterm according to remaining (residual) maturity, but original maturity may be used if data by remaining maturity are not available. Where feasible, a breakdown of debt by foreign and domestic components according to residence should be provided; this will normally be possible for debt in the form of loans. A number of other breakdowns are acceptable, including those by currency of issue or by instrument.
3.99 Classification and definition of debt according to the guidelines of the GFSM, External Debt Statistics: Guide for Compilers and Users (for external debt), or regional guidelines is encouraged. The use of such guidelines, or plans to use them, should be noted in the metadata.
Broad money and credit aggregates
3.100 The data category corresponding to the comprehensive framework for the financial sector—the depository corporations survey—is broad money and credit aggregates. While the coverage of the financial sector for money and credit aggregates will differ in composition across countries, the ideal coverage is that recommended in the MFSM. Narrower institutional coverage, such as only those institutions that have liabilities in the form of narrow money (M1), should be noted in the metadata.
3.101 The GDDS recommends that indicators of broad money and credit aggregates include:
money aggregates (narrow money or M1 if used in the country, or the main broad money aggregates—M2 and M3—that are used in the country);
total domestic credit (ideally broken down by government or nonfinancial public sector and by private sector); and
the foreign position of the depository corporations sector, which may be presented as gross claims on liabilities to nonresidents, or as the net foreign position.
3.102 The GDDS recommends that data on broad money and credit aggregates be compiled on a monthly basis and disseminated within one to three months of the end of the reference month.
Central bank aggregates
3.103 The GDDS recommends that data on central bank aggregates be compiled on a monthly basis and disseminated within one to two months of the end of the reference month. The only indicator specifically recommended is the monetary base. There is a range of definitions of the monetary base; the most relevant country-specific concept should be used. It is recommended that the guidelines contained in the MFSM be used to compile central bank aggregates.
3.104 The GDDS recommends that interest rates on short-term and long-term government securities be disseminated. The most representative rates will vary across countries—a three- or six-month treasury bill rate and a ten-year government bond rate are common examples of representative rates. Under certain circumstances, such as continuous fiscal surpluses, there may be no issuance of government securities. Under other circumstances, such as conditions of high inflation, there may be no issuance of long-term government securities. These facts should be noted on the metadata.
3.105 The GDDS also recommends dissemination of a policy variable rate such as a central bank lending or discount rate. The role of such a rate should be described in the metadata, or the metadata may indicate that interest rates are not used in monetary policy formulation.
3.106 The GDDS encourages the dissemination of money market or interbank market rates and deposit and lending rates. Where there are benchmark deposit and lending rates, these single rates may be disseminated. In other cases, a range of deposit and lending rates should be disseminated.
3.107 The GDDS recommends that interest rates should be available on a monthly basis. Since interest rate data are widely available from private sources, and the dissemination by official producers is not time-sensitive, there is no timeliness recommendation for the dissemination of interest rates.
3.108 The GDDS calls for the dissemination of stock market data, in the form of a share price index, in countries where a stock market exists. These should be disseminated on a monthly basis. The system recognizes that share price indices are widely available from private sources, and so the dissemination by official producers is not time-sensitive—there is, therefore, no recommended timeliness for dissemination of these indices. The GDDS nonetheless recommends that official agencies redisseminate information on share price indices.
Balance of payments aggregates
3.109 The data category for the external sector that corresponds to the comprehensive statistical framework is balance of payments aggregates.
3.110 The GDDS recommends compilation and dissemination of core indicators relating to
imports and exports of goods and services;
the current account balance;
the overall balance.
3.111 Classification of balance of payments components according to the BPM5 is encouraged.
3.112 The GDDS recommends balance of payments indicators be compiled on an annual basis and disseminated within six months of the end of the reference year. The system encourages the production and dissemination of quarterly balance of payments indicators.
External debt and debt-service schedules
3.113 The GDDS recommends the compilation and dissemination of quarterly data on outstanding external debt, covering public and publicly guaranteed external debt, broken down by maturity, with a timeliness of one to two quarters after the end of the reference quarter.
3.114 The GDDS also recommends compilation and dissemination twice a year of the associated debt-service schedule, with a timeliness of within three to six months after the end of the reference period, showing interest and amortization payments for the coming four quarters and two semesters. The twice-yearly periodicity means that countries would have to disseminate the debt-service schedule every second quarter, rather than for two consecutive quarters.
3.115 In addition, the GDDS encourages the compilation and dissemination of data on private external debt not publicly guaranteed and the associated debt service schedule. Periodicity for these data should be annual with a timeliness of six to nine months after the end of the reference year.
3.116 It is recommended that data on external debt and debt service schedules be compiled according to the guidelines of External Debt Statistics: Guide for Compilers and Users, or regional guidelines.
3.117 International reserves serve as a tracking category, providing a more frequent and timely indicator of external sector developments than the balance of payments aggregates. The GDDS recommends that data on international reserves be compiled on a monthly basis and disseminated within one to four weeks after the end of the reference month.
3.118 The GDDS recommends dissemination of data, expressed in U.S. dollars, on gross official reserves, defined according to the guidelines in the BPM5. The dissemination of data on reserve-related liabilities is encouraged. Reserve-related liabilities usually include short-term liabilities of the monetary authorities and use of IMF credit and loans, but country-specific measures may be disseminated.
3.119 Data on merchandise trade serve as a tracking category, providing a more frequent and timely indicator of developments in the current account of the balance of payments. The GDDS recommends that data for merchandise trade be compiled on a monthly basis and disseminated within 8 to 12 weeks after the end of the reference month.
3.120 The GDDS recommends that data on total imports and total exports be disseminated within the indicated timeliness; if dissemination of breakdowns of imports and exports by major commodity with a slightly longer time lag is encouraged. The metadata should indicate whether imports and exports are recorded at c.i.f. or f.o.b. valuations. In countries where disseminated data for total imports and total exports exclude certain items (for example, trade with certain trading partners or certain classes of commodities), the excluded items should be noted in the metadata. Classification according to International Merchandise Trade Statistics: Concepts and Definitions12 is encouraged.
3.121 The GDDS recommends that spot exchange rates be available to the public on a daily basis. If these are readily available in the media or through on-line systems, public redissemination by official agencies may be limited to monthly, or preferably weekly, end-period and period-average rates.
3.122 The GDDS includes four categories of sociodemographic data—population, education, health, and poverty. Section C of Table 3.1 summarizes the basic components that are recommended to be compiled and disseminated for each category. These do not represent the full range of statistics that are relevant for setting or monitoring social policies, nor do they reflect the full range of data-gathering activities in which official agencies may be engaged. The GDDS does not, for example, include categories for housing, criminal justice, or scientific and cultural activities. Nor does it include environmental statistics at the present time.
3.123 While the present recommendations for the sociodemographic data are subject to future elaboration and amendment, the included categories and basic components represent important areas of statistical activity, and the information produced is of great importance to the operation of governments, to the activities of nongovernmental and international organizations, and to civil society in general.
3.124 The sociodemographic component of the GDDS is concerned with the compilation and dissemination of a broad range of information within each data category, rather than a specific set of indicators. Some indicators are derived from various sets of data in order to measure a specific phenomenon. For example, GDP per capita requires an estimate of GDP and total population. GDP is derived from a comprehensive system of national accounts, and current population estimates require a complete set of demographic information, including births, deaths, and net migration. Thus, a single indicator depends on a large infrastructure of statistical information.
3.125 This same infrastructure can be used to produce other important indicators, such as life expectancy at birth, the total fertility rate, and infant and child mortality rates. The GDDS is more concerned with this infrastructure than with the production of individual indicators. While it is likely that most statistical agencies will compile and disseminate many special purpose indicators, such as GDP per capita, their ability to do so depends on having accurate and timely estimates of the underlying data.
3.126 The GDDS makes no specific recommendations concerning which social or demographic indicators should be compiled or disseminated by participating countries. Countries should construct indicators to meet their own national needs based on good statistical practices. Many countries have also committed to monitoring and achieving the MDGs—a set of targets and indicators commonly accepted as a framework for measuring development. The complete list of indicators is presented in Appendix III. MDG indicators related to each of the four sociodemographic sectors are identified in Section C of Table 3.1.
3.127 Some countries have also identified indicators to monitor national poverty reduction efforts (in the course of preparing Poverty Reduction Strategy Papers—PRSPs—with the IMF and World Bank) and development policies. Availability of MDG and PRSP indicators, including their periodicity and timeliness, should be noted in the metadata. In addition, some international organizations request regular reporting of specialized data within their area of responsibility. Plans to upgrade statistical systems and outputs to include MDG and PRSP indicators or to meet international reporting requirements may be noted in the metadata under plans for improvement.
3.128 The basic components for the sociodemo-graphic data that are listed in Section C of Table 3.1 include information on inputs of resources (financial as well as human) into the social area, insofar as it may provide a useful link to public expenditure policies. This feature is particularly appropriate for the GDDS in that it is meant to provide a comprehensive framework spanning a broad range of interrelated policy areas, both in the social and economic spheres.
3.129 The organization of the GDDS for macroeconomic and financial statistics is based on the concept of comprehensive frameworks, such as the national accounts, that provide well-recognized economic and financial measures. No equivalent comprehensive frameworks exist for sociodemographic data, but guidelines do exist for compilation, standard classification systems, and examples of “best practice” that are frequently cited and widely used by statisticians to organize the collection and presentation of social and demographic statistics. Appendix II provides a selected list of references.
3.130 The four data categories for the sociodemo-graphic data and their basic components are (i) population, (ii) education, (iii) health, and (iv) poverty.
3.131 Demographic statistics are concerned with describing the size, composition, and location of the population. Data are usually obtained from a complete enumeration of the population (which may include nonresidents living abroad, resident aliens, or refugees, as defined by law or statistical practice), supplemented in years between censuses by information on births, deaths, and migration collected through registries of vital statistics or through the use of surveys and from information derived from other administrative records.
3.132 Most countries will find it useful to compile data on population size disaggregated by age, gender, and location; the number of births (by gender), deaths (by age and gender), and inward and outward migration (by age and gender). Analysts often need population estimates at subnational levels. Derived indicators that are of great importance in analyzing demographic trends and anticipating future population change include age-specific mortality rates and the total fertility rate.
3.133 Censuses often produce additional information. For example, data on housing stocks, self-reported education and employment status, and access to water and sanitation may also be recorded. Such data can be used to supplement and cross-check information from other sources.
3.134 Education statistics are generally produced and disseminated by the agencies that oversee public education services, including those that accredit education facilities, and train or license teachers, as well as schools, colleges and universities that provide education. Responsibility for data collection and reporting may be centralized in one agency or scattered among many.
3.135 In practice, data are obtained primarily from administrative reporting systems of the education ministry and national surveys, but statistical reports also may be compiled and disseminated by another agency, such as the national statistical office. The metadata should record the work of the agencies responsible for the primary production of education statistics and their first public release. The internal process of collecting administrative data is not of concern to the GDDS, except insofar as it affects the coverage, quality, or integrity of the resulting statistics.
3.136 In Section C of Table 3.1, data are organized in three subcategories: inputs, which measure the physical and financial resources available to the education system; process, which records the flow of students through the system; and outcomes, which measure attainment and learning achievement.
3.137 The education sector comprises all levels of the formal education system: primary, secondary, and tertiary, along with preprimary and vocational training if possible. The use of the International Standard Classification of Education is recommended for international reporting of data. Disaggregation by age, gender, grade or level of education, and type of educational program, along with the disaggregation of data by subnational or regional units, as appropriate, is encouraged. If adult literacy and nonformal training programs are significant, data on these activities should also be reported, along with measures of literacy and numeracy in the population.
3.138 In nearly all countries the state is the predominant supplier of education, and published statistics may record only the activities of public education programs. But in many countries private, religious, or military schools also play an important role. As far as possible, data should be collected from such entities and included in statistical reports.
3.139Health statistics are generally produced by the agencies that oversee public health services, including those that accredit health facilities and train and license health workers, as well as those that provide primary health care. Responsibility for data collection and reporting may be centralized in one agency or scattered among many. The primary sources of health statistics are administrative systems, including hospital records, and national health surveys. Like education statistics, health statistics are organized in three subcategories in Section C of Table 3.1: inputs, process, and outcomes. Process measures are further divided into preventative and curative. These are analytically useful, but they may not reflect the actual structure of reporting used in a country.
3.140 Inputs include both the financial resources used by the public health system and the number and capacity of facilities and the personnel employed in providing services. A complete system of health accounts would also record private expenditures, but in practice these data are difficult to obtain. However, the number and capacity of private facilities and the number of trained personnel in private practice should be available. Information on personal behaviors that affect health status, such as the use of contraceptives or the consumption of cigarettes and alcohol, may be obtained from surveys or estimated indirectly from aggregate consumption data.
3.141 Process or service delivery measures should record the type of service provided and characteristics of the population served. Among the most important categories of preventative services provided are immunizations, mother and child health outpatient care, and reproductive health services. Other services, such as sewerage, solid waste management, water supply, and sanitary protection of the food supply, may not be considered part of the formal health care system, but all have important effects on public health. Information on the scale and scope of these services, including measures of accessibility, are important for diagnosing sources of illness and identifying unmet needs. Curative services are those whose purpose is to treat specific diseases.
3.142 Reporting systems should track inpatient admission rates and outpatient visits to health providers. It is also useful to monitor the nonhealth aspects of service delivery—for example, waiting time for service, quality of facilities, and client perceptions of the quality of service.
3.143 Outcomes record the incidence of disease and causes of death. Disaggregation by age, gender, and location, and the calculation of age-specific mortality rates by cause is encouraged. Data on contagious diseases, which can be used to identify and isolate outbreaks and direct treatment programs, should be compiled with higher frequency and greater timeliness than routine administrative statistics. Measures to assess the burden of disease, which capture the lifetime consequences of disease on productivity and quality of life, are currently under development by the World Health Organization (WHO).
3.144 Poverty is a complex phenomenon and cannot be measured along a single dimension. “Money-metric” methods, based on the income or consumption of households or individuals, are commonly used to establish a “poverty line” and measure the number of poor and the depth of poverty. However, many statistics discussed in other categories, such as educational attainment, health status, and employment status, are useful for diagnosing poverty, especially when they are collected at a sufficiently disaggregated level.
3.145 Because poverty is fundamentally an affliction of individuals and a reflection of the unequal distribution of wealth, income, education, health, and access to public and private resources, the measurement of poverty requires micro-level data. These data are commonly produced through household surveys.
3.146 Consumption is a better measure of welfare than income. Consumption by individuals, accompanied by information on the characteristics of the household and of individual household members, is the preferred basis for assessing money-metric poverty. Because real income and therefore consumption may vary according to the time of the year and price levels, it is important to adjust for seasonal effects and regional price differences when compiling income and expenditure data.
3.147 A poverty line provides a standard way of classifying the population as poor and nonpoor. It is usually based on an estimate of the minimum level of income or consumption needed to sustain a person at a minimally acceptable level. While a poverty line may be based in part on objective standards, such as the cost of a minimum calorie budget, poverty lines are generally not comparable across countries because of differences in the cost of living and cultural perceptions of poverty.
3.148 The MDG-recommended poverty line—the proportion of the population living on less than US$1 a day—and the poverty gap ratio are calculated by the World Bank, using PPPs. They therefore do not need to be covered in this section. However, if countries use US$1 a day as their national poverty line, the methodology and related metadata should be noted. Information on availability of income or consumption distribution should be noted, along with periodicity and timeliness.
3.149 Household income and expenditure surveys should collect information on the size and structure of the household, household assets, sources of income, and types of consumption expenditures. They may also record characteristics of individual household members. The more detailed the data collection and the closer it comes to measuring the resources and expenditures of individuals, the more useful it will be for identifying the poor, diagnosing the causes of poverty, and designing programs to eradicate it. However, household surveys are costly to administer, and large, complex data sets are likewise costly to analyze.
3.150 Access measures, which record the availability of important public services and use of these services, may be derived from household surveys or from the administrative records of service providers. They are most useful when recorded on a sufficiently small scale to provide some indication of the distribution and use of services. Mapping and geographic information systems, which record the proximity of people to service facilities, have proven to be a useful means of recording access information.
Quality Dimension of the Disseminated Data
3.151 The quality dimension, including development of plans for improving data quality, is closely associated with the data dimension in the GDDS. The quality dimension of the GDDS is especially important, given that the primary goal of the system is improvements in data quality over time.
3.152 The assessment of quality is addressed in the GDDS in two ways. First, the metadata for comprehensive frameworks are intended to provide detailed information on definitions, classifications, sources of data, compilation methods, and use of international guidelines. While the GDDS does not require a specific evaluation of data quality in the metadata, the information called for will allow users to draw conclusions about aspects of data quality. Second, for data categories and indicators, the quality dimension contains the following two proxies for a specific evaluation.
(i) Dissemination of documentation on methodology and data sources used in preparing statistics
3.153 The availability of documentation on methodology and data sources underlying statistics is key to users’ awareness of the strengths and weaknesses of the data. The documentation may take several forms, including summary notes accompanying release of the data, separate publications, or papers available on request from the producers. Countries participating in the GDDS are encouraged to include and highlight statements and declarations about important features of quality. These may indicate the kinds of errors to which the data are subject, sources of noncomparability over time, measures of coverage for census, and data or sampling error for survey data.13
3.154 Participating countries should provide metadata that identify the documentation and the means of accessing it, as well as plans for compiling and disseminating statements on methodology and sources where these do not yet exist.
(ii) Dissemination of component detail, reconciliations with related data, and statistical frameworks that support statistical cross-checks and provide assurance of reasonableness
3.155 To support and encourage the assessment of data quality by users, the GDDS recommends the dissemination of data on the components of underlying aggregate series, dissemination within a statistical framework, and/or dissemination of comparisons and reconciliations with related data.
3.156 Dissemination of component detail should, of course, be at a level that does not conflict with other desirable characteristics such as statistical reliability and the confidentiality of individually identifiable information. (Component detail, although used in this context to facilitate cross-checks, is also useful in its own right for in-depth analysis and research.)
3.157 With a sufficient amount of published detail below the aggregate (or indicator) level, especially in combination with documentation, a user is in a better position to judge the reasonableness of the data. For example, published detail allows a user to compare rates of change for components of time series, calculate deviations from past trend for components of time series, and calculate percentage composition.
3.158 Comparisons and reconciliations may be of several different kinds, and some may cut across statistical frameworks, such as exports and imports as part of the national accounts and as part of the balance of payments. For instance, it may happen that imports and exports of goods declared in foreign currency are valued in the national accounts using exchange rates that are not calculated in the same way as the exchange rates used in compiling balance of payments statistics. A comparison of the two sets of data would make this difference explicit.
3.159 Another example of comparisons and reconciliations across comprehensive frameworks relates to the residency criterion in the balance of payments and in the depository corporations survey. In the case of countries with significant numbers of nationals working abroad, it is sometimes difficult to draw the line, on the basis of the BPM5’s one-year rule, between temporary and long-term/permanent emigrants, the latter being nonresident while the former continue to be resident.
3.160 The issue may be even more difficult when the host government authorizes permanent or long-term emigrants to have accounts in domestic banks with the same status as resident accounts. In such a situation, the depository corporations survey and the balance of payments can be inconsistent with each other, in particular with respect to the international position of the banking sector, if the two data sets do not adopt the same treatment. However, the balance of payments statistics for this component are typically drawn from the depository corporations survey.
3.161 Comparisons and reconciliations may also cover data from different sources within the same country. For instance, unemployment data come from population censuses and household surveys, while employment data are often available from economic censuses and enterprise surveys. In general, both sets of statistics do not provide the same estimates for employment for various reasons. In such cases, reconciliations can help users to assess the accuracy of the two sets of data.
3.162 For the GDDS metadata, participating members should describe the component detail, the framework, and the comparisons and reconciliations that are disseminated. Wherever such comparisons and reconciliations do take place, the system recommends that they be made available to the public. The metadata should therefore mention whether comparisons and reconciliations are made available to the public.
Dimensions of Integrity of the Disseminated Data and Access by the Public
3.163 A flow of comprehensive, reliable, accessible, and timely statistics is indispensable to informed policymaking, and the wide availability of statistics to the general public, including investors, helps provide discipline by supporting informed public debate and market assessment. These roles for statistics lead to certain basic principles for the official statistics function within a country, which are described in the Fundamental Principles of Official Statistics (see Appendix I). These basic principles are reflected directly in the integrity and access dimensions of the GDDS.
3.164 These dimensions have been embodied in a manner that gives due consideration to the state of data compilation and data dissemination and to the uses of data across the broad range of countries to which the system applies. It is recognized that many countries are at an early stage of applying these principles. The focus of the integrity and access dimensions in the GDDS, therefore, is on developing policies and practices that are in line with the objectives of readily accessible data. The metadata for integrity and access will therefore include, for each official agency that compiles and disseminates data covered by the System, statements that cover existing policies and practices as well as plans to incorporate the specific elements in the future.
3.165 To fulfill the purpose of providing the public with information, official statistics must have the confidence of their users. In turn, confidence in the statistics ultimately becomes a matter of confidence in the objectivity and professionalism of the agency producing the statistics. Transparency of its practices and procedures is a key factor in creating this confidence.
3.166 Accordingly, the integrity dimension of the GDDS includes four practices that indicate the transparency of the producing agency’s practices and procedures. Three of them refer to administrative practices and procedures; the fourth deals with revisions and changes in methodology.
(i) Dissemination of the terms and conditions under which official statistics are produced, including those relating to the confidentiality of individually identifiable information
3.167 While the recommended practice, embodied in the Fundamental Principles of Official Statistics (UNSD, 1994), is indirect, it is nevertheless instrumental in fostering confidence in the objectivity and professionalism with which official statistics are prepared.
3.168 The terms and conditions under which statistical agencies operate may take various forms, including statistical laws, charters, and codes of conduct. In many countries, these may not be in place or may be out of date. Accordingly, a first step toward the objective would be to put such laws, charters, and codes in place. The terms and conditions incorporated in them may refer to matters such as the relationship of the statistical unit to a larger department or ministry of which it is part. For example, they may distinguish between matters on which the statistical unit is subordinate to an outside authority (for example, on budgetary and personnel matters) and those on which it is independent (for example, in statistical methodology and decisions on the publication of statistics).
3.169 The terms and conditions may also refer to a requirement that the statistical unit publish data it has collected as a safeguard against possible pressure to keep some findings from the public; the qualifications and terms of reference of the chief statistician/director; and the review of statistical programs by an independent group of experts. All of these may be viewed as safeguarding the professionalism and objectivity of the producing agency, but, of course, they do not guarantee this.
3.170 Another key aspect of the terms and conditions under which statistical agencies operate refers to procedures and processes relating to the confidentiality of individual responses from persons, businesses, and organizations to official inquiries. These procedures and processes may cover topics ranging from computer security to restrictions on interagency disclosure.
3.171 Dissemination of information about terms and conditions may take a variety of forms. For example, the annual report of a statistical agency may include a section on steps taken to ensure the confidentiality of individual responses, and this section may refer to the legal obligations of the agency’s employees with respect to confidentiality, security of computer systems and agency buildings, and statistical procedures for disclosure avoidance. Other publications and Internet sites may describe the basis for an agency’s data collection and dissemination activities and its practices, as in the examples below:
[Name of agency] collects, processes, and disseminates information on [country’s] demographic, social, and economic development. In addition it produces data on the country’s physical environment and natural resources. [Name of agency] was created by presidential decree, and its sphere of duties is set forth in the Law on Statistical Information, which also provides it with the technical and administrative autonomy needed to carry out its functions.
[Name of agency] is administratively subordinate to [name of ministry] but operates according to strictly professional considerations with respect to the data it produces and disseminates.
[Name of agency] has at its disposal many data about individuals and companies. These data are never released to other parties or government agencies. [Name of agency] never publishes or discloses data combinations from which individual data can be derived.
3.172 Survey forms and transmittal letters may carry statements or quote relevant legislation or codes referring to confidentiality of responses and assurances that the responses will be used only for statistical purposes and will not be divulged to others, including regulatory and tax agencies.
3.173 Compilers may find it convenient to use logos and other insignia to remind users of the terms under which their statistics are produced. For example, the official series in one country might consist of those prepared according to the prescriptions of the statistical law and might be designated. Releases prepared by another country’s government statistical service, according to its code of practice, might be marked with its banner and logo.
3.174 For the GDDS metadata, participating countries should summarize the terms and conditions applicable for each agency that produces data covered by the system and note where the laws, charters, codes, etc., are available to the public. The terms and conditions may vary across data categories, especially in countries with decentralized statistical systems.
3.175 The dissemination of information on terms and conditions, as just described, is applicable to official statistics. If a private organization that is producing any data covered by the GDDS does not disseminate such information, participating countries are encouraged to foster openness about the conditions under which key data are prepared and released. Countries could, if they wished, provide relevant information about the terms and conditions under which the private data described in the metadata are produced, with an appropriate note about the source of the information. For example, the entry could note, as appropriate, the organization’s nonprofit status, its nonpartisan approach, or its university affiliation, etc.
(ii) Identification of internal government access to data before release
3.176 In the interest of protection from undue influence on the data before their release, the GDDS recommends compiling a listing of persons or officials holding designated positions within the government, but outside the agency producing the data, who have prerelease access to the data and the dissemination of the schedule according to which they receive access. The recommended practice is intended to provide for any necessary prerelease access within government that the government deems appropriate, while giving full transparency to those practices.
3.177 Country practices differ in this regard. Some countries maintain strict embargoes on the availability of data even within the government prior to their release, while in others such procedures would be viewed as unduly restrictive and detrimental to fast and effective government reaction. Thus, rather than recommending a specific and uniform set of practices, emphasis in the GDDS is placed on the means by which the necessary degree of transparency can be assured.
3.178 Such identification of prerelease access—that is, statements of “who knows what and when”—may take a variety of forms. The identification may, for example, be included in brief notices to the public and/or in annual reports of the producer of statistics. A brief notice could be along the following lines:
Data from [name of agency] are available to all users simultaneously with limited exceptions. The exceptions are on a need-to-know basis; those with prerelease access receive advance copies of news releases (or related materials) no earlier than 48 hours before the release date and time. The exceptions are:
Head of Government
Governor of the Central Bank
Minister of Finance
In addition, other ministers, policy advisers, and, on a need-to-know basis, a limited number of ministerial and central bank staff may be briefed on a confidential basis [specify] hour(s) in advance on the day of release.
3.179 This example highlights the fact that the focus is on undue—potentially political—influence or policy advocacy. The approach is not meant to include in the listing statisticians and other employees of the producing agency, who of necessity often see data at early stages or in fragments.
(iii) Identification of ministerial commentary on the occasion of statistical releases
3.180 Ministerial commentary is not necessarily expected to maintain the same degree of objectivity or freedom from political judgment as would be expected of good practice for a producer of official statistics. Therefore, the practice that is proposed in the GDDS is to identify such commentary so that its source will be transparent to the public.
3.181 The identification of ministerial commentary on the occasion of statistical release may take several forms. One common form is for the minister (or other policy or political official) to issue a statement that is clearly separate from the statistical release—the ministerial statement under the minister’s letterhead and the statistical agency material under its letterhead or logo.
3.182 Alternatively, the statistical agency’s material may be included, but presented separately, in a release that contains both ministerial commentary and data. The agency’s material may include data, explanatory text (for example, of an unusual event affecting the data), and objective analysis; the identification as agency material may be made in various ways, including the use of source lines in tables and of the producer’s logos or other insignia.
3.183 Participating countries should describe in the metadata the procedures applicable for each agency and note any varying practices that may apply to individual data sets.
(iv) Provision of information about revisions and advance notice of major changes in methodology
3.184 Although users are interested in revisions of official statistics from several points of view, the emphasis in the GDDS is on practices related to revisions that enhance the transparency of compilers’ practices. The practices called for in the system are meant to give compilers of official statistics several ways of providing information about revisions and thus to strike a balance between users’ concerns about revisions and the resource cost to compilers of providing the information.
3.185 Relevant information about revisions stemming from changes in the underlying data may include statements about the policy followed and data about the size of past revisions. Revisions are also made to incorporate improvements or other changes dealing with any of the several aspects of statistical methodology—statistical frameworks; concepts, definitions, and classifications; accounting conventions, source data; and compilation procedures. A revision may make reference to any or all of these, as appropriate.
3.186 For example, there may be a policy about how often new concepts, definitions, and classifications are incorporated for each data set—for example, only at five-year intervals. With respect to the incorporation of additional source data, the policy may be with respect to the period over which new sources are introduced—for example, a continuing basis (revising monthly data until an annual, more comprehensive survey becomes available) or a policy of making no routine revisions. Data on the size of past revisions may refer to estimates such as dispersion and bias.
3.187 Such information may be summarized in news releases, for example, and may be presented more fully in comprehensive reports or separate periodic evaluations. Revision practices about specific comprehensive frameworks and data categories should be described in Tables A (Section V “Compilation Practices”) and B (Section I, “Data Characteristics,” under “coverage”) respectively, in Chapter 4. This element under Data Integrity (Table C in Chapter 4) should describe the agency’s general policy with regard to revisions of the data that it compiles and disseminates.
3.188 The advance notice of changes in methodology should deal with major changes as defined above—for example, change in base year, major expansions of sample size, introduction of alternative data sources, introduction of new weighting schemes for indices, introduction or change in the methods of seasonal adjustment, and reclassification of transactions or industries.
3.189 The notices may take a variety of forms, ranging from short statements in news releases to presentations and papers in public forums. Participating countries are encouraged, as well, to provide easy access to information about methodological changes when they occur (for example, by telephone, fax, or Internet contact able to answer questions about revisions).
3.190 For the metadata, each data-disseminating agency should describe its revision policy for each data category and, where relevant, indicate for each data category the size of past revisions (at least for major aggregates), or other relevant information about revisions (such as the sources of revision and/or components that show the largest revisions). The procedures, if any, for providing advance notice of major changes in methodology should be described. The metadata should also describe, where appropriate, plans for implementing the recommendations of the GDDS with regard to data revisions and advance notice of changes in methodology.
Dimension of Access by the Public
3.191 Dissemination of official statistics is an essential feature of statistics as a public good. Ready accessibility and equal access are principal requirements for the public, including market participants.
3.192 The access dimension of the GDDS is based on two practices—dissemination of advance release calendars and simultaneous release to all parties—that facilitate ready and equal access. In characterizing access, “timely” is often added to “ready and equal” to form a triad of desired attributes.
3.193 The lapse of time between a reference date (or close of a reference period) and the dissemination of the data is captured in the “timeliness” element of the data dimension, which encompasses both the time needed for statistical processing (or compilation) and the time needed to prepare for dissemination (for example, printing). Accordingly, the following two practices recommended for access assume that the timeliness objectives are in place.
(i) Dissemination of advance-release calendars
3.194 Advance-release calendars highlight sound management and transparency of statistical compilation and provide data users with information needed to take a more active, organized approach to their work. Data users are strongly in favor of advance-release calendars—financial market analysts and commercial resellers who work with data from a broad range of countries find them a particularly important tool in planning their activities. In recent years, a growing number of national statistical offices have issued advance-release calendars—in several countries, these calendars are mandated by codes of conduct, policy directives, or agency administrative manuals.
3.195 The GDDS recognizes that many participating countries will not have had experience with advance-release calendars. Scarcity of resources available to statistical agencies and shortcomings in survey and other data collection mechanisms may make it difficult to identify precise release dates in advance. In addition, for countries that are not participating in international financial markets or that do not have well-developed local financial markets, the needs of data users may not be highly time-sensitive. The GDDS therefore recommends the following with regard to advance-release calendars:
For comprehensive frameworks and indicators for which annual periodicity is recommended, a specification of dates no later than which the data will be released. The specification of no-later-than dates could be based on the prior years’ results.
For other data sets, ranges of dates, such as three to five days, during which data are expected to be released.
3.196 The calendars should cover the year ahead and be disseminated in the disseminating agency’s highest-frequency publication, in press releases, and via electronic media, if available. Participants are encouraged to identify an office or person who could provide the latest information about the likely date of release, including release of data for which periodicity and timeliness are irregular and newly disseminated data.
3.197 Release calendars can carry a cautionary note that the dates indicated are “expected,” “targets,” etc. Participants are encouraged to post their advance-release calendars on the IMF’s DSBB together with notes describing events, such as a computer failure, that make it impossible to meet the release dates previously indicated. This serves to reduce the likelihood that pressure to meet release dates will have negative effects on other aspects of data quality.
(ii) Simultaneous release to all interested parties
3.198 The GDDS recommends that data be released to all interested parties at the same time. This is not intended to refer to access by government ministries and agencies; such prerelease access is governed by conditions set out in the description of integrity above.
3.199 For media and commercial data vendors, simultaneous release may be interpreted as including access, under embargo conditions, to all on equal terms. The embargo conditions that are actually imposed may vary in strictness with the value of the data to users, including the financial markets, and the competitiveness of the news media and other data distributors, as well as other factors. In more relaxed settings, data may be provided to all accredited media representatives who request early access on condition that they neither use the data for gain nor transmit their stories/analyses until after general release to the public.
3.200 The act of release is meant to refer to the first availability of the data to the public. It may center on one dissemination format—for example, a press release or Internet posting providing summary data. As soon as practicable, other formats (for example, diskettes and access to electronic databases) should be made available, as should accompanying detail and/or explanations in other formats.
3.201 The definition of “simultaneous” must be determined by the situation; in some countries, simultaneous release is being defined with increasing strictness for high-profile data. For the GDDS, the objective with regard to simultaneous release may be met by providing at least one publicly identified and accessible location where data are available to all on an equal basis at the time of first release.
3.202 Countries are encouraged to make releases in as many formats as possible, especially electronic formats if available, consistent with the extent of public interest in the data. Electronic formats greatly increase the number of parties and geographic locations to which simultaneous access can be provided.
3.203 Participating countries should describe in the GDDS metadata the procedures of each data-disseminating agency that ensure simultaneous release; variations among data categories should be noted. These procedures may vary according to the sensitivity and other characteristics of the data. For example, the procedures for high-profile data may be stricter than for others, and data that require extensive technical explanation may involve the availability at the time of release of technical staff of the producing agency to answer questions.
There is no comprehensive framework for sociodemographic data.
Commission of the European Communities–Eurostat, International Monetary Fund, Organization for Economic Cooperation and Development, United Nations, and World Bank, System of National Accounts 1993 (Brussels/Luxembourg, New York, Paris, and Washington, 1993); and Eurostat (Statistical Office of the European Communities), European System of Accounts: ESA 1995 (Brussels/Luxembourg: Office for Official Publications of the European Communities, 1996).
Adrian Bloem, Robert J. Dippelsman, and Nils O. Maehle, Quarterly National Account Manual: Concepts, Data Sources, and Compilation (Washington : International Monetary Fund, 2001).
International Monetary Fund, Government Finance Statistics Manual 2001 (Washington, 2001).
International Monetary Fund and others, External Debt Statistics: Guide for Compilers and Users (Washington, 2003).
International Monetary Fund, Monetary and Financial Statistics Manual (Washington, 2000).
International Monetary Fund, International Investment Position: A Guide to Data Sources (Washington, 2002); available via the Internet at http://www.imf.org/external/np/sta/iip/guide/iipguide.pdf.
The inclusion of financial derivatives in the list of breakdowns stems from the publication in 2000 of the financial derivatives supplement to the BPM5: International Monetary Fund, Financial Derivatives, A Supplement to the Fifth Edition (1993) of the Balance of Payments Manual (Washington, 2000).
International Labor Organization, International Monetary Fund, and others, Consumer Price Index Manual: Theory and Practice (Geneva: ILO, 2004)—available via the Internet at http://www.ilo.org/public/english/bureau/stat/guides/cpi/index.htm; and International Labor Organization, International Monetary Fund, and others, Producer Price Index Manual: Theory and Practice (Washington: IMF, 2004)—available via the Internet at http://www.imf.org/external/np/sta/tegppi/index.htm.
Classifications of Expenditure According to Purpose (New York, United Nations, 1999). United Nations classification can be accessed through the UN website: http://unstats.un.org/unsd/pubs/gesgrid.asp.
International Labor Organization, Current International Recommendations on Labour Statistics (Geneva, 1985). See also ILO, Surveys of Economically Active Population, Employment, Unemployment and Underemployment (Geneva, 1990).
United Nations, International Merchandise Trade Statistics: Concepts and Definitions (New York, 1998).
The size of past revisions, which is often in the list of aspects of quality, is included as a component element of the integrity dimension, drawing on its role as an inidicator of the transparency of conditions under which data are produced.