1. Introduction

2. The SDDS and the Fiscal Sector

3. Best Practices for Fiscal Statistics

4. Experiences in Countries Implementing the GFSM 2001

5. Moving to Quarterly General Government Data

6. Conclusions

Box

1. Countries' Experiences with the Accrual Recording of Government Finance Statistics

I. Introduction

1. During the Third Review of the Fund's Data Standards Initiatives in March 2000, Executive Directors stressed that the Special Data Dissemination Standard (SDDS) should continue to encourage the adoption of data quality improvements and agreed with the proposal that, for SDDS subscribers implementing accrual-based reporting systems, the periodicity and timeliness of fiscal data be on a "best efforts" basis (see BUFF/00/43). These special arrangements on fiscal data will expire at the conclusion of this Fifth Review. Since the Third Review, the IMF has disseminated the Government Finance Statistics Manual 2001 (GFSM 2001), and the staff has reviewed country experiences with developing fiscal statistics using accrual-based information.

2. In this context, this paper provides the rationale for the proposed addition of a targeted timeliness flexibility option to the SDDS requirements for the central government operations (CGO) data. The proposal aims to address the problem that the current prescription, monthly CGO data with timeliness of one month, serves as an impediment to achieving the data quality improvements that occur as subscribers move towards the compilation of fiscal statistics according to international best practice as set forth in the GFSM 2001 or an equivalent standard.

3. The proposal seeks to facilitate the trend to produce general government operations (GGO) statistics with a higher periodicity than currently prescribed in the SDDS as a best practice. If a subscriber disseminates, with a one-quarter lag, quarterly GGO data in line with the GFSM 2001 or an equivalent standard, a targeted timeliness flexibility option would be allowed for monthly CGO data (on a cash, accrual, or modified cash/accrual basis, consistent with current SDDS requirements). The new targeted timeliness flexibility option would replace the special arrangements, approved by the Executive Board under the Third Review, for subscribers implementing accrual-based reporting systems for fiscal data. There are no proposed changes to the SDDS requirements for subscribers that have not taken steps to strengthen their fiscal data through the implementation of the GFSM 2001. That is, subscribers that disseminate GGO data only on a cash basis would not be allowed the targeted timeliness flexibility option.

4. This proposal does not have consequences for the General Data Dissemination System (GDDS). The GDDS recommends that data for CGO be disseminated on a quarterly basis with a one-quarter lag and that data on GGO, which is an encouraged GDDS data category, be disseminated on an annual basis with nine-months lag. The GDDS encourages participants to migrate to the standards set out in the GFSM 2001 at a pace that is compatible with the countries' state of statistical development and resource availability.

5. Section II describes the current SDDS requirements in the fiscal sector. Section III reviews best practices for fiscal data, based on the GFSM 2001, which recommends the adoption of accrual-based information. Section IV describes the experiences of countries thus far in implementing the GFSM 2001, including accrual reporting. Section V discusses the advantages and challenges of moving to quarterly general government data. Section VI concludes with arguments that support the proposal in the main text to modify the SDDS prescription for the CGO data category to add a targeted timeliness flexibility option.

II. The SDDS and the Fiscal Sector

6. The SDDS was designed to encourage the provision to the public (especially financial markets) of comprehensive, accessible, and reliable economic and financial statistics, reflecting best practices in data dissemination. Specifically, the SDDS prescribes the following for the fiscal sector:

• dissemination of annual data on GGO with a timeliness of six months;

• dissemination of monthly data on CGO with a timeliness of one month as a tracking indicator, providing a more frequent and timely indicator of fiscal operations than is usual with GGO data;

• the CGO data should cover, in addition to budgetary accounts, as wide a range of central government units (social security and extra budgetary units and accounts) as is feasible;

• dissemination of the deficit/surplus concept in use in the country; and the principal determinants of that balance;

• dissemination of aggregate financing data and, where relevant, a breakdown between domestic and external financing:
  
  a) domestic financing should be divided between the banking system and other domestic sectors, and
  
  b) in other cases, financing transactions may be presented by type of instrument, currency of issue, or other relevant characteristics; and

• dissemination of quarterly data on central government debt with a timeliness of one quarter.

7. Currently, most subscribers use the classifications and definitions of A Manual in Government Finance Statistics, 1986 (GFSM 1986) for disseminating data for the prescribed fiscal categories.

8. As noted in the introduction, special arrangements on fiscal data for subscribers implementing accrual accounting were approved at the time of the Third Review of the Fund's Data Standards Initiatives and will expire at the conclusion of this Fifth Review. For those subscribers, the periodicity and timeliness of fiscal data dissemination has been on a best effort basis. Except for these special arrangements, the fiscal requirements are those described in the Guide to Data Dissemination Standards (1996).

III. Best Practices for Fiscal Statistics

9. Following publication of the System of National Accounts 1993, the Fund began to revise the fiscal statistics standard (GFSM 1986) and to harmonize it with the national accounts and other macroeconomic statistical standards. The GFSM 2001, which embodies international best practices, is a significant step forward in the international effort to improve the comprehensiveness, transparency, and analytical usefulness of fiscal data reporting.

10. The GFSM 2001 calls for fiscal statistics on an accrual basis, while maintaining information on a cash basis. Accrual accounting, widely used in business, records economic activity when economic value is exchanged or transferred (transactions) or is created, transformed, or extinguished (other economic flows). The GFSM 2001 also introduces a comprehensive analytical framework that integrates balance sheet information with transactions and other economic flows. This framework can be used to present key fiscal variables:

• An integrated set of statements of stocks and flows:
  
  a) the government's net worth at the beginning and end of the accounting period; and
  
  b) the contributions to the changes in the balance sheet through transactions and other economic flows (revaluations and other volume changes).

• Several balances, instead of the single balance (the deficit/surplus) emphasized in the GFSM 1986:
  
  a) the net operating balance (revenue minus expense), which measures the impact of government operations on its net worth;
  
  b) net lending/borrowing, which measures the extent to which the government provides (draws) financial resources to (from) the other sectors of the economy and the rest of the world;
  
  c) net worth (total assets minus total liabilities); and
  
  d) the net change in the stock of cash (net sum of cash received).

11. The GFSM 2001 recommends the compilation and dissemination of these statements and balances for the general government sector and the broader public sector. The implementation of the fully integrated system presented in the GFSM 2001 will take time, and progress will be determined by the differing needs and circumstances of each country.

IV. Experiences in Countries Implementing the GFSM 2001

12. Since the Third Review, staff has examined countries' experiences with the development of accrual accounting and accrual-based statistics, in line with the GFSM 2001 or an equivalent standard. Complete accrual recording, as recommended in the GFSM 2001, is in place in two countries (Australia and Iceland) thus far, and several other countries in Europe and Latin America have made progress to adopt accrual-based data (Box 1).

13. Based on these experiences, staff has observed that compiling accrual-based statistics is significantly more resource intensive than preparing information solely on a cash basis. Under an accrual-based system, more time and resources are needed to prepare balance sheets and an operating statement, as well as a cash statement. Investigations need to be carried out to retime the recording of activities to ensure the precision and accuracy of the data. A thorough set of notes must accompany the accounts. Audit statements, which enhance accuracy, take time to prepare. Because accrual-based statistics must be reconciled with the cash data, the cash-based data also would be delayed.

14. The SDDS requirement that monthly CGO data be reported with a one-month lag creates strains at the turn of the fiscal year, when resource pressures arising from the compilation of accrual-based data are most intense. Typically, a larger volume of transactions connected with the full spending of budgetary allocations and the existence of complementary periods for spending contribute to complications and delays. At the turn of the fiscal year, accountants and statisticians require a clear cut-off date to close the books, and therefore the compilation of monthly statistics must wait for these source data, and is less timely than other months of the year.

V. Moving to Quarterly General Government Data

15. At present, general government data usually are presented on an annual basis. However, for fiscal policy formulation, execution, and monitoring, quarterly general government data are preferable. In order to meet the needs of fiscal policy (and also as an input to quarterly national accounts), the compilation and dissemination of quarterly general government data, following the GFSM 2001 framework, is desirable.

16. European Union (EU) member states already provide the European Commission with quarterly data on general government tax revenue and on social contributions and benefits on an accrual basis (European System of Accounts 1995).¹ The prescribed timeliness for these data is one quarter. In addition, at the end of September 2002, many EU members started providing the European Commission with more comprehensive data covering quarterly general government revenue, expenditure, and net lending/borrowing.² While a trial period is foreseen (ending in 2005), most of the countries are already submitting the entire data set within a 90-day deadline.

17. The Australian Bureau of Statistics plans to release quarterly fiscal statistics for the general government sector beginning with the quarter ending in December 2003. The release will contain statistics for the statement of government operations only. Data will be released for the consolidated general government for the reference quarter, and data for previous quarters will be released simultaneously for the subsectors of general government by jurisdiction. It is planned to release these statistics approximately two months after the reference quarter.

VI. Conclusions

18. The adoption of international best practices in the compilation of fiscal data presents some challenges for SDDS subscribers. Fiscal policy analysts will still need monthly data on CGO as an important tracking indicator and will require cash data for liquidity analysis. However, as subscribers move to present fiscal statistics on an accrual basis, it will be even more difficult to meet the existing SDDS monthly timeliness requirements for CGO at the turn of the fiscal year. Moreover, in response to user demands (as demonstrated in the EU member countries and Australia), quarterly data on the general government with quarterly timeliness are becoming increasingly available.

19. In support of the efforts to implement the GFSM 2001 and to pursue quarterly GGO dissemination, the staff has proposed the addition of a targeted timeliness flexibility option for CGO data (whether on a cash, accrual, or modified cash/accrual basis). If a subscriber disseminates quarterly GGO data, with a one-quarter lag, in line with the GFSM 2001 or an equivalent standard, a flexibility option would be available for monthly CGO data. This targeted flexibility option would be allowed for the last month of the fiscal year (up to three months lag) and the first month of the new fiscal year (up to two months lag). In order to make use of this flexibility option, a subscriber would need to begin disseminating quarterly GGO data for at least the last quarter of the fiscal year in which the option is exercised. (For example, when the fiscal year is the same as the calendar year, to use the option for monthly CGO data for December 2003 a subscriber would need to disseminate quarterly GGO data for the last quarter of 2003.)

20. This proposed option facilitates the movement toward international best practices in fiscal statistics, but leaves the SDDS requirements unchanged for those subscribers that are not yet ready to implement accrual recording. This proposal would replace the special arrangements for subscribers implementing accrual-based reporting systems for fiscal data approved by the Executive Board under the Third Review in March 2000.