Report on the Observance of Standards and Codes—Fiscal Transparency Module

This report provides an assessment of fiscal transparency practices in Rwanda against the requirements of the IMF Code of Good Practices on Fiscal Transparency. First, there is a description of practice, prepared by the IMF staff on the basis of discussions with the authorities and their responses to the fiscal transparency questionnaire, and drawing on other available information. Second, there is an IMF staff commentary on fiscal transparency in Rwanda. Important improvements have also been made in budget preparation processes.


This report provides an assessment of fiscal transparency practices in Rwanda against the requirements of the IMF Code of Good Practices on Fiscal Transparency. First, there is a description of practice, prepared by the IMF staff on the basis of discussions with the authorities and their responses to the fiscal transparency questionnaire, and drawing on other available information. Second, there is an IMF staff commentary on fiscal transparency in Rwanda. Important improvements have also been made in budget preparation processes.

I. Introduction1

1. This draft report provides an assessment of fiscal transparency practices in Rwanda against the requirements of the IMF Code of Good Practices on Fiscal Transparency. The assessment has two parts. First, there is a description of practice, prepared by the IMF staff on the basis of discussions with the authorities and their responses to the fiscal transparency questionnaire, and drawing on other available information. Second, there is an IMF staff commentary on fiscal transparency in Rwanda.

II. Description of Practices

A. Clarity of Roles and Responsibilities

2. The coverage of the central government budget is restricted to central government excluding the pension fund. General government, as defined in the IMF’s manual of Government Financial Statistics (GFS) is a larger concept than central government. In Rwanda, general government is composed of central government and 106 districts. The central government budget for 2003 included, as budget titles (or “Votes”), the following: parliament, the presidency, the Prime Minister’s Office, 15 ministries, 11 provinces, and the city of Kigali. The central government’s budget provides transfers to districts, via the Ministry of Local Government and Social Affairs (MINALOC). The central government budget also shows transfers to central government autonomous agencies (établissements publics), such as the National University of Rwanda (NUR), other higher education institutes, various Offices and Commissions, and hospitals/clinics. Revenues from fees of établissements publics are not included in the budget. Nor is direct donor support (e.g., to the NUR). Similarly, any revenues retained by ministries are not captured in the central government budget. The government pension fund (Caisse Sociale Rwandaise, or CSR), is not included in the central government budget; the CSR does, however, provide information to the Ministry of Economic Planning and Finance (MINECOFIN) annually. Four special funds are partly included in the central government budget (see paragraph 5).

3. The fiscal responsibilities of lower levels of government are specified in law. Laws establishing provinces, districts, urban authorities, and the city of Kigali were adopted in 2000 and 20012. The 11 provinces are arms of central government, administered by provincial governors (préfets) appointed by the President of the Republic. Provinces, whose budgets are shown in the state budget presented to the National Assembly, are responsible for executing central government expenditures in the regions. In the government’s decentralization policy, it is envisaged that the elected Councils of the 106 rural and urban districts will be responsible for preparing and executing their own budgets. Within urban districts (which are autonomous in administration and finance), there are 1,544 sectors and 9,175 cells. Although the economic sectors for which districts are responsible are spelt out in law—expenditure assignments include primary and secondary education, health centers, agricultural and forestry activities, road and water networks of the district—the devolvement of responsibility from central to district/town governments is evolving progressively.

4. The budgets of districts are not consolidated at central level. Districts’ domestically generated revenues are mainly from fees, property taxes, and taxes on rental income. To supplement these revenues, in 2002, the government created two funds: an Equalization Fund, which earmarks 1.5 percent of central government revenues for districts’ recurrent budgets, and a Common Development Fund, which is planned to allocate 10 percent of central government revenues to districts’ development budgets. In 2003, allocations were currently less than this percentage, owing to a lack of sufficient qualifying projects, low administrative capacity of local government units, and the lack of a functioning reporting and control system. Revenues raised by districts are not reported to the central government on a regular basis. Nor is direct donor financing of districts; at least one bilateral donor provides substantial resources directly to districts, outside the government’s budget.

5. Not all central government spending is included in the annual budget. In addition to the Common Development and Equalization Funds, two special funds have been established by law—the Road Fund and the Fund for the Victims of the Genocide (FARG).3 All revenues and expenditures of the Road Fund are included in the state budget. For the FARG, resources provided by donors are not captured in the budget. A number of other expenditures are made outside the main treasury account. First, expenditure is made from own revenues of établissements publics (commissions and institutes that receive fees) and from donor resources, outside the budget. For 2003, such revenues, including grants, amounted to 5–6 percent of domestic revenues, with the largest being for a Kigali hospital and various higher-learning institutes. Second, there is spending from ministries’ and provinces’ bank accounts that are replenished by transfers from the central treasury account. The Treasury does not monitor this spending; it records the transfers as “spending.” This affects the timing and classification of actual disbursements, more than their size.

6. The privatization process is transparent. A privatization secretariat was established in 1996. Of the 72 state-owned enterprises slated for privatization in 1997, by end-2002, 43 had been sold, mostly smaller or medium-sized firms. In 2003, the privatization of some of the larger firms, including utilities (telephone, electricity, and gas) and tea companies—which had lagged behind the initial schedule—was to take place. The process of privatization of state-owned enterprises is as follows. After a financial audit of the enterprise, invitations for tenders are widely publicized in the media. The evaluation of bids and recommendations are presented to cabinet for final decision by an interministerial technical committee. Decisions by the government to retain a minority share in enterprises are made on a case by case basis and details of the sale are publicly available. Privatization receipts, net of the costs required to cover restructuring, are held on deposit in a separate account at the Banque Nationale du Rwanda (BNR). At this stage, they are not regarded as budget revenues, since there is no formal policy on the use of privatization proceeds. There are also uncertainties concerning the size and timing of privatization revenues. However, privatization receipts are included in budget reporting.

7. Partial ownership of several commercial banks by the government poses significant fiscal risks, which are not reported in the budget documents. The government remains a significant shareholder in the largest banks.4 Their regulation is by the central bank, the BNR.5 Commercial banks operate with an increasing level of independence; policies of directed lending and government intervention are no longer applied. However, their balance sheets are fragile, with nonperforming loans amounting to 30–40 percent of their portfolios and with most banks not meeting capital adequacy requirements. A restructuring process was beginning in 2003. The 2003 budget allocated the equivalent of 0.6 percent of GDP, with the risk that this amount will prove to be inadequate. Eventual privatization of commercial banks is envisaged.

8. The central bank has day-to-day operational independence for the implementation of monetary policy and it does not carry out quasi-fiscal activities. The BNR is not institutionally independent—it is an autonomous agency under the MINECOFIN. The Act establishing the BNR specifies that it has the role of “advising the government on the definition and conduct of economic and financial policy.”6 Although the government could influence BNR policies, since it appoints the Governor and all Board members and the Minister of Finance appoints the Committee of Auditors that supervises the BNR, it appears that the BNR establishes monetary policies with considerable independence.

9. The formal relationships between the BNR and the MINECOFIN are clear. The BNR Law establishes the basis for coordination between the central bank and the MINECOFIN. First, the government has an overdraft facility at the BNR and may borrow, for up to 240 days, an amount equivalent to 11 percent of previous years’ domestic revenues. Although the BNR charges the government a market-based interest rate on overdrafts, the central bank does not pay interest on cash balances. Second, the BNR acts as the government’s agent for issuing treasury securities. Third, the residual of net profits is transferred to the main treasury account after the close of the fiscal year.7 Coordination between the BNR and the MINECOFIN for monetary/fiscal policy requires more frequent collaboration to reach common understandings on short-term projections of net credit to government (from the banking system) and of government imports.

10. Government activities are mostly distinguished clearly from public nonfinancial corporations, but some quasi-fiscal activities continue. Privatization has steadily reduced the number of public nonfinancial corporations. Legislation has reinforced the distinction between government activities and those of nonfinancial public enterprises,8 reducing the scope for quasi-fiscal activities. Government intervention (by ministerial decree) is restricted to a narrow range of activities, largely the setting of prices of a limited set of sensitive goods and services and the management of strategic stocks for petroleum and food.9 The Multi-Sector Regulatory Law establishes a basic regulatory framework for utilities, in anticipation of steps toward privatization. Whilst the principle of cost-recovery pricing policies for utilities has been accepted, actual prices for electricity, gas, and water are inadequate for placing the relevant utility companies in a strong financial position. For the electricity and gas parastatal (Electrogaz), its management was contracted out to the private sector in 2003.

11. Government regulation of the nonbank private sector is conducted openly and procedures appear to be applied in a nondiscriminatory way. Recent steps to clarify the regulation of nonbank private sector activities include the simplification of business license requirements and revisions to the labor code, which removed restrictions on the movement of labor and employment of women. The state-controlled Chamber of Commerce has been abolished and replaced by an independent body representing private sector organizations.

12. Government equity holdings are not reported. Nonetheless, the MINECOFIN maintains information on actual equity holdings—the government has retained minority shareholdings in a number of nonfinancial enterprises that have completed the privatization process. Also, as noted above, the government has considerable minority share-holdings in commercial banks.

13. A new constitution was adopted by referendum in May 2003; its specifies the main roles of the executive, legislative, and judicial branches. The previous legal framework was the 1991 Constitution and the Fundamental Law of August 1993 and its amendments. In the 2003 Constitution, a multiparty system of government is envisaged. Elections are expected later in 2003. The legislature is to consist of two chambers, with a senate formed by appointments or indirect elections.10 The role of parliamentary committees and their internal rules relevant for the budget process will depend on organic laws, yet to be adopted. The role of the senate in budgeting is unclear: the constitution does not explicitly state that that the senate approves the annual budget, although the senate is required to provide an opinion on the annual budget to the President of the Chamber of Deputies. The executive is headed by an elected president of the republic, who is also head of state. He/she appoints a prime minister and other cabinet ministers (upon the proposal of the prime minister).

14. Parliament has the right to modify the composition of the proposed budget, but not to change the overall fiscal deficit. The 2003 Constitution allows parliament to modify the executive’s draft budget. However, if amendments have the potential to reduce government revenue or increase state expenditures, they must indicate accompanying proposals for raising the required revenues or making savings equivalent to the anticipated expenditures.

15. The legal framework for budget management is incomplete and outdated, but new laws are envisaged. In early 2003, the main legal basis underlying the budget process was the 1979 Decree Law on the Budget and Public Accounting (confirmed by Law No. 01/82 of 1982). Some of its unclear provisions were also included in the annual budget laws of recent years, in an attempt to specify the accountabilities of those involved in the budget preparation and execution processes. To remedy these deficiencies, in mid-2003, the government was developing a new legal and regulatory framework for fiscal management. In particular, as required by the 2003 Constitution, an Organic Budget Law (OBL) was being drafted.11 This will include certain desirable practices—for example, the preparation of a medium-term budget framework (MTBF)—which had already been developed in the absence of a formal legal framework. Modifications to the 1998 external audit law are also to be drafted in 2003. Once these new laws are adopted, new financial regulations and accounting instructions are needed to clarify the financial management framework. Currently, the financial management regulations applicable to central and local governments are either unclear or inconsistent, making it impossible to derive nation-wide government accounts.

16. Institutional arrangements for fiscal management are evolving, in an effort to clarify responsibilities. The MINECOFIN—formed from the merger of the Ministry of Finance and the Ministry of Planning in 1997—has the responsibility of managing the country’s budget and finances. The organizational structure of the ministry is laid out in Key divisions include: the budget department (recurrent budget preparation, authorization of commitments, and cash planning); the treasury department (authorization of payments, management of government bank accounts); and the public accounts department (maintaining the public accounts, collecting nontax revenues). A semiautonomous agency, CEPEX, formed from the Investment Directorate from the former Ministry of Plan, handles the preparation and monitoring of the development budget.12 The Rwanda Revenue Authority (RRA), an autonomous agency established by law in 1997,13 collects all the major taxes, but not nontax revenues.14 Both on the expenditure and tax sides, responsibilities and accountabilities under the legal and organizational arrangement prevailing in early 2003 were unclear.

17. Taxes and exemptions have a legal basis, but some tax laws are unclear. The constitution states that no taxation can be imposed, modified or removed except by law, and that no exemption from or reduction of tax may be granted unless authorized by law. All the major taxes, and procedures to collect them, are based on specific laws that are readily accessible.15 The VAT Law, adopted in 2001 to replace sales taxes, is clear and supported by extensive administrative documentation. In contrast, the income tax law is not always clear. Also, the absence of national accounting standards makes it difficult to unambiguously determine taxable profit. More generally, the adoption of three official languages16 makes it difficult to interpret tax laws. Although the criteria for guiding the administrative application of some tax laws have become clearer since the establishment of the RRA, it is uncertain how well tax laws are applied in practice.

18. The “investment code” provides discretion in according tax advantages to investors or to specific beneficiaries (e.g., NGOs, with respect to duty-free importations). Tax incentive schemes, aimed at accelerating investment, are provided by the Rwanda Investment Promotion Agency (RIPA). Tax incentives are currently included in the “investment code”—Law No. 14/98 establishing the RIPA. There remains some scope for discretion in determining eligibility for investment incentives, including which businesses can operate in duty-free zones. The Minister having jurisdiction over commerce and industry has considerable powers under the RIPA Act, thereby weakening the powers of the Minister responsible for finance to oppose new tax concessions under the RIPA Act.17

19. While it has no formal legal protection, it was indicated that the government has not interfered with the operation of the national tax administration. While the RRA is financed from the central government budget, it has been granted considerable independence in its operations and maintains a separate salary structure subject to the approval of the RRA’s Board of Directors. While taxpayer rights are not embodied in law, a broad statement of taxpayer rights is posted on the RRA website and in RRA offices.

20. A code of behavior, applicable to all civil servants, has not been prepared. In most civil service employment contracts, there are no explicit provisions relating to ethics. However, in the case of the RRA, a Code of Staff Conduct and Disciplinary Procedures was issued in June 2001. This code includes a requirement to declare asset holdings. Although the government has discussed a national program for good governance,18 there are as yet no formal anticorruption program or legislation.

B. Public Availability of Information

21. The budget documents presented to parliament cover most central government fiscal activities, but do not provide revenues and expenditures of general government. The budget is set in the context of the Poverty Reduction Strategy Paper (PRSP), a publicly available document. Also, the budget speech by the Minister of Economic Planning and Finance is published on the MINECOFIN’s website. However, published material does not as yet contain details of expenditures by autonomous agencies, nor the budgets of lower levels of autonomous governments, notably the 106 districts (even in aggregate form).

22. The budget documents provide parliament with a medium-term analysis of the main budget targets. Besides the draft annual Budget Law, parliament receives a “Background to the Budget” document that lays out recent economic performance and the MTBF, including projections of the main macroeconomic aggregates (real economic growth, inflation, and so on) and the major budgetary aggregates—revenues, expenditures, the overall deficit, and deficit financing. A summary table shows the main aggregates for the new budget year, the two years preceding the budget year, and the two years following the budget year. Annexes to the “background to the budget” document include PRSP priority programs, analysis of recurrent expenditures, details of the medium-term expenditure projections and, since 2003, off-budget revenues. There is, however, no discussion of the evolution of aggregate debt—a key indicator for medium-term fiscal analysis.

23. Defense expenditures are treated in some detail in the budget documentation. Like other ministries, the defense budget does not capture the ministry’s own revenues, which are reported to be small. In 2001, expenditures from the Ministry of Defense’s foreign exchange account at the BNR amounted to about 13 percent of its total budgeted expenditure. Like the bank accounts of other spending ministries, this account is replenished from the main treasury account at the BNR (see paragraph 44).

24. Information on contingent liabilities is not reported. The existing legal framework does not require this information to be made available to parliament or the public. The MINECOFIN maintains information on loan guarantees granted by the central government. Other explicit contingent liabilities, such as indemnities or legal payments, are not monitored systematically. Associated with decentralization, there are implicit contingent liabilities—central government financial assistance may be required should a district default on its borrowing. Currently, the MINECOFIN has very little control in this area.19

25. Estimates of tax expenditures and quasi-fiscal activities are not published. Nonetheless, the MINECOFIN has established a database that includes estimates of tax expenditures from import duties, excises, and the VAT. The database tracks the legal basis for each exemption granted under existing laws. Quasi-fiscal activities do not appear to be extensive—the government does not, in general, use state-owned enterprises for fiscal policy purposes (for exceptions, see paragraph 10).

26. Information on gross public debt and government assets is not published. The authorities maintain comprehensive information on gross public debt. However, neither external nor domestic debt data are published. In the absence of a viable public accounting system, complete information on government financial assets is not available. An inventory of selected nonfinancial assets was being conducted in 2003, but asset valuation issues will not be addressed in the immediate future.

27. Fiscal data are now published on the MINECOFIN website. Beginning at end-October 2002, the authorities began publishing quarterly data on budget execution, the aim being to publish such data with one month’s lag.

28. There are no formal commitments for regular publication of fiscal data and advance release date calendars are not announced. Existing laws, dating from the time of a one-party government, do not set out reporting requirements. However, the new OBL, once adopted, is expected to lay out fiscal reporting requirements. Rwanda is not yet meeting the IMF’s general data dissemination standards, owing in part to the need to reestablish basic statistical capacity.

C. Open Budget Preparation, Execution, and Reporting

29.The annual budget has traditionally focused on financial compliance, but in recent years, a program budget approach has been initiated. The budget preparation process is guided by a budget circular that establishes expenditure ceilings for ministries and provinces, based on a medium-term macroeconomic framework. The program classification that has been adopted since 2000 has allowed the identification of priority and nonpriority budget programs. These are consistent with the broader sectoral priorities identified in the PRSP. Annexes to the annual budget law show revenues by economic classification and expenditures by administrative, program/subprogram, economic, and functional classifications, which are broadly consistent with GFS international standards. Presently, the information provided is not sufficient to assess the extent to which performance is consistent with stated policy objectives.

30. The recurrent and development budgets are not integrated. However, some steps are being taken toward integrating the two budgets. In particular, the budget document now provides details of both the recurrent and development budgets. The development budget was previously limited to highly aggregated programs and projects. It now shows programs/subprograms, donor, and type of funding (domestic budget, loan, grant). The development budget coverage is complete for loan-financed projects, but some grant-financed projects, especially those financed by NGOs, are excluded. Whereas the Budget Department of the MINECOFIN coordinates the preparation and the early part of execution of the recurrent budget, CEPEX plays this role in the case of the development budget. Separate databases are maintained.

31. The domestic balance and the overall balance of central government are the main indicators of the fiscal position. The domestic balance captures revenues and expenditures under the control of the central government—it excludes the disbursement and use of foreign resources. The general government overall balance—the aggregate of central and local government balances—is not monitored, as an adequate reporting system of districts’ budget execution has not been developed. The overall public sector balance—the general government balance plus the net balances of public enterprises—is not examined.

32. Medium-term projections of each expenditure program are prepared in the context of the MTBF. Experience thus far with such projections indicates that the second-year projections do not automatically become the first year of the new MTBF twelve months later. This is mainly because: (i) revenues, especially grants, and/or the macroframework for the first year, may deviate considerably from initial assumptions; (ii) the medium-term expenditure projections do not include a formal planning contingency reserve; and (iii) fixed costs are very high, thereby reducing flexibility for changing nonpriority expenditures and remaining within overall ceilings. In particular, the PRSP details policy objectives and priority programs, which results in rigidity for most goods and service spending. Government salaries and employment are determined centrally and cannot be changed easily. Also, debt servicing and other costs (e.g., food for prisoners, army) are also “fixed.” As a result, most spending is predetermined.

33. New policies are clearly described in the medium-term budget framework paper. In the budget submission process, spending ministries are invited to prepare a mission statement, a strategic overview, policy objectives, key issues facing the ministries, constraints to achieving objectives, and proposals for dealing with them. This may lead to the identification of new policies, which are discussed in cabinet, then parliament. Despite these procedures, other new policies have been introduced late in the budget cycle, irrespective of whether they are in the budget framework. For example, in preparing the 2003 budget, the initial expenditure ceilings were overridden by 10 percent.20

34. The main fiscal risks are not fully disclosed in the budget documentation. Discussions of fiscal risk are limited to the potential shortfall of domestic revenues and donor assistance. More comprehensive analyses of the sensitivity of the estimates to changes in economic variables or other risks are unavailable. The annual budget’s contingency reserve is very small and would not be able to handle expenditure increases necessitated when major risks are realized.

35. Internal control, although administratively cumbersome, appears to be effective for controlling expenditures. Every expenditure commitment requires a priori control (a “visa”) by the Budget Department of the MINECOFIN. After goods and services are received by spending ministries, a second control is exercised by the Treasury. With some exceptions, payment is made by the BNR from the Treasury’s main account. Since the implementation of quarterly cash limits in 2001, this rather heavy system of centralized a priori expenditure checks appears to have brought expenditure arrears under control. Also, the expenditure process has been computerized—in the so-called “SIBET” system—with online links with all central government ministries in Kigali and similar links being established with provinces. This will facilitate electronic expenditure control.

36. Internal audit in spending ministries is weak and focuses on pre-audit of documentation. Internal audit units, generally consisting of only one person, have recently been established in spending ministries. However, instead of performing the internal audit function—checking that financial regulations are being respected, reporting to management, and suggesting remedial measures if regulations are not respected—the units are largely performing an internal control function: checking documentation required for the a priori controls on spending commitments. At the level of the MINECOFIN, besides its own internal audit unit, there is a high-level office for Inspection Générate et Audit (IGA). Its mandate is to conduct audits of any public entity and to report the results of its inspections to the Minister of Economic Planning and Finance. However, its work appears to be largely overlapping with that of the external audit body, the Office of the Auditor General (OAG). With the help of the OAG, a draft Manual of Financial Management Procedures for central government was developed in 2002. This does not include reference to the SIBET system. Moreover, its approach to public financial management is not entirely consistent with another manual, prepared under MINALOC’s aegis, to guide financial management in districts.

37. The National Tender Board (NTB) has been operating since 1998. The legal framework to support its operations—including procurement regulations—was submitted to parliament for consideration in early 2003.21 The NTB has been operating under a presidential decree and has issued and implemented procurement guidelines consistent with international standards. The NTB oversees the tendering of all larger procurement contracts, and assists ministries, autonomous agencies, provinces, and districts in building procurement capacity and implementing proper procedures. Staff have been trained at both central government and district levels, although additional resources may be required to ensure the smooth functioning of local governments’ financial management.

38. The accounting system is neither comprehensive nor integrated. In an effort to develop the public accounting function, the Department of Public Accounts was split from the Treasury Department in 2001, and public accountants have been placed in all line ministries and provinces. Although it was envisaged that the SIBET system would be extended to include an accounting module, in early 2003, this had not been done. Moreover, two quite different charts of accounts have been drafted—one for central government and another for districts. For central government, the draft chart of accounts is based on the nine classes of accounts typically used in francophone countries. For local governments, a simplified accounting system, based on an accounting framework typically used in anglophone countries, was being implemented in districts in 2003. Until a harmonized double-entry accounting system is operational at all levels of government, a reliable basis for assessing payment arrears cannot be provided.

39. The fiscal situation is monitored by a monthly “flash reporting” system. These reports are not derived from a double-entry accounting system that comprehensively covers all central government transactions. Rather, monthly “flash reports” are compiled from statistical information derived from a number of sources within the MINECOFIN, its autonomous agencies, and the BNR. Although information on revenues, current expenditures, and deficit financing are generally readily available, a major shortcoming is the periodicity and timeliness of development budget expenditures financed by donors. Although reporting requirements are known, project managers and/or donors fail to provide timely and comprehensive information on actual disbursements. The recurrent and development budgets are handled by separate institutional units, and reporting systems have not been synchronized. Actual expenditures by provinces are also not captured in the accounting system—only transfers to provinces.

40. A mid-year examination of the budget is undertaken; this may lead to modifications to annual budget authorizations late in the year. The timing with which the supplementary budget is presented to parliament has been variable. In general, the draft supplementary budget is presented toward year-end—it is more a proposal to legitimize virements in actual expenditures, rather than a plan for the second half of the year. Present legislation does not require the government to present a formal mid-year report to parliament, with or without a supplementary budget.

41. A consolidated set of accounts for central government has not been audited for over a decade, thereby impeding the effectiveness of the external audit agencies. The consolidated accounts for central government were last formally verified (“judged”) by the Court of Accounts in 1989. Final accounts for 1999 were prepared by the MINECOFIN and submitted to the Court of Accounts, but these were not presented to the legislature. The Auditor General (AG) has conducted audits of a number of ministries, agencies, and public enterprises (see paragraph 46). Like the Court of Accounts, the AG has been thwarted in his work of auditing the consolidated annual accounts of central government.

D. Assurances of Integrity

42. Statements of accounting policy are not included in the budget. The accounting system is essentially cash-based. On the expenditure side, the SIBET system records expenditures at each step (commitment, payment order, and cash payment). In principle, there is no end-year carryover for unpaid commitments or outstanding payment orders.22 In practice, unexecuted payment orders at year-end can be substantial, partly due to difficulties in forecasting available cash balances. In the absence of a viable accounting system and the nonpreparation of consolidated annual accounts, it is unclear whether there is a de facto complementary period for payment of previous-year bills.23

43. The data on which the annual budget projections are based are generally reliable for revenues and recurrent spending. Earlier tendencies to overestimate budgeted revenues and the attendant gaps between the targets set in the budget appear and actual deficits have diminished. Indeed, tax revenues have even exceeded budget projections recently, reflecting conservative estimates of expected gains from new revenue measures and the positive impact of administrative reforms. The outcomes for current expenditures have generally been broadly in line with budget projections. In contrast, the development budget projections are a poor guide to actual disbursements, which generally fall well below budget projections. This mainly reflects the delays in the disbursement of foreign financing and the lack of capacity to implement projects. Until full financial accounts of government are prepared, it is impossible to provide parliament with full explanations of the reasons for the divergences between budgets and outcomes.

44. Regular reconciliation of all government bank accounts is not done. As a result, there are discrepancies between the monetary accounts and fiscal accounts, casting a doubt on the quality of reported revenues and expenditures. At end-2001, the nearly 1,500 government bank accounts were maintained, being denominated in both Rwanda francs and in foreign currency. About half were held at the BNR and the others were in commercial banks. In principle, ministries and public bodies are prohibited from opening bank accounts in private banks without authorization from the Minister of Finance. However, some accounts have been opened without this authority. The main treasury account (the so-called OTR account)24 is used for the major transactions; it is regularly reconciled with accounting records. The three main revenue accounts—transit accounts of the RRA prior to the transfer of revenues into the OTR account—are not reconciled regularly, although in early 2003 the Treasury was planning to do so. None of the other accounts, which include dormant accounts, many active project accounts, ministries’ accounts, provinces’ accounts, autonomous public agencies’ accounts, are reconciled with accounting data on a regular basis.

45. The quality of expenditures recorded in monthly flash reports is undermined by recording transfers as expenditures. Ministries and provinces have been permitted to open their own bank accounts, out of which they pay for certain expenditures (e.g., mission travel, electricity). These accounts are replenished by transfers from the Treasury’s OTR account. However, the Public Accounting Department currently records the transfer as an expenditure, before the payment to a nongovernmental entity has taken place.

46. The national audit body is currently not independent of the executive, nor does it provide timely reports to the legislature and public on the financial integrity of government accounts. The 1991 Constitution endorsed the Court of Accounts—under the judiciary branch of government—as the external body of the country. In 1998, the OAG was established by law.25 During 1998–2003, both external audit bodies were unabled to perform their principal task of auditing the central government’s consolidated annual accounts, since the Public Accounting Department of the MINECOFIN had generally not prepared draft annual accounts. The 1998 OAG law calls for the AG, who is appointed by the President of the Republic subject to the approval of cabinet, to deliver his/her annual report to the President of the Republic. There was no obligation to publish the report. The 2003 Constitution transfers the external audit function from being under the executive/judiciary branches to one primarily serving parliament. In particular, parliament is to become the primary recipient of the OAG’s annual report.

47. Since its establishment in 1998, the OAG has been strengthened and has audited several individual agencies. However, the legislature has not yet been called upon to follow up on external audit reports. Thanks primarily to donor-funded training programs, the OAG has expanded its qualified staff considerably. During 2002, the OAG audited 34 public sector entities, including 10 ministries and 13 projects. In line with the 1998 OAG Act, these audits were submitted to the President of the Republic, the Speaker of the National Assembly, and the Chief Justice of the Supreme Court. Although the AG’s reports were reviewed by the parliamentary budget committee, they were not made available to all parliamentarians and the public. Although the new constitution obliges public officials to implement the recommendations of the AG, the mechanisms for enforcing this have yet to be worked out. Moreover, the constitution does not explicitly state that the OAG’s annual report must be published.

48. External scrutiny of macroeconomic models and assumptions underlying the budget are not carried out. Background papers for the budget describe the broad assumptions used in budget projections. However, the underlying assumptions and behavioral relationships underlying the projections of the aggregates are not made available to parliament. At a detailed level, most expenditures are budgeted on an incremental basis, without explicit assumptions for the various programs. Revenues are projected tax by tax, based on assumptions for the tax base, tax rates, changes in exemptions (if any), and improvements in tax administration.

49. The national statistics agency is not institutionally independent and does not verify the quality of fiscal data. Most statistical information is compiled by the Statistics Department of the MINECOFIN. In 2002, the department prepared a strategic plan for the development of a national statistical system, covering all areas, including government financial statistics.26 The absence of a formal legal framework for national statistics and the department’s lack of financial autonomy are noted in the strategic plan. Unlike the tax collection function, where an independent agency has been set up, the statistics collection and dissemination function is still an integral part of the MINECOFIN. However, the day-to-day work of the department proceeds without political interference.

III. IMF staff Commentary

50. The authorities have taken a number of actions that have clarified the role of government. The starting base in the mid-1990s was very low, as a lot of the institutional capacity for effective budget management was destroyed during the tragic events of 1994. Since then, the state has been withdrawing from commercial sectors, thereby limiting the role of government in the economy to one focused primarily on the provision of essential public services. The PRSP process has been helpful in this respect, as it has clarified the priorities and choices of sectors where the government is to intervene. Privatization of state enterprises has been done in a transparent manner.

51. Important improvements have been made in budget preparation processes. Following the devastating events of 1994 and until 1997, it was not possible to present an annual budget to parliament before the start of the fiscal year. The institutional capacity to prepare, execute, and report on budgetary developments had to be rebuilt. Since 1998, an annual budget has been adopted by parliament. As from 2000, the budget classification structure has been improved and expanded, in line with international standards. The detailed recurrent expenditures are now classified by administrative unit, program/subprogam, and economic and functional categories. Efforts have been made to present the development budget on the same basis. These improvements allow stronger links to be built between budget appropriations and the overall policy objectives of government, as incorporated in the PRSP. Medium-term expenditure projections, consistent with available resources, have been introduced. These enable stakeholders in the PRSP process to identify changes in government outlays over a three-year time frame.

52. More information on public finances is available publicly. A number of documents, such as the PRSP, the Minister of Finance’s budget speech, and quarterly budget execution reports, are available to the public. Accompanying the annual budget law is a clearly written background document that provides parliament with essential information on the orientation and objectives of the annual budget. This lays the basis for focused policy debates in the National Assembly. Both branches of government have a number of websites, resulting in an increasing amount of information being made available to the public. The computerization of expenditures, from appropriation to payment, has improved the availability of expenditure data.

53. The new constitution, adopted in May 2003, paves the way for direct presidential and legislative elections, as well as providing a firmer basis for the budget process. The staff welcomes the constitution’s provisions relating to the restriction of parliament’s power to change the budget—namely that it cannot introduce revenue-reducing or expenditure-raising measures unless it also accompanies them by expenditure-reducing or revenue-increasing measures. This is a transparent provision that enhances fiscal discipline. The new constitution, by not referring to the Court of Accounts, effectively abolishes this body, whose role was similar to that of the AGO. The reorientation of the OAG, from a body under the presidency to one primarily serving parliament, is in line with the IMF Code of Good Practices on Fiscal Transparency. It will be important to ensure that the AG’s annual reports are readily made available to the public.

54. Although these accomplishments are significant, much remains to be done to bring Rwanda up to a high standard of fiscal transparency. The most urgent issue is to address the pervasive weaknesses in government accounting. It is of great concern that a consolidated annual report of government financial transactions has not been audited since the Court of Accounts’ report covering 1989. Without timely annual reports, based on solid accounting data, the external audit function cannot operate satisfactorily, and the public is deprived of knowing what is happening with taxpayers’ monies. The staff does not consider the institutional changes within the MINECOFIN—the splitting of the Treasury and Public Accounting Departments—as necessarily helpful for consolidating the accounting function. There has also been undue focus on completing an inventory of physical assets, which detracts from the more-pressing need to prepare comprehensive, timely, and quality financial data on budget execution. In order to build a comprehensive and integrated accounting system, it will be essential to:

  • Decide on the type of accounting system that is best suited to Rwanda’s needs, especially in view of the increasing transfer of responsibility for budget execution and accounting to spending ministries, provinces, and districts.

  • Adopt a simplified chart of accounts for use in both central and local governments.

  • Use ledger accounts for each transaction, thereby allowing the use of public monies to be traced and provide a basis for external audit of annual accounts.

  • Introduce a computerized accounting, in the context of an integrated financial management information system. This would include extending the SIBET system to all provinces and using it as a tool for expenditure control.

  • Leave aside for the present time, the reporting of the value of nonfinancial assets.

  • Strengthen the Public Accounts Department, to enable it to exercise a strong oversight responsibility, not only for enhancing the skills of accounting personnel, but also for setting and maintaining accounting standards.

  • Recruit and train accountants, ensuring that remuneration is competitive.

55. Closely associated with the need for producing and disseminating high-quality data is the need to rationalize the excessive number of government bank accounts. In 2002, there were nearly 1,500 government bank accounts, which is excessive for a small country. The government has already begun a process of rationalizing government banking arrangements, starting with the closure of dormant accounts and strengthening treasury oversight of all bank accounts. The staff encourages this process to continue, by channeling more and more payments through the main treasury account. In particular, ministries’ and provinces’ imprest accounts in banks should be closed and replaced by ledger accounts, with payments being made directly from the OTR account at the BNR. Even prior to this action, transfers to imprest accounts should not be recorded as expenditures.

56. There is a need to increase the coverage of fiscal data. The authorities are fully aware of the longstanding problem of incomplete information regarding donor-financed expenditures, especially for the development budget. As decentralization proceeds, this problem will also become acute at district level, unless the problem is addressed decisively. A second area where data coverage needs improvement is the reporting of all revenues, expenditures, and indebtedness of ministries, provinces, autonomous agencies, and districts. Mechanisms are needed to ensure regular reporting to the MINECOFIN, including balance sheets, especially for autonomous agencies.

57. The improvement of data quality and the provision of additional fiscal data to the public are highly desirable. The progressive incorporation of central government entities into a treasury single account arrangement should be envisaged. Such an arrangement facilitates the much-needed reconciliation of banking and accounting data, without which the quality of fiscal data may remain low. Although basic data quality standards should be met prior to publication, the staff is not opposed to the publication of lower-quality data, as this could possibly stimulate improved data quality. Some data that are already available, such as public debt data, could be disseminated more widely immediately. This could be followed by fuller reporting of contingent liabilities and estimates of tax expenditures. Reporting of all government assets would be useful, but this will take time. Identification of quasi-fiscal activities, and reporting them to parliament, should also be a longer-term objective.

58. There is a need to establish a strong internal audit function and clarify the role of the government inspectorate. The effectiveness of internal audit units in spending ministries and provinces needs to be improved. This could be achieved by reorienting their current focus on internal compliance control to internal audit, i.e., ensuring the effectiveness of management systems in spending ministries and provincial administrations. Internal auditors play a key role in ensuring adherence to the centrally issued financial management and accounting procedures, including reporting obligations. The IGA could reorient its work away from that currently done, and place greater emphasis on developing guidelines and procedures for internal audit. The IGA could also train, inspect, and supervise internal audit staff in spending ministries and provinces. Effective internal audit at local government level is also needed. The role of the IGA in this area is not yet defined. The IGA could play a role at district level rather than being involved with inspections of public enterprises, which could be left to the AG or enterprises’ own external auditors.

59. The role of parliament and its procedures for budget-making need to be clarified and developed. The role of the senate in the budget process, and the coordination mechanisms with the Chamber of Deputies, are yet to be spelt out. Although the senate is competent to vote on laws relating to defense and security (Article 88 of the constitution), it is unclear whether it is to vote on the annual budget in its entirety. It is important that each chamber of parliament quickly adopt the organic laws establishing their own internal regulations (Article 73 of the constitution). These will specify the powers of the Bureau of each chamber, as well as the number, duties, and powers of various parliamentary committees. The mechanisms for resolving potential disputes of the two chambers over the budget also need to be specified. Presently, it is unclear how the Chamber of Deputies would incorporate in the budget an opinion of the senate that differs from its own.

60. With the adoption of the 2003 Constitution, there is a need to adopt clear laws to underpin budget management. The first steps in this process is to complete the envisaged drafting of a new Organic Budget Law and modify the 1998 Act establishing the AGO. The OBL would clearly specify the powers of the Minister of Finance and establish clear accountability provisions for the key players in budgetary management. In particular, the roles of ministers, who decide on budgetary policies, and those of secretary generals of ministries, who are responsible for implementing the government’s agreed policies, need to be clarified. As part of their budget management responsibilities, a secretary general should ensure that the accounting function in their ministry is effective—to enable the preparation of the ministry’s annual report. This is important, since the head of spending ministries and provinces need to be accountable before parliament’s budget committee(s).

61. The legal framework needs to be accompanied by clear financial regulations and accounting instructions. A separate public accounting law is not required. Rather, the OBL should specify that the minister of finance is authorized to issue clear financial regulations to govern budget execution, accounting, internal control, and reporting. The public accounting framework, including the chart of accounts and the procedures/instructions on how to implement them at all levels of government, should also be under his authority.

62. The OAG’s report should be available to the public. The 2003 Constitution does not explicitly require the publication of the AG’s report—it merely requires its submission to parliament. A publication requirement could be made explicit in the revisions to the 1998 Law establishing the OAG. This law also needs to be modified to specify that the AG is appointed by parliament, not the president of the republic, and that he/she reports to parliament.

63. The OAG and parliamentary committees need to be strengthened. Although remarkable progress has been made in training staff of the OAG since its establishment in 1998, it is essential that the OAG has adequate capacity to audit each ministry’s accounts. The audits should not just examine financial compliance, but also ways that a ministry’s management could be improved to achieve the objectives of economy, efficiency and effectiveness. It is even more important, given the steps taken toward introducing democracy in Rwanda, to ensure that the parliamentary committees of the two chambers are adequately resourced—not only to ensure adequate follow up of the AG’s reports, but also to provide analysis of proposed annual budgets.

64. Fuller information should be provided to parliament, including full information on budget outcomes and explanations of why they deviate from the budget appropriations. A fuller discussion of program outcomes would be facilitated if the recurrent and development budgets were fully integrated. In the longer term, information on fiscal risks and sensitivity analysis of the budget projections to deviations in key assumptions would be desirable.

65. Although the immediate emphasis should be on improving budget execution, especially government accounting, refinements to the budget preparation process would also be useful. Greater efforts need to be made to plan expenditures in advance, rather than adopting changes in them in cabinet at a late stage of the budget cycle. This could be addressed by including a planning contingency in the budget preparation process. Initial spending ceilings would exclude a fixed amount (e.g., 2 percent of total expenditures). Later, when the draft budget is submitted to cabinet, ministers could increase expenditures for urgent needs not previously foreseen, by allocating the contingency set aside for this purpose.

66. Emphasis should gradually turn toward accounting for performance of budget programs, answering questions such as “were program objectives achieved?;” “Are additional budget resources needed to achieve (sub)program objectives?;” “what is the cost?;” and “is it affordable within the revenue constraint?” To enhance transparency in this area, the development of performance indicators should be envisaged. However, given the need for additional administrative and human capacity, a gradualist approach in this area is suggested.

67. Financial management systems of local government finances need further development. To address weaknesses at local government level,27 donors are assisting districts to improve budget preparation and management. However, this is being done in the context of unclear roles of the MINALOC and the MINECOFIN. It would be useful to clarify each ministry’s responsibilities for ensuring sound financial management, including nationwide fiscal reporting. It would also be useful to subject districts’ borrowing to firm control by the minister of finance. Additionally, since the financial management and accounting systems being put in place at district level are not compatible with those of central government, it is urgent that steps be taken to harmonize the two disparate systems.

68. The staff emphasizes the importance of “getting basics right first.” The importance and urgency of developing the accounting system is stressed, along with the need to establish a new legal framework to accompany the 2003 Constitution. The provision of timely, comprehensive and high-quality fiscal data is the top priority. Expenditure control mechanisms, and internal and external audit also need to be strengthened. Only at a much later stage should there be a move toward adopting accrual accounting or output-oriented budgeting that the most-advanced countries of the world are still refining. In the meantime, strong fiscal discipline at all levels is needed to achieve the overarching objectives of the authorities as specified in the PRSP.


Initial discussions on fiscal transparency were held in Kigali in November 2001 by a mission composed of Ms. D. Rehm and Ms. C. Roehler. Discussions between the authorities and Ms. Rehm continued during March/April 2002. This report was updated and completed by Mr. I. Lienert in March 2003.


Law No. 43/2000 Establishing the Organization and Functioning of Provinces; Law No. 4/2001 Establishing the Organization and Functioning of Districts; Law No. 05/2001 Establishing the Organization and Functioning of Urban Authorities; and Law No. 07/2001 Establishing the Organization and Administration of the City of Kigali.


The Road Fund was established in 1989 by Law No. 20/89 (as amended). The National Fund for Assisting Needy Victims of the Massacre and Genocide Committed in Rwanda during December 31, 1994 to October 1, 2000 was established by Law No. 02/98 of January 1998.


The government has a 49.4 percent share in the Banque Commerciale du Rwanda, a 50 percent share in the Banque du Kigali (including ownership by government-owned entities), a 4 percent share in the Banque Continentale Africaine du Rwanda. The government also has a 47 percent direct stake in the Banque Rwandaise du Developpement, which is also partly owned by government-owned enterprises or banks.


Law No. 26/06/99 Relating to Regulations Covering Banks and Other Financial Institutions.


Article 4 of Law No. 26/07/1997 establishing the BNR, whose website is


According to the BNR Law, 20 percent of the BNR’s net profits are allocated to a general reserve fund and a further portion—decided by the discretion of the Board of Governors—is dedicated to special reserves.


Guiding legislation includes Decree-Law No. 39/75 regarding Public Enterprises.


Law No. 15/2001 Amending and Completing Law No. 35/91 concerning the Organization of Internal Trade.


A new OBL, together with a new Public Accounting Law, were drafted in the 1999–2001 (to replace the 1979 Decree Law). However, they were withdrawn from parliament, pending adoption of the new constitution.


The Central Projects and External Finance Bureau (CEPEX) was created in the late 1990s, largely taking on the responsibilities of the Investment Directorate of the merged ministries of finance and of planning.


Law No. 15/97 Establishing a Rwanda Revenue Authority, November 1997.


A decision has been made to transfer the responsibility for collecting nontax revenues to the RRA. The public accounting department inherited this role following the splitting of this department from the Treasury.


The main laws regarding taxation are: (i) Law No. 08/97 Direct Taxes on Different Profits and Professional Income (June 1997) as amended; (ii) Law No. 06/01 Law on the Value-Added Tax (June 2001); (iii) excise tax laws; (iv) Law No. 07/98 Duties on Imported Products (August 1998) as amended; (v) Law No. 09/97 Code for Fiscal Procedures, which covers tax collection and penalty provisions.


The 2003 Constitution specifies that Kinyarwanda, English, and French are official languages. All laws presented to parliament are in these three languages.


The MINECOFIN has only one representative on RIPA’s governing board.


“National Program for Strengthening Good Governance for Poverty Reduction,” MINALOC, May 2002.


Districts are allowed to borrow for their development budgets (Article 118 of Law No. 04/2001). However, the minister of local government, not the minister of finance, fixes regulations governing the control of financial activities of districts (Article 143 of same law).


The additional expenditures were mainly for the cost of financing the constitutional referendum and subsequent national elections (additional expenditures for demobilizing troops had already been foreseen).


The NTB has been operating for three years despite the lack of legal provisions for its operations and the consequent inability to impose penalties. The NTB’s website is found at and contains information on outstanding tenders and past awards, and procurement procedures.


The financial regulations state that expenditure commitments are to be closed on November 15 and payment order issuance on December 15.


The financial regulations require any unpaid bills relating to a previous year to be paid out of the arrears allocation of the new budget year, in line with strict cash accounting principles.


OTR = Ordonnateur-Trésorier du Rwanda, who has the responsibility to verify payment order requests from spending ministries and to instruct the BNR to pay the suppliers of goods or services. The OTR is the head of the Treasury Department.


Law No. 05/98 establishing the OAG.


See Plan Stratégique de Developpement du Système Statistique National, August 2002.


See “Fiscal and Financial Decentralization Policy,” a document prepared jointly by the MINALOC and the MINECOFIN.

Rwanda: Report on the Observance of Standards and Codes—Fiscal Transparency Module
Author: International Monetary Fund