Rwanda
Technical Assistance Report-Government Finance Statistics
Author: International Monetary Fund. Statistics Dept.
Publication Date:
28 Mar 2018
eISBN:
9781484349021
ISBN:
9781484349021
Language:
English
Keywords:
ISCR; CR; government; data; work program; GFS compilation process; GFS data; AFR's mission representative
Download PDF (895.9 KB)

This Technical Assistance Report discusses the findings and recommendations made by the IMF mission to assist authorities in Rwanda in aligning the compilation and dissemination of government finance statistics (GFS) in accordance with the Government Finance Statistics Manual 2014. It was recommended that the Ministry of Finance (MINECOFIN) should officially assign GFS compilation duties to a specific subset of its staff. The mission believes that it is important for the MINECOFIN to make a “permanent” assignment of GFS compilation and dissemination duties so that the work can evolve systematically. In the mission’s view, the Macroeconomic Policy Unit appears to be well placed to perform this task. However, the unit should be allocated appropriate resources to take on the responsibility.

Abstract

This Technical Assistance Report discusses the findings and recommendations made by the IMF mission to assist authorities in Rwanda in aligning the compilation and dissemination of government finance statistics (GFS) in accordance with the Government Finance Statistics Manual 2014. It was recommended that the Ministry of Finance (MINECOFIN) should officially assign GFS compilation duties to a specific subset of its staff. The mission believes that it is important for the MINECOFIN to make a “permanent” assignment of GFS compilation and dissemination duties so that the work can evolve systematically. In the mission’s view, the Macroeconomic Policy Unit appears to be well placed to perform this task. However, the unit should be allocated appropriate resources to take on the responsibility.

I. Executive Summary

A government finance statistics (GFS) technical assistance (TA) mission comprising Brooks Robinson (IMF East African Regional Technical Assistance Center (AFE) GFS Advisor), Ismael Ahamdanech Zarco (GFS expert), and Clément Ncuti (GFS expert), visited Kigali, Rwanda during January 18–29, 2016. The mission was conducted jointly with the IMF’s African Department (AFR) to ensure that fiscal and debt data development, compilation, and dissemination are aligned with AFR’s Policy Support Instrument (PSI) program monitoring needs. Senior Economist Tobias Roy was AFR’s mission representative.

The mission’s main objectives were to: (i) Assess progress on fulfilling the nation’s fiscal and debt statistics development (Government Finance Statistics Manual 2014 (GFSM 2014) implementation) plan; (ii) assist authorities in aligning the compilation and dissemination of GFS in accordance with the GFSM 2014; (iii) assist authorities in compiling general government finance statistics for financial years (FYs) 2013/14 and FY 2014/15 and develop related metadata documentation; (iv) reconcile externally-financed investment spending in Rwanda’s Integrated Financial Management and Information System (IFMS) with underlying financial flows, specifically project loans and grants; (v) assist authorities in developing plans to establish an integrated (Data and Metadata Exchange (DMX)) system for disseminating Rwanda’s fiscal and debt statistics; and (vi) assist authorities in planning to compile high-frequency data that are required by the IMF’s Statistics Department (STA) and AFR as it transitions from a GFSM 1986 to a GFSM 2014 reporting framework.

The mission is part of an AFE – East African Community (EAC) Secretariat collaboration (GFS capacity building) program that aims to assist EAC Partner States in achieving the fiscal data requirements associated with the East African Monetary Union (EAMU) Protocol.

A key conclusion of this mission is that the Ministry of Finance (MINECOFIN) should officially assign GFS compilation duties to a specific subset of its staff. Currently, a GFS Technical Working Group (TWG) that is comprised of planning, classification, and compilation committees of changing membership work on an ad hoc basis to consider GFS compilation and dissemination issues. The mission believes that it is important for the MINECOFIN to make a “permanent” assignment of GFS compilation and dissemination duties so that the work can evolve systematically. In the mission’s view, the Macroeconomic Policy Unit (MPU) appears to be well placed to perform this task. However, the unit should be allocated appropriate resources to take on the responsibility.

The mission acknowledges authorities’ nearly complete efforts to automate IFMS to produce annual and high-frequency GFSM 2014-compliant data for all of budgetary central government (BCG), most extrabudgetary units (EBUs), and all local governments (LGs) on a timely basis. However, efforts should continue to add remaining EBUs, social security funds (SSFs), and development projects to IFMS, and to finalize bridging from a GFSM 1986 to a GMFS 2014 framework.

Rwandan authorities are far along in the process of collecting, validating, and compiling GFS source data to meet the GFSM 2014 standard. At the same time, AFR continues to gather GFS from authorities that are consistent with the GFSM 1986 standard in order to monitor policy targets. The latter reverse-mapping exercise is becoming increasingly problematic, and it may prove to be advisable and efficient for AFR to consider updating its monitoring process so that fiscal statistics based on the GFSM 2014 standard can be used. Rwanda and the IMF are in the final phase of the current PSI program, and the mission believes that the quality of the data MINECOFIN can now produce, should allow AFR to transition to GFSM 2014-based statistics for program monitoring purposes at any moment, including in the near term. AFR’s representative on the mission team engaged with authorities during the mission to identify possible transition points.

While the mission team was impressed with the overall smoothness of compiling annual GFS, certain ad hoc interventions remain at the periphery of compilation processes. Absorption of these ad hoc intervention processes into core compilation procedures and implementation of the recommendations outlined below will help ensure that Rwanda satisfies fiscal data requirements associated with the EAMU Protocol and will help position Rwanda to meet fiscal Special Data Dissemination Standards (SDDS).

The mission’s main recommendations are:

  • Decide on a permanent assignment of GFS compilation and dissemination duties so that the work program can proceed effectively.

  • Complete automation of the GFS compilation process through IFMS by integrating source data from the Rwanda Revenue Authority (RRA), and by incorporating EBUs, SSFs, and development projects that are still outside of IFMS.

  • Work with the Statistics Department (STA) and AFR to implement an integrated DMX fiscal data dissemination system that will efficiently address future reporting requirements to multiple users.

  • Agree on a near-term transition plan for discontinuing reporting to AFR on a GFSM 1986 basis so that authorities can cease reverse data mapping.

  • Expand fiscal reporting to include stocks of financial assets and liabilities.

  • Finalize and begin implementing the GFS data quality improvement work program that was drafted in November 2015.

The mission warmly thanks the authorities for their excellent hospitality and support, which contributed greatly to the success of the mission.

II. Introduction

1. In response to a request from Rwandan authorities and in consultation with the IMF’s African Department (AFR), Brooks Robinson, IMF East African Technical Assistance Center (AFE) Government Finance Statistics (GFS) Advisor, Ismael Ahamdanech Zarco (GFS expert), and Clément Ncuti (GFS expert) conducted a technical assistance (TA) mission to Kigali during January 18–29, 2016. It was a joint mission with the IMF’s African Department (AFR) to ensure that fiscal and debt data development, compilation, and dissemination are aligned with AFR’s Policy Support Instrument (PSI) program monitoring needs. Tobias Roy (IMF Senior Economist) represented AFR during the mission. The work is part of an AFE – East Africa Community (EAC) Secretariat collaborative program that aims to assist EAC Partner States in fulfilling fiscal data requirements associated with the East African Monetary Union (EAMU) Protocol. Appendix I presents a list of officials met during the mission.

2. The overall objective of the mission was to review Rwanda’s compilation and dissemination of GFS and provide recommendations for improvement to align current practices with the Government Finance Statistics Manual 2014 (GFSM 2014) analytical framework. The specific tasks included:

  • Encourage authorities to decide on the permanent assignment of GFS compilation and dissemination duties so that the work program can proceed effectively.

  • Assess authorities’ progress in meeting requirements for the nation’s fiscal and debt statistics development (GFSM 2014 implementation) plan.

  • Assist authorities in the compilation and dissemination of general government finance statistics datasets for financial years (FYs) 2013/14 and FY 2014/15.

  • Assess authorities’ efforts to reconcile data on external financing for development projects in the form of loans and grants with actual project spending that is derived from the Integrated Financial Management and Information System (IFMS).

  • Seek agreement with authorities on plans for implementing an integrated Data and Metadata Exchange (DMX) and OpenData Platform (ODP) data dissemination system that will be more efficient than the current reporting system.

  • Facilitate an agreement between authorities and AFR for a near-term transition to GFSM 2014-based reporting for PSI program monitoring purposes so that reverse data mapping to a GFSM 1986 basis can be discontinued.

3. The remainder of the report reflects the following structure. Part III discusses the legal framework and the importance of assignment of duties for GFS compilation and dissemination. Part IV provides an assessment of authorities’ progress in meeting the GFSM 2014 implementation plan. Part V describes progress at compiling GFS for FY 2013/14. Part VI discusses the reconciliation of external resources data with IFMS reporting of development projects’ expenditure. Part VII reports on a new initiative to streamline Rwandan authorities’ reporting process. Part VIII discusses AFR’s reporting needs and options for a transition to GFSM 2014-based data. Part IX reports on Rwanda’s requirements for resources, training, and TA. Part X is the conclusion.

III. Legal Framework for Assignment of GFS Compilation Duties

4. At the national level, key legislation concerning the budget includes the Organic Law on State Finances and Property and a new Ministerial Order on Financial Regulations (N°001/16/10/TC of 26/01/2016). The Organic Law on State Finances and Property (Law No. 12/2013, issued on September 12, 2013) establishes principles and modalities for sound management of state finance and property and for the planning, execution, monitoring, and reporting of the budget. It assigns to the minister of finance the responsibility for carrying out the provisions of the law. The Ministerial Order provides further guidance on State finances. It prescribes the structure and functioning of public financial management, preparation and implementation of the government budgets, accounting and reporting of all financial transactions, and financial control in line with the service delivery objectives covered in the government plans and programs. The Order, too, assigns these responsibilities to the minister of finance.

5. At the regional level, the East Africa Monetary Union (EAMU) Protocol of November 2013, ratified by Law No. 24/2014, issued on August 5, 2014 provided for the establishment of EAC monetary union by 2024, and sets out provisions for harmonization and coordination of fiscal policies and laws concerning the production, analysis, and dissemination of statistical information.

6. Both the Organic Law on State Finances and Property and the Ministerial Order assign responsibilities for fiscal reporting, timelines, and frequency to the MINECOFIN, while the EAMU Protocol requires the use of a harmonized framework for fiscal reporting. These laws leave no doubt that MINECOFIN has responsibility for compiling and disseminating GFS. The outstanding issue is which department or office within MINECOFIN should be permanently assigned to this responsibility. This decision needs to be taken soon to maintain the momentum associated with the new compilation of improved GFS.

7. The mission believes that MINECOFIN’s Macroeconomic Policy Unit (MPU) is a well-qualified candidate for assuming the permanent responsibility for GFS compilation and dissemination duties. Assigning the GFS compilation function to the MPU would have two advantages: First, being embedded in all aspects of policy planning, the MPU is well-connected with all data-producing agencies (though it shares this advantage with the Accountant General Department (AGD)). Second, and perhaps more importantly, integrating the function in the MPU will leverage the use of GFS data for policy planning and assessment—by the MPU itself, but also by external counterpart agencies that are stakeholders in the policy formulation process and that liaise regularly with the MPU. If the

MPU is selected for this assignment, then appropriate resources should be provided to ensure that the unit enjoys success in conducting its new duties.

Recommendation

  • As a short-term priority, MINECOFIN should take a decision concerning which of its departments or offices should be permanently tasked with compiling and disseminating GFS.

IV. GFSM 2014 Implementation Plan

8. Rwanda established a GFS Technical Working Group (TWG) and adopted a fiscal and debt statistics development (GFSM 2014 implementation) plan in 2014. The TWG has functioned well since its establishment, and it has implemented many of the plan’s components. Still pending is the alignment of financial asset and liability items with the GFSM 2014 standard within Rwanda’s chart of accounts (CoA) that was approved for use in developing the nation’s 2015/16 budget—a very important achievement. Most importantly, as a result of the current mission, general government finance statistics for FY 2013/14 have been successfully compiled in accordance with GFSM 2014 guidelines.

9. However, other tasks are pending. For example, historical series should be translated into the new fiscal reporting framework; GFSM 2014-compliant public sector debt statistics should be compiled; and efforts to compile fiscal data in accordance with the Classifications of the Functions of Government (COFOG) should be completed.

10. Appendix II presents Rwanda’s implementation plan and reports on the status of tasks as of January 2016.

Recommendation

  • Authorities should accelerate efforts to complete the TWG’s roadmap.

V. Compilation of FY 2013/14 General Government Finance Statistics

11. One of the mission’s key tasks was to assist authorities in the compilation and dissemination of general government finance statistics datasets for FYs 2013/14 and FY 2014/15. The authorities pointed out, however, that the most benefit could be derived from the mission if the FY 2013/14 dataset was compiled to maximum perfection and the process documented well during the mission. Applying lessons learned during the exercise, authorities would then compile a general government finance statistics dataset for FY 2014/15, and submit the dataset for review to the mission team. In addition, authorities indicated that FY 2014/15 GFS source data were still being audited. The benefit in waiting for audited source data would be to avoid later revisions, even if it entailed a slight delay in GFS compilation.

12. The mission team provided guidance to authorities on compiling a general government finance statistics dataset for FY 2013/14 in a GFSM 2014 Statement of Operations framework.1

Table 1.

FY 2013/14 GFS Source Data by General Government Subsector and Major Economic Classification*

article image

--The full forms for the acronyms in this table are: AAFR-Audited Annual Financial Reports; AGD-Accountant General Department; DMFAS-Debt Management and Financial Analysis System; IFMS-Integrated Financial Management Information System; NBR-National Bank of Rwanda; GPMU-Ministry of Finance and Economic Planning’s (MINECOFIN’s) Government Portfolio Management Unit; RB-Rwanda’s Budget; RRA-Rwanda Revenue Authority; and TD-MINECOFIN’s Treasury Department. Source: Mission team.

Table 1 shows the sources of data that were used for the compilation by general government subsector and by major economic classification. The table indicates that for budgetary central government (BCG) tax Revenue data were from RRA (Rwanda Revenue Authority) and most nontax Revenue data were from the National Bank of Rwanda (NBR);2 Expense data were from the Debt Management and Financial Analysis System (DMFAS) and IFMS; Net acquisition of nonfinancial assets (NANFA) data were from IFMS and RB (Rwanda’s Budget); Net acquisition of financial assets (NAFA) data were from IFMS, NBR (Depository Corporations Survey (DCS)) and MINECOFIN’s Government Portfolio Management Unit (GPMU) and Treasury Department (TD); and Net incurrence of liabilities (NIL) data were from DMFAS. For extrabudgetary units (EBUs) source data for all economic classifications were derived in combination from audited annual financial reports (AAFR), the Accountant General Department (AGD), and from IFMS.3,4 For the two existing social security funds (SSFs), data for all economic classifications were derived from AAFR.5 IFMS was the data source for all economic classifications for Local Governments (LG). Appendix III reflects the institutional unit coverage by general government subsector.

13. It was necessary to adjust the source data to ensure conformance with the GFSM 2014 standard. The remaining subsections explain specific methodological procedures that were applied to transform the source data into GFSM 2014-compliant GFS.

A. Budgetary Central Government

14. For BCG, Revenue data had not been captured in IFMS. Monthly data on tax Revenue were compiled by RRA. A bridge table was developed by authorities, with assistance from an earlier IMF TA mission, to translate RRA monthly tax Revenue data into a GFSM 2014 framework. Current Grants data were obtained from NBR reports, while capital Grants data were based on Rwanda’s budget. All revenue data are on a cash basis.

15. All BCG Expense data were captured in IFMS and are available on a real-time basis. MINECOFIN records all BCG Expense data in IFMS, and can generate aggregated data on a monthly basis. Two datasets were generated from IFMS, based on time of recording. Data on an accrual-like basis (ordonnancement or payment order basis) were used for GFS reporting purposes.6 Cash basis data (payment) were used to compute a value for “float” (i.e., Other accounts payable). The “float” is the difference between total Expense on an ordonnancement basis and total Expense on a payment basis. Although data on Interest Expense were recorded in IFMS, Interest data from DMFAS were used as the source to compile GFS to ensure data accuracy.7

16. To compile Expense data on Grants from the BCG to other levels of government, a customized report was generated from IFMS by aggregating spending by EBUs and by LGs. The sum of spending by EBUs in IFMS’s “national mode” constituted Grants from the BCG to EBUs; similarly, the sum of spending by LGs in IFMS’s “national mode” constituted Grants from BCG to LGs.8

17. BCG data for domestically-financed NANFA were extracted from IFMS. For FY2013/14, IFMS did not include data on the NANFA financed by external Grants or Loans. To compile data for externally-financed NANFA, MINECOFIN summed disbursements of capital Grants (from RB) and capital Loans (from DMFAS). Data on disposals of nonfinancial assets were obtained from MINECOFIN’s Treasury Department (TD) and GPMU.

18. Data sources for NAFA were obtained in two parts. First, data on the acquisition of financial assets (Debt securities, Loans, and Equity and investment fund shares) for BCG were obtained from IFMS. Data on the acquisition of Currency and deposits were obtained from NBR’s DCS. Second, data on financial asset disposals were obtained from MINECOFIN’s GPMU.

19. Data on BCG’s NIL were also obtained in two parts. First, incurrence of liabilities data were obtained from DMFAS. Second, although data on repayment of liabilities are recorded in IFMS, DMFAS data were used for GFS reporting to ensure accuracy and timeliness.

B. Extrabudgetary Units

20. Table 1 shows that source data for EBUs for all major economic classifications were obtained from AAFR, AGD, or IFMS. As already noted, GFS for the EBU that operated outside of IFMS and did not report to the AGD (the Rwanda Utility and Regulatory Authority (RURA)) were compiled directly from the institutional unit’s AAFR. The AGD compiled GFS for those EBUs that operated outside of IFMS but that reported to the AGD.

21. For EBUs that report within IFMS, Rwandan authorities produced an EXCEL file from IFMS that presented tabulations by detailed economic classifications. Some of the IFMS nomenclature used for these classifications was inconsistent with the GFSM 2014 standard (e.g., Grants were classified as Other current transfers and vice-versa); therefore, the mission team recommended that authorities apply correct nomenclature. In addition, there were a few cases of misclassifications that the mission team provided guidance for correction; e.g., Disposal of nonfinancial assets was classified as Other revenue, NAFA data were classified under NANFA, and certain Loans were classified as Other accounts payable. Data for Revenue, Expense, and NANFA were presented on a flow basis, while data for the NAFA and the NIL were presented on a stock basis. Flow data were derived for the latter two classifications as the difference between opening and closing stock values. These adjustments brought the source data into conformity with the GFSM 2014 standard.

22. The GFS that were compiled from AAFR, AGD, and IFMS were aggregated to produce total GFS for EBUs.

C. Social Security Funds

23. AAFR were used to compile GFS for the two existing SSFs (RSSB (Rwanda Social Security Board) and MMI (Military Medical Insurance)). Generally, these financial statements were comprehensive and reflected high-quality data, which was evidenced by a relatively small Statistical discrepancy between Net lending/borrowing and Net financing. However, the SSFs AAFR are available only with a significant time lag. Reducing this time lag would help improve the timeliness with which GFS can be compiled. Also, the GFS TA mission team detected a grant that was extended by BCG to MMI that was not recorded as revenue in MMI’s financial report.9 While the mission team corrected this oversight, it is important for Rwandan authorities to communicate with MMI to ensure that this or similar grants be reported in future financial reports.

D. Local Governments

24. As explained in paragraph 12, and as seen in Table 1, the source data for all economic classifications for LG were derived from IFMS. In fact, the LG IFMS source data were comparable in form to EBUs’ IFMS source data. Therefore, the methodological procedures that were used to compile GFS for EBUs from IFMS source data were used to compile LG GFS using IFMS source data (see paragraph 21).

25. Appendix IV presents the results of the above-described compilation process, which is a consolidated GFS dataset for the general government sector for FY 2013/14.10 Considering the final column of the first page of the table, observe that Total revenue (line 1) is 1,577.7 billion FRw; Total expense (line 2) is 950.0 billion FRw; the Net operating surplus (line 3) is 627.8 billion FRw; the NANFA (line 4) is 706.4 billion FRw; and Net borrowing (line 5) is 78.6 billion FRw. Turning to the right-most column of the second page of the table, note that Net financing (line 3) is 120.2 billion FRw, which results from 14.1 billion FRw in NAFA (line 4) less 134.3 billion FRw in NIL (line5). Consequently, the Statistical discrepancy (line 2) is 41.6 billion FRw. The size of the Statistical discrepancy, which is 2.6 percent of Total revenue and 4.4 percent of Total expense, indicates that the dataset balances well and is of very good quality. The mission team concludes that, if the above-described compilation methodologies are used to compile a general government finance statistics dataset for FY 2014/15 (with required modifications), then authorities will produce a very good quality GFS dataset for dissemination.

Recommendations

  • Authorities should continue efforts to incorporate RRA data into IFMS

  • Authorities should prepare a Statement of Sources and Uses of Cash for FY 2013/14.

  • Authorities should continue efforts to incorporate into IFMS EBUs and SSFs that operate outside of IFMS currently.

VI. Reconciliation of External Resources and Project Expenditure

26. Creating timely and consistent records for development projects’ expenditure and their related external resource flows has always been difficult for authorities. These difficulties are quite common in countries receiving external development resources. Recording gaps and delays can translate into substantial discrepancies in the fiscal accounts, undermining their integrity. A separate difficulty arises with timing discrepancies between external disbursements and the project expenditure they are supposed to finance. This can create a significant problem for fiscal management.11

27. Recording errors leading to discrepancies can occur on both the expenditure and the financing level. First, during the budget planning phase donor-provided information on expected project disbursements does not always distinguish properly between Grants and Loans financing. Second, given time lags of audited project spending reports, it is difficult to capture actual spending execution accurately and assign it to the correct period. This has been a particular problem with independent projects that fall under specific line ministries, but are not integrated into the ministries’ reporting systems—in the past, spending by these projects has often been recorded “as budgeted,” which can deviate from actual spending.

Third, donor disbursements are frequently made as “direct payments” to foreign-based suppliers of goods and services, leaving no record in the domestic payment system. To correctly capture such direct payments, fiscal data compilers depend entirely on timely and accurate information provided by donors.12 Fourth, Grants in particular are occasionally channeled to projects through cash remittances outside the banking system. And fifth, Grants in kind are sometimes hard to capture and to reconcile with project expenditure plans due to valuation problems.

28. Over the past years, the authorities have successfully engaged with donors to gradually close these reporting gaps and minimize the potential for discrepancies. To enhance the accuracy of project spending plans for the budget, MINECOFIN reconciles information from datasets provided by three different sources (planning departments responsible for the Public Investment Plan; consolidated disbursement plans by donors; and spending projections by line ministries). Monitoring of project execution has been facilitated greatly by the creation of Single Project Implementation Units (SPIUs) that are embedded in line ministries: Starting with the FY 2015/16 budget, independent project spending will be reflected in IFMS based on monthly reports, which will result in more reliable yearly and quarterly information.13 However, any improvement in the compilation of project transactions will depend critically on the reliability and timeliness of the data reported by projects and donors to the SPIUs.

29. The centralization of project accounts in the NBR has optimized data coverage for all cash disbursements. Following a request by the government ten years ago, donors have shifted project accounts from commercial banks and other financial institutions to the Central Bank, which allows for timely and accurate tracking of disbursements. The remaining problem of recording direct payments to suppliers has been narrowed down to those cases where the disbursement is initiated by donors— disbursements requested by project coordinators would now be registered in the IFMS, provided the information is passed on correctly to the SPIUs. In addition, because the AGD receives actual spending data for self-accounting projects, including direct payments relating to technical assistance, intra-year and annual financial statements for self-accounting projects can be used to account for external financing.

30. Notwithstanding these significant improvements, some challenges remain. Despite repeated appeals by authorities, the quality and timeliness of information about direct payments initiated by donors still varies (and more so with Grants than with Loans disbursements). Moreover, under the GFSM 2014 presentation, the proper valuation of Grants in kind provides a challenge, as this is needed for the correct assessment of NANFA. This is particularly important in the case of technical assistance received, as these services are often provided directly by donor agencies and are difficult to measure in nominal monetary terms.

Recommendations

  • Authorities should continue extending the range of independent projects that are covered by IFMS.

  • Authorities should establish a Single Project Implementation Unit at the Ministry of Infrastructure.

  • Authorities should continue exploring new and innovative ways for incentivizing donors to provide timely information on direct payments to suppliers.

VII. Planning a DMX-Based Data Dissemination System

31. Currently, Rwandan authorities disseminate similar fiscal statistics separately to the IMF’s STA and AFR. In the future, the authorities will be required to disseminate to the EAC Secretariat also. As a new initiative, STA is considering an effort to develop an integrated dissemination system using a DMX- and ODP-based dissemination process, which would be designed as follows:

  • STA and AFR would develop a Microsoft EXCEL file that would be designed to include GFSM 2014 flow and stock data that meet STA, AFR, and EAC Secretariat requirements, plus certain idiosyncratic fiscal data required only by AFR. Data series in the file would be associated with GFSM 2014, AFR, and DMX codes.

  • Rwandan authorities would populate the EXCEL file using data that they compile.

  • Rwandan authorities would use the EXCEL file to populate an ODP database.

  • The ODP database would be made available to STA, AFR, and the EAC Secretariat; these data users could fulfill their specific fiscal data requirements using the ODP database.

32. The mission concluded that the best feasible strategy for designing the new integrated dissemination system would be to have STA develop a file with GFSM 2014 flow and stock series that meet STA’s and the EAC Secretariat’s needs. This file would later be augmented by AFR with idiosyncratic fiscal series that are required to meet AFR’s data reporting requirements. Coding experts at the IMF would incorporate DMX codes for each series in the file. Rwandan authorities should experience no difficulty in populating this file. However, the authorities may require refresher training in order to produce the aforementioned ODP database.

33. Mission team members developed a roadmap for producing the just-described EXCEL file, for conducting ODP refresher training, and for initiating data dissemination using the new integrated system (see Appendix V). Rwandan authorities agreed to conduct ongoing discussions about the roadmap with STA, AFR, and the EAC Secretariat.

Recommendation

  • Authorities should participate actively during consultations that are associated with the development of the new integrated DMX data dissemination system.

VIII. Meeting AFR’s Fiscal Statistical Requirements

34. The envisaged improvements of GFS under the GFSM 2014 methodology will go a long way to improve data reporting to AFR. The related increase in frequency, timeliness, and coverage of fiscal data reporting will enhance AFR’s core work, including economic surveillance, policy analysis, and program monitoring.

35. In terms of priorities, efforts should first focus on establishing a robust data reporting system for complete GFS of the BCG on a monthly basis.14 The delineation of institutional units in the government, which was developed and rigorously applied in the FY 2013/14 compilation exercise, provides sufficient granularity in the reporting system as to augment reliably BCG data on GFSM 2014 standards with the additional information needed to report on fiscal program target outcomes previously agreed between the IMF and Rwandan authorities.

36. Going forward, extending data coverage will strengthen economic policy assessment and facilitate cross-country analysis. Once BCG and central government reporting have been established, AFR’s economic surveillance insights would benefit greatly from standardized general government data, which will allow cross-country comparisons and aggregate regional analysis.15 Ideally, this extension of regular data provision at the general government level would be in place for the next Article IV consultation with Rwanda, which is expected to take place in October 2016. However, this is a very ambitious target.

IX. Resources, Training, and Technical Assistance

37. As agreed with authorities, a mission will return to Kigali during July 2016 to assess progress on implementing the recommendations presented in this report, and to provide general support for Rwanda’s implementation plan. The main objectives of the mission will be to:

  • Conduct a review of the FY 2014/15 general government finance statistics dataset.

  • Provide hands-on GFS compilation training to MINECOFIN staffers and to selected source data providers.

  • Assess efforts to finalize and begin implementing a GFS data quality improvement work program that was drafted in November 2015.

Continued TA will focus on the compilation of high frequency fiscal and debt statistics.

38. The mission team confirmed certain planned TA activities with authorities—both in Rwanda and at the regional level (see Table 2 below).

Table 2.

GFS Technical Assistance Activities During 2016

article image
Source: Mission team.

X. Conclusion

39. This mission resulted in the successful compilation of consolidated general government finance statistics for FY 2013/14. The documentation produced during this exercise (see Appendix VI) should facilitate the smooth compilation of general government finance statistics for FY 2014/15 in the near term, and the mission team awaits an opportunity to review the dataset. Rwandan authorities will also benefit from this short time series of consolidated general government finance statistics as these statistics may help inform future policy decisions.

40. Also, the mission was successful in producing GFS for BCG on a monthly basis for FY 2013/14. Unfortunately, neither monthly nor quarterly GFS for central government could be produced due to the lack of high-frequency data for all EBUs. Other important accomplishments during the mission were: (i) Development of a plan for an integrated DMX- and ODP-based data dissemination plan; (ii) discussions with authorities efforts required to meet AFR’s future data demands; (iii) completion a Microsoft PowerPoint presentation on the new FY 2013/14 consolidated general government dataset; (iv) completion of a Government Finance Statistics Yearbook questionnaire using the new consolidated general government finance statistics dataset for FY 2013/14; (v) provision of recommendations on the sectorization of eight new institutional units; and (vi) development of a “Synopsis of GFSM 2014” to sensitize authorities to the new GFS dataset.

41. If MINECOFIN’s leadership will now assign responsibility for the continued compilation of fiscal and debt statistics, then Rwanda can build on the successes experienced during this mission and begin to fulfill systematically its GFSM 2014 implementation plan. Importantly, the assigned staff can work to respond to the 11 recommendations set forth in this report and repeated in their entirety below. The mission team looks forward to returning to Kigali to continue this important work in July 2016.

Recommendations

  • As a short-term priority, MINECOFIN should take a decision concerning which of its departments or offices should be permanently tasked with compiling and disseminating GFS.

  • Authorities should accelerate efforts to complete the TWG’s roadmap.

  • Authorities should continue efforts to incorporate RRA data into IFMS.

  • Authorities should prepare a Statement of Sources and Uses of Cash for FY 2013/14.

  • Authorities should continue efforts to incorporate into IFMS EBUs and SSFs that operate outside of IFMS currently.

  • Authorities should continue extending the range of independent projects that are covered by IFMS.

  • Authorities should establish a Single Project Implementation Unit at the Ministry of Infrastructure.

  • Authorities should continue exploring new and innovative ways for incentivizing donors to provide timely information on direct payments to suppliers.

  • Authorities should participate actively during consultations that are associated with the development of the new integrated DMX data dissemination system.

  • Expand fiscal reporting to include stocks of financial assets and liabilities.

  • Finalize and begin implementing the GFS data quality improvement work program that was drafted in November 2015.

Appendix I.—List of Official Met During Mission

article image

Appendix II. Status Update of GFSM 2014 Implementation Plan

Project Objectives

article image

Project Outputs

article image
article image
article image

Appendix III. Institutional Structure of Rwanda’s General Government

Budgetary Central Government

article image
article image
article image

Extrabudgetary Units

article image
article image

Social Security Funds

article image

Local Governments

article image
article image

Appendix IV. FY 2013/14 Consolidated General Government Finance Statistics Dataset

article image
article image

Appendix V. Roadmap for Implementing DMX Integrated Data Dissemination System

article image

Acronym Key:

AFR – IMF’s African Department

EAC – East African Community Secretariat

DMX – Statistical Data and Metadata Exchange

GFSM – Government Finance Statistics Manual

ODP – OpenData Platform

STA – IMF’s Statistics Department

Appendix VI. Documentation of the GFS Compilation Process for FY 2013/14

1. Budgetary Central Government

a. Tax revenue

Data on tax revenue were obtained by rearranging raw Rwanda Revenue Authority (RRA) data on tax revenue collections into a GFSM 2014 framework by using a mapping schedule of the RRA classification into the GFSM 2014 classification. Data on tax revenue were compiled into the file: “2013_14_RRA_Domestic Revenue.”

b. Grants

Grants are broken down into two categories of grants received by the Budgetary Central Government (BCG): (1) Current grants and; (2) Capital grants. Current grants data were obtained from the National Bank of Rwanda (BNR)’s foreign exchange cash flow statement (PDT) while data on capital Grant is obtained from the budget (Finance law). Data on grants were compiled and aggregated into the file: “2013_14_Rwanda_BCG_GFS_Dataset_Final.”

c .Other revenue

There are two sources for other revenue data: RRA and BNR. RRA data were rearranged in a GFSM 2014 framework using a mapping schedule of the RRA classification to the GFSM 2014 classification. BNR data in the expenditure and receipts of the treasury’s account report under the line “non tax revenue/RNF” were aggregated into the GFSM 2014 other revenue category of “Sales of goods and Services” under Administrative fees. All other non tax revenue data in the BNR’s expenditure and receipts of the treasury’s account report was aggregated into the GFSM 2014 other revenue category “Miscellaneous and unidentified revenue”.

Receipts from reimbursements for Peace Keeping Operations (PKO) were recorded into the GFSM 2014 other revenue category “Sales of goods and Services” under Incidental sales by nonmarket establishment.

Data on other revenue were compiled into the file: “2013_14_Rwanda_BCG_GFS_Dataset_Final.”

d. Expenses

Data on expense for: (1) compensation of employees, (2) use of goods and services, (3) subsidies, (4) social benefits, and (4) other expense are obtained from IFMS. However, although data on interest payments is also recorded in the IFMS, data used for GFS compilation purposes was obtained from DMFAS.

In order to clearly distinguish data on grants from BCG to other levels of government respectively for the local government (LG) and extra-budgetary units (EBU); a customized report was generated from IFMS showing total spending of LG and EBU in the national mode of IFMS. The aggregate spending for LG and EBU in the national mode of IFMS was then reported under BCG grants to other levels of government to LG and EBU respectively.

Data on expense were compiled into the file: “2013_14_Rwanda_BCG_GFS_Dataset_Final.”

e. Net Acquisition of Non financial assets (NANFA)

Data on NANFA are broken down into NANFA domestically and externally financed. Data on NANFA domestically financed were obtained by identifying data separately for acquisition of nonfinancial assets and disposal of nonfinancial assets. Data on acquisition of non financial assets were obtained from IFMS while data on disposal of nonfinancial assets were obtained from the MINECOFIN’s Government Portfolio Management Unit (GPMU). Data on NANFA externally financed were derived by summing up capital grants disbursements (obtained from Rwanda’s budget) and the incurrence of capital loans (obtained from DMFAS).

Data on NANFA were compiled into the file:“2013_14_Rwanda_BCG_GFS_Dataset_Final.”

f. Net Acquisition of Financial Assets (NAFA)

NAFA is reported in two distinctive financial assets instruments, namely: (1) currency and deposits; and (2) loans. Data on currency and deposits were obtained by computing the change in the government’s deposits data from the Depository Corporation Survey (DCS) of the banking sector. Data on acquisition of financial assets in the form of loans were obtained from the IFMS, while data on disposal of financial assets were obtained from MINECOFIN’s GPMU.

Data on NAFA were compiled into the file: “2013_14_Rwanda_BCG_GFS_Dataset_Final”

g. Net Incurrence of Liabilities (NIL)

NIL distinguishes external and domestic incurrence and repayment of liabilities. Data on domestic NIL were reported in three distinct liabilities instruments, namely: Debt securities; loans; and other accounts payable.

  • Data on issuance and repayment of debt securities were obtained from DMFAS.

Data on loans were computed as the sum of the change in government claims on the government in the DCS, excluding debt securities held by banks, and the net issuance of nonbank loans data from DMFAS.

  • Data on other accounts payable were computed as the difference between total IFMS spending on an “Ordonnancement” basis and on a “Payment” basis.

Data on external NIL that distinguishes current and capital loans issuances and repayments, were obtained from DMFAS.

Data on NIL were compiled into the file: “2013_14_Rwanda_BCG_GFS_Dataset_Final.”

2. Local Government (LG)

All data on LG revenue, expense, and financing were obtained from IFMS and aggregated into a report produced by the Accountant’s General Department (AGD). To compile LG data into a GFSM 2014 framework, data from the AGD report on LG were rearranged in the following way:

  • On Revenue – data reported in the AGD report as taxes on payroll and workforce were reclassified into taxes on income, profits, and capital gains since LG (and for that matter no other government unit) collects taxes on payroll and workforce. Data reported in the AGD report as transfers from central government units and transfers from districts were recorded as grants, while transfers from independent projects were recorded as transfers. Grants received from local individual and organizations were recorded as transfers. Capital receipts were recorded as sales of non-financial assets while proceeds from borrowings were recorded as incurrence of liabilities.

  • On Expense – data reported in the AGD report as transfers to central government units and transfers to districts were recorded as grants, while transfers to independent development projects were recorded as transfers. Grants to local individuals and organizations were recorded as transfers.

  • On Financing – data reported in the AGD report on cash at bank, cash in hand, accounts receivables, and accounts payable were stocks data that had to be converted into flows by differencing opening and closing stock values.

Further to these rearrangements to comply with a GFSM 2014 presentation, LG data on Revenue were adjusted in the following way:

Tax revenue collected by RRA on behalf of LG amounting to FRw 6.8 billion that was initially recorded as grants from BCG to LG were recorded as taxes on income, profits, and capital gains for LG.

  • Grants from the Local Authority Development Agency (LODA) to LG amounting FRw 30.7 billion that was initially recorded as grants from BCG to LG were recorded as grants from international organization to LG.

Data on the local government were compiled into the file: “2013–14_Rwanda_LG&EBU_GFS_Dataset_Final.”

3. Extra Budgetary Units

For all data on EBUs in the IFMS, Revenue, Expense, and Financing are obtained from IFMS while data on EBUs not in the IMFS are aggregated together with data on EBUs not in the IFMS into a report produced by the Public Accounts Unit within the AGD. To compile EBU data into a GFSM 2014 framework, data from the AGD report on EBU were rearranged the same way they were rearranged for LG with the exception of taxes, since EBU do not collect any taxes.

Data on extrabudgetary units were compiled into the file: “2013–14_Rwanda_LG&EBU_GFS_Dataset_Final.”

4. Social Security Funds

There are two institutional units which integrate the subsector Social Security Funds: Rwanda Social Security Board (RSSB) and the Military Medical Insurance (MMI). To compile the accounts of both entities, the financial statements have been used.

Data on social security funds were compiled into the file: “2013–14_Rwanda_SSF_GFS_Dataset_Final.”

a. Rwanda Social Security Board (RSSB)

For the medical scheme, financial statements show inconsistencies that have not been explained. For revenue and expense, the statement of comprehensive income has been used and contains all data needed. In the case of the pension scheme, the contributions receivable are not divided between employer and employee contributions. However, as the law establishes that the employee contribution is 3% of the salary and employer contribution is 5%, the total has been split as 0.375 (3/8) being employee contributions and 0.625 (5/8) being employer contributions. In the case of the medical scheme, both employee and employer make the same contribution, so the total contributions received by the medical scheme have been split in two equal parts to have employer and employee contributions. Regarding NANFA, the statements of financial position for 2012/13 and 2013/14 have been used, calculating the net acquisition as the difference in the stock of non-financial assets at the end of both periods. While the total can be compiled from the tables, splitting of the different NANFA categories must be performed by reading notes from the financial statement.

For NAFA and NIL, the statements of financial position for 2012/13 and 2013/14 have been used in the same way as for NANFA. It is important in this case to bear in mind that two adjustments, with data coming from the statement of comprehensive income, have to be performed. As it is known, transactions in 2013/14 are equal to stocks at the end of 2013/14 minus stocks at the end of 2012/13minus other economic flows in 2013/14. There are two other economic flows, in particular profit gains and losses, to be taken into account:

  • Change in fair value of financial assets at fair value through profit or loss (33,780,659,546 RWF for the pension scheme in 2013/14; 1,265,594,703 RWF for the medical scheme).

  • Realized gain on disposal of assets: (4,194,492,697 RWF for the pension scheme and 9,409,282 RWF for the medical scheme).

b. Military Medical Insurance (MMI)

There is an important grant (in the form of a transfer of land valued at 3.7 billion FRw) given by BCG to MMI, which is not recorded as revenue of the MMI (in the financial statement, only the increase in nonfinancial assets is shown). The amount (nonfinancial assets transferred by the BCG to MMI) has to be recorded also as MMI grant revenue. It is worth mentioning that this grant is also not recorded as an expense in BCG’s accounts; an adjustment was added to BCG expense to recognize the grant. On the other hand, BCG’s estimates of NANFA are compiled from IFMS, which only accounts for cash transactions. Therefore, an adjustment was made to BCG’s NANFA to account for the disposal of the land to MMI.

For other items, the compilation of GFSM 2013/14 data from the financial statements is straightforward and follows the methodology indicated above for the RSSB.

Calendar year 2014 was used for the compilation of GFSM 2013/14 because MMI’s fiscal year is equal to the calendar year. Ideally, in order to have higher-quality statistics for FY 2013/14, a pseudo fiscal year could be created by averaging data from the 2013 and 2014 calendar year financial statements.

1

Although the source data for Revenue were provided on a cash (when received) basis by the Rwandan Revenue Authority (RRA), data for Expense, Net acquisition of nonfinancial and financial assets, and Net incurrence of liabilities were tabulated from IFMS on a “payment order” basis, which is an accrual-like accounting concept. This inconsistency in recording basis prohibited the compilation of a Statement of Sources and Uses of Cash for FY 2013/14. However, given sufficient time and the fact that IFMS can tabulate fiscal data on a payment (cash) basis, it is possible to produce a Statement of Sources and Uses of Cash.

2

Among nontax revenue, only interest and penalties on late payments of taxes are reported by the RRA.

3

Mission team members and the authorities collaborated to compile GFS from the AAFR.

4

Most EBUs are accounted for in IFMS. However, certain EBUs operate outside of IFMS, but report to the AGD. Finally, one EBU, the Rwanda Utility and Regulatory Authority(RURA), operates outside of IFMS, does not report to the AGD, and must be accounted for by compiling GFS from an AAFR.

5

SSFs are comprised of the Rwanda Social Security Board (RSSB) and Military Medical Insurance (MMI).

6

A payment order is usually issued and recorded in the IFMS at the moment the reporting BCG agency has availed itself of the goods or services provided. This data point represents the closest approximation to the accrual basis (defined as “change of economic ownership”) that the IFMS can provide. Payment is consistent with a cash basis of recording.

7

DMFAS is managed by MINECOFIN’s Debt Management Department.

8

IFMS is configured in two modes: A “national mode” and “local mode.” The “national mode” covers all transactions in the annual Finance Law, and reflects for EBU and LG spending that is financed by BCG. The “local mode” captures total spending.

9

Also, the BCG grant was not reflected in IFMS data.

10

Consolidation in this dataset is limited to Grants.

11

For example, Rwanda has experienced situations in which project grants were disbursed as planned, but execution of projects was temporarily halted due to unexpected external factors. This led to an unplanned accumulation of project deposits in the reporting period. In the following period, project execution was resumed at an accelerated pace to make up for the delays, leading to a sizeable drawdown of project deposits. Since no corresponding external grant inflow was recorded in the same period, project execution would unexpectedly increase the overall fiscal deficit even if it was completely financed by grants. While situations like this greatly complicate fiscal policy management, they are not, in a strict sense, a data reconciliation issue, as they can occur even if all flows and transactions are reported and recorded correctly.

12

Even when the disbursement information is provided fast, it may differ from project coordinators’ disbursement requests, requiring further time-consuming reconciliation.

13

An SPIU has yet to be established for the important Ministry of Infrastructure.

14

Initially, the reporting frequency could be bi-annual or quarterly, allowing for stabilization of monthly data through revisions before they are submitted. As in any statistical reporting system, it is understood that best- effort timely data submissions are provisional in nature and can (should) be amended on the power of additional new information.

15

The timeliness of annual and high-frequency reporting for the general government sector can be facilitated and assured through the use of existing data collection templates that were designed by MINECOFIN.

Same Series

Other IMF Content

Other Publishers

Cover IMF Staff Country Reports
Rwanda: Technical Assistance Report-Government Finance Statistics
Author: International Monetary Fund. Statistics Dept.
Volume/Issue:
Volume 2018: Issue 088
Publisher:
International Monetary Fund
ISBN:
9781484349021
ISSN:
1934-7685
Pages:
37
DOI:
https://doi.org/10.5089/9781484349021.002