Rwanda
Technical Assistance Report-Government Finance Statistics

This Technical Assistance Report discusses the findings and recommendations made by the IMF mission to assist Rwandan authorities in compiling annual government finance statistics (GFS) for the general government for FY2015/16 and high-frequency GFS for the budgetary central government and central government. It was observed that the annual GFS compilation is on track and even exceeding the expectations. The compilation of high frequency data should now be the focus of the authorities’ efforts. Quarterly Central Government data are expected to be compiled within 60 days for FY2017/18. Monthly BCG data are already compiled within 30 days from end of the period. In both cases, the Rwandan authorities have made some initial progress.

Abstract

This Technical Assistance Report discusses the findings and recommendations made by the IMF mission to assist Rwandan authorities in compiling annual government finance statistics (GFS) for the general government for FY2015/16 and high-frequency GFS for the budgetary central government and central government. It was observed that the annual GFS compilation is on track and even exceeding the expectations. The compilation of high frequency data should now be the focus of the authorities’ efforts. Quarterly Central Government data are expected to be compiled within 60 days for FY2017/18. Monthly BCG data are already compiled within 30 days from end of the period. In both cases, the Rwandan authorities have made some initial progress.

I. Executive Summary

An IMF AFRITAC East (AFE) government finance statistics (GFS) technical assistance (TA) mission carried out by Ismael Ahamdanech Zarco (IMF GFS expert), visited Kigali, Rwanda during February 6–17, 2017. Mike Seiferling (IMF GFS expert) joined the mission during the second week (training to Rwandese officials). During the first half of the mission, Yalenga Nyirenda participated, in the context of a mentoring program, to learn from the experiences of Rwanda on IFMS implementation and GFS compilation. The mission was part of an EAC Secretariat – AFE joint GFS Capacity Development Program, which seeks to align compilation and dissemination of GFS and public sector debt statistics (PSDS) with international standards.

The mission’s main objectives were to: (1) assist the authorities in compiling annual GFS for the general government for FY 2015/16 and high-frequency GFS for the budgetary central government (BCG) and central government (CG); (2) assist the authorities in extending quarterly PSDS up to the third quarter of 2016; and (3) provide a one-week GFS training program. All these objectives were met during the mission. The mission main achievements were:

  • Compilation of FY 2015/16 GFS.

  • Compilation of high frequency FY 2015/16 GFS.

  • Updating Central Government PSDS.

  • Analyzing the technical working group (TWG) composition, results-based management (RBM) plans and the statistical data and metadata exchange (SDMX) file for data transmission to the IMF’s African Department (AFR).

  • Providing one week training to Rwandese officials involved in GFS compilation.

Annual GFS compilation is on track and even exceeding the expectations (with a compilation period shorter than nine months). The compilation of high frequency data should now be the focus of the authorities’ efforts. Quarterly Central Government data are expected to be compiled within 60 days for FY 2017/18. Monthly BCG data are already compiled within 30 days from end of the period. In both cases, the Rwandan authorities have made some initial progress. However, the statistical discrepancies remain too high to consider such data of good quality. A correct recording of capital grants and net acquisition of nonfinancial assets (NANFA) is crucial to solve this problem and should be the main priority for the authorities going forward.

The mission’s main recommendations are:

  • Inform AFE and the EAC Secretariat of the composition of the updated TWG at their earliest convenience.

  • Allocate more human resources to GFS compilation.

  • Accelerate the tracking and recording in IFMS by Single Project Implementation Units (SPIUs) of disbursements and subsequent NANFA in the context of Capital Grants for each project.

  • Fine-tune the methodology used to compile data on extra-budgetary units (EBUs) and local governments (LG) GFS to reduce discrepancies. This could be done by sharing and analyzing the reports received by the Accountant General Department (AGD).

  • Audited annual financial reports (AAFR) both for the Rwanda Social Security Board (RSSB) and the Military Medical Insurance (MMI) should be transmitted in Excel to save time and to observe possible mistakes in the GFS compilation process.

  • Continue to explore new and innovative ways to encourage external donors to provide timely information on their direct payments to suppliers.

  • Finalize the streamlining of monthly data aggregation for EBUs.

  • Continue to work with UNCTAD so that accrued interest on T-Bills can be calculated.

  • Discuss the final RBM Plan with the TWG for approval.

  • Finalize the SDMX file, and in particular introduce the splits that AFR signaled as necessary for its analysis.

  • Transmit the SDMX file to AFR to ascertain that it meets its needs.

The mission team transmits its warm appreciation to the authorities for the hospitality and cooperation extended during the mission, which contributed significantly to the mission’s success.

II. Introduction

1. An IMF AFRITAC East (AFE) government finance statistics (GFS) technical assistance (TA) mission carried out by Ismael Ahamdanech Zarco (IMF GFS expert), visited Kigali, Rwanda during February 6–17, 2017. Mike Seiferling (IMF GFS expert) joined the mission during the second week (training to Rwandese officials). During the first half of the mission, Yalenga Nyirenda participated, in the context of a mentoring program, to learn from the experiences of Rwanda on IFMS implementation and GFS compilation. The mission was part of an EAC Secretariat – AFE joint GFS Capacity Development Program, which seeks to align compilation and dissemination of GFS and public sector debt statistics (PSDS) with international standards. Appendix I presents a list of officials met during the mission.

2. The mission´s main objective was to review Rwanda’s compilation of FY 2015/16 GFS data (both annual and high frequency) and provide recommendations for fine tuning further compilation and solve any problem found.

3. A July 2016 TA mission to Rwanda helped the authorities to finalize data for FY 2013/14 and 2014/15. The current mission worked on the FY 2015/16 data, which implies an improvement on timelines for annual data compilation of five months compared with the previous FY.

4. Besides the compilation of annual general government finance statistics, the mission’s main tasks were:

  • Assist the authorities in compiling high-frequency GFS.

  • Update the composition of the Technical Working Group (TWG).

  • Extend the availability of public sector debt statistics (PSDS) data to last quarter of 2016.

  • Finalize the Results Based Management (RBM) Plan for GFS, the Financial Balance Sheets and the PSDS (see appendix VII).

  • Review the SDMX template, which is used to provide data to the IMF AFR Department.

  • Provide a one week training to officials involved on GFS compilation (see Appendix VIII).

5. The remainder of this report is organized as follows. Section III describes the composition of the TWG. Section IV describes compilation of FY 2015/16 general government GFS. Section V covers the compilation of high-frequency GFS data for the budgetary central government (BCG) and central government for FY 2015/16. Section VI describes compilation of public sector debt statistics. Section VII covers the work on the RBM Plan. Section VIII describes the advances on the SDMX file. Section IX shows the main outcomes of the training course. Finally, section X contains the conclusion.

III. Composition of the Technical Working Group

6. The mission discussed with the authorities the composition of the updated Technical Working Group. The TWG is composed of staff of MINECOFIN, RRA, Rwanda Statistical Office and BNR.

7. While the composition of the TWG involves all units linked to GFS compilation, it has to be said that almost all compilation efforts fall under one person who is not exclusively dedicated to this task. In this context, allocating at least one more person (even if not full time) for the compilation effort would improve the general data quality and might avoid some errors that can be caused by the overloading the one compiler.

8. Finally, the authorities committed to send the composition of the updated TWG to AFE and the EAC Secretariat in the coming days.

Recommendations

  • Send the composition of the updated TWG to AFE and the EAC Secretariat at their earliest convenience.

  • Allocate more human resources to GFS compilation.

IV. Compilation of FY 2015/16 General Government Finance Statistics

9. One of the mission’s key tasks was to assist the authorities in the compilation of general government finance statistics for FY 2015/16. The compilation methodology was similar to that for FYs 2013/14 and 2014/15. Table 1 below shows the FY 2015/16 GFS source data by major economic classification for the general government subsectors.

Table 1.

FY 2015/16 GFS Source Data by General Government Subsector and Major Economic Classification*

article image
Source: Mission team.

--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; BNR-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.

10. The following subsections briefly describe adjustments made to the FY 2015/16 source data to line with the GFSM 2014.1 Appendix II shows the institutional coverage for the General Government subsectors.2

A. Budgetary Central Government

11. BCG data for FY 2015/16 is of high quality, with figures following the trends of previous FYs and with relatively low statistical discrepancy or billion FRw 17.2, which is 1.2 percent of total revenue. This discrepancy can be considered acceptable as the first step of the compilation process. However, reducing it should actually be one of the main goals of the authorities.

12. A remaining challenge in Rwanda GFS is the compilation of net acquisition of nonfinancial assets (NANFA). For the domestically financed NANFA, the source data were from IFMS, while the information on foreign financed NANFA was obtained as the sum of capital loans and capital grants received from foreign governments and international organizations. Owing to a lack of timely and accurate data on inflows of capital grants, the reported actual inflows of capital grants have always been assumed to be equal to the budgeted amount of capital grants (See Box 1). This lack of information in the budget execution on inflows of capital grants and on its matching with changes in deposits may be, in the mission opinion, the reason for the statistical discrepancies. Data on disposal of NANFA were obtained from MINECOFIN’s Treasury Department (TD) and Government Portfolio Management Unit (GPMU).3

Discrepancies Arising from the Mismatch between Expenditure and Financing for Foreign Capital Grants and Loans

Compilation of GFS for BCG has always been a challenge owing to a lack of accurate and timely data on external Capital Grants and capital Loans disbursements. While data on foreign financed capital Loans disbursements are generally available with a substantial lag (up to 6 months), data on foreign financed Capital Grants are usually not available.

Foreign financed Capital Grants and capital Loans disbursements come in two forms: (i) cash transfers, and (ii) direct payments. On one hand, cash transfers are cash disbursements into project accounts lodged at the BNR and are traceable in the government’s deposits in BNR’s DCS. Direct payments, on the other hand, are payments made by creditors directly to contractors for the delivery of services or acquisition of nonfinancial assets. These payments are usually made to foreign contractors, and information on the timing and amounts of the payments very much depends on the creditors’ willingness to disclose such information.

Because not all information on Capital Grants and capital Loans financing flows is directly available and timely, compilation of GFS for BCG assumes that actual disbursements are equal to budgeted amounts for Capital Grants and for capital Loans, and revisions are incorporated after actual information becomes available.

Given available information on financing flows, and to match Expenditure and Net Financing flows, Expenditure for foreign financed NANFA is derived as the sum of foreign financed Capital Grants and capital Loans disbursements. There can be, however, in a given fiscal year, either cash disbursements to projects relating to previous fiscal years or drawdowns on deposits from cash transfers of previous fiscal years. These cash disbursements (accumulation in deposits) or drawdowns on deposits relating to previous fiscal years are reflected in the current year’s changes of government’s Currency and deposits in BNR’s DCS, creating discrepancies arising from the timing mismatch between Expenditure and Net financing for foreign financed Capital Grants and capital Loans.

To correct this discrepancy, when known, changes in the current year’s (or month’s) government’s deposits in BNR’s DCS relating to previous fiscal years foreign Capital Grants and capital Loans transactions are subtracted (in the case of Currency and deposits accumulation) or added (in the case of Currency and deposits drawdowns) to foreign financed NANFA. This adjustment alleviates the problem but does not eliminate it, especially in the case of monthly data.

13. In order to ensure consistency of transactions on expense from IFMS recorded on an accrual-like basis and transactions on revenue, NAFA, and NANFA recorded on a cash basis, a float was recorded under Other accounts payable. This float is the difference between expense recorded on a cash basis in IFMS and expense recorded on “commitment basis” (accrual-like) in IFMS.

B. Extra Budgetary Units

14. Compilation of GFS data for EBUs uses IFMS, AGD reports and AAFR (for RURA) as data sources. The GFS that were compiled from AAFR, AGD, and IFMS were aggregated to produce total GFS for EBUs.

15. The statistical discrepancy for the EBUs (before consolidation adjustments) is FRw 918 billion without considering RURA or 0.9 percent of total revenue. While it is not too high, it is remarkable that for RURA (where the mission made the bridge from the AAFR to GFS) that the discrepancy was only FRw 39 million of 0.04 percent of total revenue. For all other EBUs, the bridge from reports to GFS was done by the AGD. The mission thinks that analyzing the EBUs reports could be of help to fine-tune the methodology applied to compile GFS for this sub-sector and subsequently might reduce the discrepancies. Due to a lack of time this task was not carried out.

C. Social Security Funds

16. AAFR were used to compile GFS for RSSB (Rwanda Social Security Board) and MMI (Military Medical Insurance).

17. Regarding RSSB’s AAFR, there have been important improvements as these reports have been available within 6 months from end of the period (the previous availability was 12 months). While these advanced reports include only preliminary figures, they show comprehensiveness and integration. As a result, the statistical discrepancy is low (less than 0.1 percent of total revenue). Moreover, RSSB AAFR includes now CBHI (Community Based Health Insurance) and increase therefore the coverage of the Social Security Sub-sector.

18. Another important improvement is related to the MMI data. Before, the AAFR referred only to the calendar year, which made comparability difficult, but now it is produced on a fiscal year basis, in line with all other GFS data source. The delay for this source has also been improved so it is now available at the same time that RSSB AAFR.

19. However, as the mission could verify and typing the figures of the AAFR from Word to Excel. This exercise is time consuming and may imply errors that affect the quality of final data. Having the AAFR in excel would help to economize time and reduce possible mistakes.

D. Local Governments

20. The IFMS is the data source for local governments. The mission discovered a small compilation issue that increased the statistical discrepancy, i.e., an issue of calculation of grants (an amount received from RRA was deducted). However, as this amount was not previously added as grants, the deduction was increasing the statistical discrepancy. Once this problem was fixed, the statistical discrepancy was reduced to FRw 7 million, which is less than 0.02 percent of total revenue of LGs.

21. Appendix IV presents the consolidated GFS dataset for general government for the FY 2015/16.4 The last column of the first page of the Table shows that total revenue (line 1) is FRw 1,845.1 billion; total expense (line 2) is FRw 1,259.9 billion; the net operating surplus (line 3) is FRw 585.2 billion; the NANFA (line 4) is FRw 725.1 billion; and net borrowing (line 5) is FRw 139.9 billion. Considering the last column of the second page of the Table, note that net financing (line 3) is FRw 157.5 billion, which results from FRw 193.3 billion in NAFA (line 4) less FRw 350.7 billion in NIL (line5). As a result, the statistical discrepancy (line 2) is FRw 17.6 billion, or less than 0.3 percent of GDP. The size of the statistical discrepancies does indeed indicate a very good quality of the GFS dataset that should give the authorities comfort for dissemination.

Recommendations

  • Accelerate the tracking and recording in IFMS by Single Project Implementation Units (SPIUs) of disbursements and subsequent NANFA in the context of capital grants for each project.

  • Fine-tune the methodology used to compile EBUs and LG GFS to reduce discrepancies. This could be done (by next mission) by sharing and analyzing the reports received by the AGD.

  • The AAFR both for RSSB and MMI should be transmitted in excel to economize time and reduce possible mistakes in the GFS compilation process.

V. High-Frequency General Government Finance Statistics for BCG for FY 2015/16

A. Monthly GFS for BCG

22. Monthly GFS for BCG were compiled for FY 2015/16 using the same methodology as for the compilation of annual GFS for BCG.

23. However, as can be seen in Appendix V, the results are far from being satisfactory in quality terms. The statistical discrepancies are too high for the data to comply with a minimum quality standard.

24. The problem is related with capital grants, capital loans and externally financed NANFA—this is the same problem faced in the annual compilation of GFS for BCG (See Box 1). This problem, partially attenuated in the case of annual data, is exacerbated for monthly data. The reason is that, while in the annual case the timing differences between the inflows of capital grants and capital loans and deposits drawdowns can be compensated between different months, but such compensation does not occur in monthly data compilation.

25. As stated in previous reports, this situation will improve once the SPIUs track all disbursements related to projects and the subsequent foreign financed NANFA and record both flows in IFMS. Although this recording will be performed only for disbursements from international organizations, it is expected that the discrepancy will be reduced significantly.5

B. Quarterly GFS for Central Government

26. The AGD is in the process of streamlining the aggregation and the consolidation of monthly financial statements of EBUs. The information will be available for GFS compilers within forty-five days from end of the period, so quarterly GFS for central government will be compiled within 60 days following the close of the quarter, and therefore meet requirements of international standard.

27. As for RURA, financial statements are only compiled on an annual basis. Therefore, until quarterly information is available, the economic activity of RURA may need to be estimated, and then introduced into the quarterly GFS for EBUs.

28. It is worth noting that the high statistical discrepancies of monthly data will also affect the quality of quarterly data. In this sense, without a correct tracking of capital grants and capital loans, it is expected that the quality of quarterly data will also be low due to high statistical discrepancies.

Recommendations

  • Continue exploring new and innovative ways for encourage external donors to provide timely information on their direct payments to suppliers.

  • Finalize the streamlining of monthly data aggregation/consolidation for EBUs.

VI. Extending Compilation of Central Government Public Sector Debt Statistics

29. The mission worked with the authorities to extend the compilation of quarterly debt statistics for BCG to last quarter of 2016. The debt data follow the classifications required by the World Bank – IMF Quarterly Public Sector Debt Database.

30. Debt data are compiled on a face value basis. While for loans and bonds accrual of interest can be calculated, this is not the case for T-Bills. This problem prevents the calculation of debt at nominal value. The authorities are in contact with the United Nations Conference on Trade and Development (UNCTAD) to solve this problem.

Recommendations

  • Continue working with UNCTAD so accrued interest on T-Bills can be calculated.

VII. Finalizing the RBM Plan

31. The Rwandan delegation developed a RBM Plan for GFS, Balance Sheets and PSDS during a workshop held in Zanzibar in July 2016.

32. The mission, in cooperation with the Rwandan authorities, has revised and finalized this RBM Plan. It has included some additional items as the extension of PSDS compilation coverage to General Government and it has adjusted the deadline for other items to make it more realistic. Finally, the mission has indicated the items already achieved. The updated RBM Plan can be found in Appendix VII.

Recommendations

  • Discuss the final RBM Plan with the TWG for approval.

VIII. Filling the SDMX File

33. In 2016, previous missions to Rwanda developed an Excel file with SDMX codes (see Appendix VII) that fulfills data needs of the African Department of the IMF (AFR). AFR itself participated in the design of such file.

34. The mission has filled in such table both for monthly and annual data. However, the data for some item´s split that AFR signaled as necessary have not been made available to the mission. According to the authorities, these splits will be introduced in the SDMX Excel file in the coming weeks.

35. Once this is done, the next step should be to submit the table to AFR so it can check if it meets the AFR’s needs. Finally, the SDMX codes should be introduced with the support of the technical division at the IMF.

Recommendations

  • Finalize the SDMX file, and specially introduce the splits that AFR signaled as necessary for its analysis.

  • Transmit the SDMX file to AFR for it to check if it meets AFR’s needs.

IX. Training Course

36. During the second week of the mission (13–17 February), a GFS training was delivered to fifteen Rwandan officials. The agenda of the training can be found in Appendix VIII.

37. The training combined theoretical lectures with practical examples using Rwanda´ data and compilations methods used. This combination of practice and theory was very well received by the participants, who scored the training with an average of 4.6 out of 5.

X. Conclusion

38. This mission resulted in the successful compilation of consolidated general government finance statistics for FY 2015/16 with a substantial improvement of timeliness. The GFS compilation methods used in the previous missions were used and fine-tuned. In addition, the mission compiled monthly GFS for BCG for FY 2015/16.

39. The mission also compiled debt statistics for Central Government for first three quarters of 2016, which implies a significant improvement in timeliness. Other important accomplishments during the mission included: (i) the review of the TWG; (ii) the finalization of the RBM Plan for GFS, Balance Sheet and PSDS compilation; (iii) the revision and pre-filling of the SDMX file to be transmitted to AFR and; (iv) the delivery of a one week training to Rwandans officials.

40. As closing point, the mission has confirmed the authorities’ engagement in the compilation, improvement and dissemination of GFS and PSDS. This is verified by aspects as the improvement of timeliness for annual GFS and PSDS and by the implementation (or the process of implementation) of most of the recommendations left in previous missions. However, the mission has also confirmed that the treatment of capital grants and capital loans and subsequent NANFA remains as an important issue that challenges quality of high-frequency data (and, to a lesser extent, also annual figures). The solution of this problem is crucial for obtaining high-quality data. Finally, the mission has also acknowledged that more human resources should be allocated to the compilation of GFS and PSDS. The mission team looks forward to returning to Kigali to continue this important work in May 2018.

Recommendations

  • Send the composition of the updated TWG to AFE and the EAC Secretariat at their earliest convenience.

  • Allocate more human resources to GFS compilation.

  • Accelerate the tracking and recording in IFMS by Single Project Implementation Units (SPIUs) of disbursements and subsequent NANFA in the context of Capital Grants for each project.

  • Fine-tune the methodology used to compile EBUs and LG GFS to reduce discrepancies. This could be done by sharing and analyzing the reports received by the AGD with the next mission.

  • The AAFR both for RSSB and MMI should be transmitted in Excel to economize time and possible mistakes in the GFS compilation process.

  • Continue exploring new and innovative ways for incentivizing external donors to provide timely information on direct payments to suppliers.

  • Finalize the streamlining of monthly data aggregation for EBUs.

  • Continue working with UNCTAD so accrued interest on T-Bills can be calculated.

  • Discuss the final RBM Plan inside the TWG for approval.

  • Finalize the filling of the SDMX file, specially introducing the splits that AFR signaled as necessary for their analysis.

  • Transmit the SDMX file to AFR for them to check that it meets their needs.