III. Selected Expenditure Management and Fiscal Data Issues32
62. Malawi’s fiscal programs have been frequently derailed by expenditure overruns. These overruns have reflected policy decisions and lack of political commitment to the program, but they have also been a consequence of the weak capacity of expenditure management and control. The weak expenditure management inevitably impacts on the quality of the reported fiscal data. In recent years, the authorities have made improvements to the expenditure management systems, supported by technical assistance; however, progress has not been as fast as expected, with some systemic weaknesses remaining. This may reflect the limited ownership and few incentives to strengthen expenditure management.
63. An important recent initiative to strengthen expenditure management is the initiation of weekly cash management meetings between the Treasury, Accountant General’s Department (AG) and the Reserve Bank of Malawi (RBM). Senior management participate in these meetings and take decisions on funding levels and associated borrowing implications. However, decisions taken are constrained by the timeliness and quality of information on fiscal developments. This illustrates the impact of the data quality problems on the decision-making.
64. This chapter provides an overview of the main issues in expenditure management (Section B), and the progress made on tracking pro-poor expenditure (Section C). Moreover, it highlights a number of issues regarding the quality of the fiscal data (Section D).
B. Expenditure Control and Management
65. This sections focuses on the systems at the heart of budget execution: The Credit Ceiling Authority (CCA) system, the Commitment Control System (CCS), and, looking-forward, the planned implementation of the Integrated Financial Management Information System (IFMIS).
Credit Ceiling Authority System
66. The main instrument for budget execution is the CCA system, a zero-balance cash release system. It was introduced in May 2000 to eliminate the need for line ministries to maintain large deposit balances in commercial banks, and to give the Treasury tighter control over the execution of the budget. While the system has addressed these two objectives successfully, some weaknesses remain, preventing the CCA system from realizing its full potential in strengthening expenditure management and control.
67. Twice a month, the Treasury and the AG Department issue credit ceiling authorities and supplementary credit ceiling authorities for wages, other recurrent transactions (ORT) and development expenditure.33 The combined credit ceilings provide the maximum expenses the line ministries can incur, with any unused credit to be carried forward into the following month. The line ministries can issue checks drawn on commercial banks up to the total amount authorized under the credit ceilings. On a daily basis, the four commercial banks participating in the system are reimbursed by the RBM for checks that have been cashed. The commercial banks are remunerated for providing this service.
68. The CCA system was introduced to avoid the build-up of idle deposits by line ministries in commercial banks. Previously, the Treasury would release cash to line ministries that would be deposited in commercial bank accounts. The Treasury had little oversight over how much money was actually spent by line ministries—at least in the short run—and over time there was a large buildup in line ministries’ bank accounts.34 With the introduction of the CCA system, in principle all line ministries’ commercial bank accounts should have been closed and the available balances transferred back to the centralized government account, the so-called MG account no. 1 in the RBM. In practice, a number of line ministries maintain commercial bank accounts for project funds disbursed by donors and for Treasury funds, and some may even keep non-tax revenue outside the oversight of the Ministry of Finance.35 Nevertheless, commercial bank deposits have declined significantly—in April 2002 they amounted to 0.4 percent of GDP down from 1.7 percent of GDP in May 2000—although some government activities still remain outside the CCA system.
69. However, the CCA system is compromised by a lack of coordination. The operation of the CCA system is supported by a committee of officials from the Treasury, the AG Department, and the RBM meeting on a monthly basis. The committee has an important coordinating role to ensure the smooth functioning of the system. One problem identified earlier was the issuance of supplementary CCAs by the AG, which were not properly coordinated and captured in the system, although steps have now been taken to address this. However, some confusion remains regarding the exact ceilings issued, as revealed by the, at times inconsistent, information on this presented by the three agencies involved.
70. A weakness relates to the reimbursement by the RBM of commercial banks’ daily claims. The RBM has now begun to maintain holding accounts organized by economic type for each ministry to intermediate the transfer of funds from the line ministry accounts to reimburse the commercial banks. However, each ministry has typically more operating accounts than there are holding accounts (e.g., different operating accounts for ORT in each cost center versus one holding account in the RBM). When the RBM receives a claim from a commercial bank, the reimbursement against a particular operating account may therefore exceed the funding limit for this account without breaching the holding account limit in the RBM.36 This system could be strengthened by introducing corresponding holding accounts for all the operating accounts and enforce the spending limits for these.
71. Another problem relates to the carry-over of funding between fiscal years. The rules are clear: budgetary funds that have not been spent by the end of the fiscal year will be available for one month to settle checks issued toward the end of the fiscal year. Thereafter, any funding that has not been spent should revert to the MG account no. 1. However, when the CCA system was introduced in May 2000 the full balance outstanding by the end of the fiscal year was carried forward into the new fiscal year (2000/01) amounting to more than MK 1.5 billion. Line ministries tapped into this by reducing the outstanding balance by MK 1 billion during 2000/01 (see Figure III.1). Moreover, the instruction from the AG in May 2001 clarifying the treatment of unused credit between fiscal years was not effectively implemented.
Figure III.1.Malawi: Credit Ceiling Authority System, 2000/01-2001/02
Source: Malawian authorities.
72. Despite its weaknesses, the CCA system is an improvement of cash management and budget execution relative to the previous situation. However, the potential benefits from the system could be fully realized with little additional effort. To strengthen the control functions built into the system, it is essential that the RBM effectively enforce the funding limits associated with the line ministry accounts. Efforts should also be made to facilitate the timely compilation of data on reimbursements by economic and administrative classification. This is expected to substantially strengthen the quality of expenditure data.
Commitment Control System
73. The CCS was introduced to provide the authorities with a handle on commitments to prevent these from evolving into arrears. The emphasis in the CCS as implemented, however, is more on reporting commitments rather than on controlling their incurrence. The system has also been marred by implementation problems impacting on both the quality and the timeliness of reports. This may have led policymakers not to utilize the information on commitments and arrears generated by the system.37
74. The CCS is essentially a ministry-based commitment recording system. It requires line ministries to maintain a commitment register and ensure that commitments are not incurred in excess of the available funding (although this aspect of the system does not seem to be strongly enforced). The Treasury generates a monthly management report aggregating this information, including information on commitments, accounts payable and arrears.
75. The CCS reports have been compiled with some time lag and the quality of the data reported by line ministries suffers from shortcomings. It is not uncommon that the reported commitments for particular ministries are not consistent between monthly reports (i.e., stocks and flows are inconsistent). This reflects the weak recording of commitments at the level of line ministries, as well as a need to provide more training to line ministry accountants in maintaining these. However, many of the observed data problems could have been identified and corrected by the Treasury when compiling the monthly reports.
76. The CCS system will prevent the accumulation of arrears if commitments are approved only subject to availability of funds under the respective budget lines allocated in the CCA system. However, the quality and timeliness of the generated reports must be substantially improved, which will require a more comprehensive approach. Foremost is the need to train staff in line ministries to maintain proper recording and control of commitments and arrears. This may require the Treasury to introduce credible penalties for line ministries that fail to control commitments and submit returns, perhaps by making reporting compliance one of the criteria against which the performance of controlling officers is judged. But it will also require more vigilance on the part of the Treasury in monitoring the quality of the information produced by the system. Finally, if policymakers were actually seen to make use of information, this would provide incentives to line ministries for improvement.
Integrated Financial Management System
77. To improve the expenditure management situation, the government with the support of the World Bank in 1995/96 initiated work toward introducing an IFMIS. This is a comprehensive financial management system that will, when fully operational, allow a major strengthening of expenditure management. However, the implementation of IFMIS has been marred by delays. Recently, the UK Department of International Development and the World Bank have stepped in with additional funding and this appears to have given some new momentum. Although repeatedly postponed, IFMIS has been introduced in four pilot sites namely the Treasury, the AG, the Ministry of Agriculture, and the Ministry of Education, with a further roll-out planned to proceed shortly.
78. The World Bank in December 2001 pointed out several problems with the implementation at the pilot sites. First, the implementation had not been carried out in a systematic and satisfactory manner. Second, the software application would need adjustments to fully comply with the current business processes and accounting procedures, particularly relating to cash management (such as the issuance of CCAs, reimbursement of checks paid by commercial banks, and issuance of receipts for fees and charges). Third, there is a need for further refinement in budgeting and accounting procedures to be introduced concurrently with IFMIS. The AG is now working on addressing these issues, but is planning to proceed with the roll-out as originally planned.
79. As a long-term solution to strengthen expenditure management and control, a properly implemented IFMIS will be of great importance, and the roll-out should provide momentum to improve expenditure management. However, in the short term the need to address problems as they occur—by introducing focused solutions—is unavoidable. Ongoing attempts at improving the existing systems should therefore continue in parallel with the rollout of IFMIS. The successful implementation of IFMIS is also dependent on the initial data being correct, which requires proper reconciliation between banking and expenditure data as well as an up-to-date commitments register.
C. Pro-poor Expenditure Tracking
80. Given its well developed program-based budget presentation, Malawi is in a good position to monitor pro-poor expenditure. The authorities have already made commendable progress in developing and implementing a system to monitor pro-poor expenditure. A strong feature of the system is that it uses the existing reporting systems and classifications. However, the critical issue will be to garner political support to ensure the effective implementation. It is encouraging, therefore, that the pro-poor expenditure programs were identified in the 2002/03 budget and endorsed by parliament in line with the PRSP.38 A list of the programs included in the budget are shown in Appendix I.
81. Discussions with the authorities during the last year focused on the appropriate tracking of pro-poor expenditure under the HIPC Initiative.
First, the budget will identify particular programs and sub-programs as priority pro-poor expenditure. The identification of pro-poor expenditure reflects the PRSP prioritization, although only programs that are directly poverty alleviating will be included in this definition. To fully implement the PRSP further refinement of the program or sub-program budget classification may be required.
Second, actual expenditure in 2000/01 on the identified pro-poor programs will constitute the baseline for establishing additionality under the HIPC Initiative. The additionality requirement recognizes the fungibility of financing and is intended to ensure that HIPC debt relief does not replace pro-poor expenditure already being undertaken. The baseline is based on actual expenditure reports submitted by line ministries for 2000/01.
Third, the authorities have established a HIPC account as a subsidiary account to the MG account no. 1. The purpose of this account is only to record the inflow of HIPC debt relief; the funds will not be tied to any particular expenditure. Accounting for debt relief has proven difficult, partly reflecting the authorities’ very poor record keeping of loan terms and conditions, and the multiple channels through which creditors provide debt relief. A few creditors reimburse Malawi after the full debt service has been paid. To some extent, for recording purposes this is the most straight-forward form of debt relief, since it can be recorded as a grant when the reimbursement is received.39 However, most creditors outright forgive debt service payments (some up to 100 percent). This requires Malawi to properly record as debt relief the amount of debt service forgiven. The authorities have indicated that the establishment of the HIPC account as an accounting device has been helpful in terms of focusing attention on the need to strengthen the recording of debt relief.
Fourth, to reduce the risk of virement of pro-poor spending into nonpriority spending and to provide additional control when releasing funds, the Treasury is considering to amend the CCA system by introducing separate CCAs to release funding for pro-poor expenditure programs.40 The amendment may not fully prevent virement since, to reduce the administrative burden, the CCAs most likely will not be drawn on separate bank accounts.
Fifth, spending on pro-poor expenditure programs will be protected, and if the budget has to be cut—for example, in the case of revenue shortfalls—this will be done by reducing funding for nonpriority expenditure programs. Introducing separate CCAs for pro-poor expenditure programs will facilitate this by providing an effective instrument to the Treasury to protect pro-poor spending while reducing funding for non-priority expenditure.
Sixth, spending on pro-poor expenditure programs will be monitored on a monthly basis using the expenditure returns from line ministries. Given the substantial interest in this by civil society, information on expenditure will be published, including in newspapers and on the website of the Ministry of Finance. In addition, the authorities are planning to, with assistance from civil society, develop output based monitoring for selected core pro-poor expenditure programs during the fiscal year.
D. Data Quality Issues
82. The weak expenditure management is reflected in the poor quality of fiscal data, as evidenced by the substantial discrepancy between above and below the line data in the fiscal table for 2001/02. Despite some recent improvements, major weaknesses seriously limit the authorities’ ability to monitor fiscal policy. Focus of this section is on the reported expenditure and financing data as quality issues on tax and non-tax data are minor.41
83. The fiscal reports prepared by the Ministry of Finance reveal large discrepancies between expenditure and financing data. To some extent, this is inevitable and merely reflects time differences in the recording of cash and check transactions (the float). However, a large part of the statistical discrepancy relates to the lack of timely reconciliation between expenditure and banking data. This is supposed to be done at the line ministry level, but is carried out with some delay—and in some ministries not at all—so that the monthly fiscal reports do not reflect this reconciliation.
84. The AG’s finalizing of the annual accounts has in the past been delayed exceeding the statutory requirement. These delays slow down the preparation of audited accounts by the Auditor General. Certainly, the late submission of the audited accounts to Parliament makes less likely any action to rectify discrepancies identified in the audited accounts.
85. The need to strengthen the quality of expenditure data has long been recognized, but, with the need to transparently monitor pro-poor programs, has now gained renewed importance. There are currently two primary sources for expenditure data: (i) data based on funding released by the Treasury; and (ii) data based on expenditure returns submitted to the Treasury by line ministries.42 The expenditure data reported in the fiscal tables are based on funding released by the Treasury rather than actual spending reported by line ministries, reflecting the weak quality and poor timeliness of the expenditure returns. By relying on funding data rather than actual expenditure, one would expect some discrepancies between expenditure and financing data. However, this should reflect timing differences so that over-funding in one month would be followed by over-spending in subsequent months and vice versa. The best source of expenditure data would be from the reimbursements through the CCA system. Currently, these data are only provided at the aggregate level, but it is expected that the RBM shortly will begin reporting monthly reimbursement data with an appropriate economic and administrative classification.
86. While there are large differences between funding and spending for both 2000/01 and 2001/02, it is difficult to assess the degree to which this reflects data problems—from the weak quality and incomplete coverage of the expenditure returns—or real differences between funding and spending. However, combined with information from the CCA system for 2001/02, the data suggest that funding for ORT has exceeded actual spending through March 2001.43 This may be explained by line ministries “saving” some of their funding for later in the fiscal year.44 Another explanation for the discrepancy, of course, could be that the spending reports are not comprehensive.
87. Although more projects have been brought into the development budget in recent years, this has not been supported by systematically capturing either expenditure or disbursements against these projects. There are two main issues: First, the debt and aid unit in the Ministry of Finance reports only data on disbursement from multilateral donors, whereas almost no bilateral disbursements are captured. Second, expenditure on development projects in the fiscal table are simply derived as the total of project grant and loans disbursements. Particularly in cases where donors require pre-financing by reimbursement of costs, development expenditure could be double-counted.45 There are also likely to be timing differences between actual expenditure and disbursement of financing. An attempt should therefore be made, over time, to capture more fully actual spending also on externally funded projects.
88. The authorities are attempting to clear any new expenditure arrears within one quarter following verification by the Auditor General. Moreover, if a line ministry accumulates arrears, these should be settled within the original budget allocation for that ministry to remove any incentives to finance expenditure through arrears accumulation.
89. The authorities’ strategy has been only partially effective, and serious concerns remain regarding the quality of the arrears data and the delays experienced in clearing the verified arrears. First, while the Auditor General’s verification report should be prepared within 45 days of the end of the quarter, none of the reports has yet been completed in compliance with this requirement.46 Second, comparing the Auditor General’s reports with arrears information from the CCS show large differences in the reported data for individual ministries (although the total stock of arrears is converging, see Figure III.2).47 Third, a number of ministries appear not to be consistently included in all the reports (affecting both the CCS and the Auditor General’s reports), which makes the change in the arrears stock difficult to interpret.48 Fourth, the data in the Auditor General’s reports are notconsistent with separate data on government arrears to the Water Boards. The Auditor General has agreed to broaden the verification exercise to reconcile line ministry data on arrears with the information from utilities, initially focusing on the Water Boards.49
Figure III.2.Malawi: Stock of expenditure arrears, 2001
Source: Ministry of Finance and National Audit Office.
90. The authorities have not been able to clear new arrears in a timely manner. This, to some extent, reflects the delays experienced in the compilation of the verification reports, but also a lack of enforcement of the clearance mechanism. For example, there have been instances where funding released to line ministries for clearance of arrears has been diverted to other uses. Moreover, it is difficult to reconcile the information on arrears cleared directly by the Ministry of Finance with the data reported by line ministries. This may be because line ministries have not updated their arrears information correctly. A memorandum of understanding was signed in 2001 between the Auditor General and the Treasury to strengthen the information flows to reduce these inconsistencies.
91. Before data on arrears can be reported with confidence, the inconsistencies in the reports must be corrected. The impact on the reported flow of arrears (net change in the stock) is unclear. However, to ensure consistent data over time, it is important that any revisions to the stock data be fully reflected across all quarterly reports so as to prevent distortions in the flow between quarters.
92. The authorities’ approach to arrears has achieved its objectives to some extent. The issue has received more attention than in the past and this has led to improvements in the recording of arrears data and also reduced the scope through which expenditure is financed by incurring arrears. Nevertheless, a concerted effort to strengthen the data is required.
93. Performance of Malawi’s fiscal program is monitored using financing data, given that these are traditionally regarded as being more reliable than revenue and expenditure data. However, there are some concerns about the financing data, particularly the quality and consistency of external and domestic debt data, adding to the uncertainties regarding arrears developments. To ensure that policy decisions are based on correct information, this is an issue that will need to be addressed.
94. There have been inconsistencies in the data on external debt service (interest and amortization) as reported by the RBM and the Ministry of Finance. This could be due to incorrect treatment of HIPC debt relief but also to lax accounting procedures.50 However, even for actual cash payments the various data presentations have not been fully consistent. Data on project disbursements are not sufficiently comprehensive and actual disbursements are likely to be underestimated. The data on repayments of special financing (such as supplier credits) are poor and not reported transparently. Particularly information on disbursements (and on the receipt of the goods financed by supplier credits) is incomplete, and the information provided on the repayment schedule has been insufficient to make reliable projections of actual repayments during the fiscal year.
95. Some discrepancies between different sources of information on t-bills and bank advances have been resolved by the RBM. However, the government issuing papers to the RBM to reduce the ways and means advance may have led to some inconsistent recording in the data on RBM financing and non-RBM financing when the RBM sold those papers in the secondary market.
96. Expenditure management and control remain weak in Malawi. However, over the last couple of years, a number of improvements have been implemented. While some systemic problems remain, these appear to be fixable with relatively little effort. This also suggest that expenditure overruns in the past have been more a reflection of policy decisions rather than the failure of expenditure management systems per se. To ensure that the remaining shortfalls are corrected, strong ownership and commitment are crucial.
97. The CCA system is clearly an improvement in terms of budget execution and cash management and the remaining shortfalls in the system could easily be overcome. Foremost of these, is the need for the RBM to maintain separate holding accounts for each bank account for which CCAs are issued and enforce the spending limits. Even though progress has been made recently in terms of timeliness of the CCS reporting, more effort should be devoted to strengthen the quality of the commitments and arrears information both at line ministry levels and in the Ministry of Finance.
98. Commendable progress has been made in setting up a system for tracking pro-poor expenditure, using the existing reporting systems and budget classification. The key issue is to maintain strong and committed implementation; this is becoming even more important given the prominence civil society attaches to this issue. Moreover, there is a need to build up strong political support for the system. There is still room to improve the transparency, including by clearly identifying PPEs in the program-based budget documents.
99. The problems of the expenditure management systems resurface in the fiscal data. There are serious data quality issues affecting particularly the expenditure and financing data. Many of these can be addressed by strengthening cooperation at the technical level between the Treasury, the AG and the RBM (e.g., on the discrepancies in the debt service data). Expenditure data may be more difficult to correct in the short term since these are relying on accounting practices in the line ministries. However, the quality of the data can only be improved through more forceful supervision by the Treasury and the AG. The initial momentum for improving the data on expenditure arrears seems to have slowed down, and serious questions remain regarding the quality of the reported data.
Agricultural extension services
Targeted input program
Primary health care
Preventive health care
Secondary health care
Rural water supplies
Family welfare services
Adult literacy education services
Rural feeder roads
Small-scale fish farming
Technical and vocational training
Small- and medium-enterprise promotion
Prepared by Thomas Baunsgaard.
Supplementary credit authority ceilings mostly relate to foreign-financed development projects and, on a smaller scale, to expenditure through Treasury funds and other line ministry deposit accounts.
As cash releases continued to be made on a monthly basis without any regard to balances in the bank accounts.
Treasury funds are deposit accounts outside the central budget maintained by line ministries for certain activities.
This explains why commercial banks on several occasions have been reimbursed in excess of the credit ceilings.
This reduces the incentive and pressure on ministries to improve the quality of the information.
To increase transparency, the Treasury could consider to explicitly identify PPE programs in the programmatic budget document (vol. 4b).
Although uncertainty relates to the timing of when debt service is being reimbursed.
This is contingent on a technical assessment to ensure that the gains from a separate CCAs will justify any additional administrative costs this may entail.
However, the authorities are encouraged to carry out a monthly reconciliation exercise between the Malawi Revenue Authority revenue reports and tax deposits in the MG account no. 1.
The expenditure returns are based on the vote books (recording vouchers processed) and not the cash books (recording actual cash spent). The latter are submitted to the AG with a time lag.
It appears that funds have been diverted away from wages, development expenditure and normal ORT to finance unbudgeted special activities.
This does not rule out arrears; before the CCA system was introduced, large bank balances and arrears co-existed.
If the original spending has been captured in the credit ceiling for domestically funded development expenditure and the reimbursement by donors of this is subsequently captured in the disbursement data and reported as externally funded development expenditure.
For example, the arrears report for end-June 2001 was only received in mid-September, the report for end-September was delayed until January 2002, and the full end-December 2001 was only received in July 2002.
This could be due to the Auditor General having determined that these were related to ineligible payment claims; however, given the magnitude of the difference (and that the Auditor General has not explicitly reported finding large differences) it is possible that there are arrears that have not been captured in the Auditor General’s reports.
While it is possible that this reflects large swings in the stock of arrears (e.g., if all arrears owed by a ministry are fully repaid in one quarter, followed by an equivalent accumulation of arrears in the following quarter), this is not very plausible.
The capture of arrears between utilities and line ministries may have been obscured by past attempts by the Ministry of Finance to carry out offset operations that have not been properly recorded. The authorities have agreed not to conduct any further offset operations.
HIPC debt relief can either be reimbursed by donors after initial payment by Malawi or simply forgiven, thus not requiring an up-front cash payment (the majority of donors have chosen this treatment). In both instances, the authorities have agreed that the amount of debt relief received will be booked as a grant in the fiscal table and this will be offset by showing the external debt service on a due basis before debt relief (i.e., both the part of the debt service that is actually paid to the donors, and the part of the debt service that is forgiven).