Technical Assistance Report-Strengthening Cash Management and Reviewing the Treasury Single Account

This Technical Assistance Report discusses recommendations for strengthening cash management and reviewing of the treasury single account in Uganda. The Cash Management Guidelines in Uganda have been approved. The coverage of the cash resource pool must be precisely defined and non-budget flows excluded. The cash flow forecasting model should continue to be developed along the lines given in previous Technical Assistance reports. The publication and dissemination of the Guidelines will provide authority to the Cash Policy Department to implement a program of training and capacity building within the central government entities in coordination with other directorates of the Ministry of Finance, Planning and Economic Development. This should be started as soon as the Guidelines are released.


This Technical Assistance Report discusses recommendations for strengthening cash management and reviewing of the treasury single account in Uganda. The Cash Management Guidelines in Uganda have been approved. The coverage of the cash resource pool must be precisely defined and non-budget flows excluded. The cash flow forecasting model should continue to be developed along the lines given in previous Technical Assistance reports. The publication and dissemination of the Guidelines will provide authority to the Cash Policy Department to implement a program of training and capacity building within the central government entities in coordination with other directorates of the Ministry of Finance, Planning and Economic Development. This should be started as soon as the Guidelines are released.

I. Introduction

1. At the request from the Ministry of Finance, Planning and Economic Development (MoFPED), East AFRITAC (AFE) fielded a technical assistance (TA) mission to Kampala during November 21 to December 2, 2016. The mission comprised Amitabh Tripathi (AFE public financial management advisor), and John Gardner (expert). It provided practical assistance to the MoFPED on strengthening cash management and implementing the Treasury Single Account (TSA).

2. Box 1 summarizes the mission’s findings and recommendations, whilst the remainder of this aide memoire provides more detailed analysis.

Uganda: Summary of Findings and Recommendations

The Cash Management Guidelines have been approved. These are consistent with the advice provided by the previous AFE mission. Together with the establishment of the high-level Cash Management Committee (CMC), the guidelines should provide an appropriate basis for formalizing the cash management function. This will also require addressing the uncertainty around the role and institutional arrangements for the new Directorate of Debt and Cash Policy (DDCP).

Cash flow forecasting is still at an early stage. The coverage of the cash resource pool must be precisely defined and non-budget flows excluded. The cash flow forecasting model should continue to be developed along the lines given in previous TA reports. The publication and dissemination of the Guidelines will provide authority to the Cash Policy Department (CPD) to implement a program of training and capacity building within the central government entities in coordination with other directorates of the MoFPED. This should be started as soon as the Guidelines are released.

The Cash Management Committee, as described in the Guidelines, should be formally established. A reporting structure should be designed by the CPD to assist with its regular report to the CMC. The CMC will be a vital element in leading the change management for making the transition from cash rationing to active cash management across government. It will provide authority to the CPD and ensure that, once cash flow forecasting is sufficiently accurate, active cash management operations can supersede the imposition of regular cash limits on expenditures. Budget execution control should then operate on expenditure commitments rather than payments.

There has been reasonable progress with Treasury Single Account implementation. Local governments are being included within the TSA but despite efforts there has been no progress with inclusion of donor funds.

Based on a review of the progress, the following key recommendations are made:

  • Issue and start implementing the Cash Management Guidelines.

  • Continue to implement the Action Plan for strengthening cash management (Appendix 1).

  • Operationalize the new DDCP as approved by the Cabinet.

  • Define precisely the cash resources pool used for cash flow forecasting and continue developing the cash flow forecasting model.

  • Design and implement a training and capacity building program by CPD for the MDAs

  • Design the CPD reporting structure to the CMC and commence formal CMC meetings.

  • Continue to bring local governments within the TSA coverage; and rationalize the Uganda Consolidated Fund/TSA bank account structure with extension of integrated financial management system (IFMS).

  • Continue to explore with receptive donors the options for piloting donor funds in the TSA.

  • Finalize the agreement between MoFPED and Bank of Uganda on banking arrangements.

A. Background

3. The mission was one of a series planned to assist the MoFPED in improving the management of its cash resources. Good progress has been recorded on TSA implementation over the last two years. The Accountant General’s Office (AGO) has confirmed its core TSA arrangements comprising of a separate Uganda Consolidated Fund (UCF), TSA holding accounts and TSA sub-accounts for accounting entities. The sub-accounts are funded when payments are ready to be made and unspent balances are swept back to the TSA holding account on a daily basis. The TSA arrangements are being progressively extended to local government but the pilot for inclusion of donor funds within the TSA has not progressed as planned.

4. The mission follows-on from the most recent AFE mission on cash management. That mission in July 2016 made recommendations and provided guidance on: (i) finalizing the cash management guidelines; (ii) improving cash flow forecasting by analyzing forecast errors and ensuring coverage of all significant cash flows consistent with IFMS scope; (iii) coordination with Debt Policy and Issuance Department on borrowing needs; and (iv) finalizing and implementing the action plan for strengthening cash management.1

5. Some progress has been recorded since July 2016. In particular, the Permanent Secretary/Secretary to the Treasury (PS-ST) has approved the Guidelines for Cash Management in Government. The Cash Policy Department (CPD) has identified the sources of forecast and actual data and have been using the previously developed cash flow forecasting templates to monitor actual revenue, expenditure and financing. They have been monitoring the cash balances and movement in the UCF, TSA holding and TSA sub-account and prepare a summary quarterly report to get an understanding of the consolidated cash position of government. The MoFPED considers it an opportune time to assess progress, identify any emerging issues and review its plans. Accordingly, it has requested AFE to provide guidance in these areas.

B. Mission Methodology

6. The scope of the mission was confirmed during an initial meeting. It was agreed that the mission would review the approved guidelines for cash management and the plans for developing the cash management function; advise on the processes and templates for constructing and managing a comprehensive cash plan; provide practical training to the staff of the CPD; and review the implementation of the TSA and plans for its extension. The mission assessed the progress and plans and identified emerging issues through document review and discussions with different stakeholders.

7. The mission acknowledges the cooperation and assistance provided by and the discussions held with the senior officials of MoFPED. The mission benefitted from its meetings with Keith Muhakanizi, Permanent Secretary/Secretary to the Treasury; Patrick Ocailap, Deputy Secretary to the Treasury; Maris Wanyera, Acting Director, Debt and Cash Policy; Lawrence Semakula, Accountant General; Kenneth Mugambe, Director Budget; and members of staff from these Directorates. It also held useful discussions with the Deputy Directors and other officials from the Banking Operations and Financial Markets Departments of the BOU. The information shared during these meetings enhanced the mission’s understanding of the progress achieved and the outstanding issues and thus informed its conclusions. The findings and recommendations in the report were validated during extensive discussions with the officials of the Debt and Cash Policy Directorate and presented in a wrap-up workshop. The mission is grateful to Robert Bellarmine Okudi, Acting Commissioner CPD, for coordinating its work.

C. Mission Outputs

8. The output from the mission comprised: practical guidance on development of the cash management function; discussion and training on cash forecasting templates; review of the TSA implementation and coverage; and a wrap-up workshop with senior officials of the MoFPED; and this aide memoire, which following this introductory section is structured as follows:

  • Section II – reviews the guidelines and institutional arrangements for cash management and makes recommendations for strengthening them.

  • Section III – reviews the cash forecasting model and suggests measures for improving cash forecasting and management.

  • Section IV – reviews the implementation of treasury single account arrangements and makes recommendations for extending its coverage.

  • Appendix 1 – Action Plan – Strengthening Cash Management.

II. Strengthening the Regulatory Framework and Institutional Aspects

9. The “Guidelines for Cash Management in Government” have been formally approved by PS-ST. As per the PFM Act (2015), the PS-ST is responsible for prescribing the framework for cash management in government. Considering that the PFM Act preceded the establishment of the Debt and Cash Policy Directorate and in the absence of adequate operational guidance in the PFM Regulations (2016), the approval of these guidelines is a significant step.

10. The Guidelines provide a framework for the operations of CPD by formalizing the cash management mandate and include the following:

  • Specification of the institutional restructuring for cash management as approved by the Cabinet;

  • Overview of the concept of government cash management and its primary objectives;

  • Government cash flow forecasting;

  • Scope of short-term borrowing and investing operations to meet anticipated temporary cash shortages and surpluses;

  • Institutional roles and responsibilities of MoFPED directorates and other stakeholders;

  • Role and composition of the CMC and its relationship with the CPD and other counterparties.

11. The mission reviewed the approved guidelines. Whilst they are broadly consistent with the draft prepared with assistance from the AFE mission in July 2016, some changes were suggested to clarify and streamline the provisions. These included: exclusion of public enterprises from the coverage of cash management; clarifying the role of the CMC; monitoring and reviewing the current process of issuing quarterly cash limits in view of improvements in cash flow forecasting; and streamlining the references to the oversight role of the Accountant General as regards to government banking arrangements.

12. The establishment of the institutional structure for cash and debt management has not progressed as anticipated. The Cabinet approved the structure and establishment of the Directorate of Debt and Cash Policy (DDCP) early this year. The approved positions were to be filled in a phased manner and an initial provision for this was made in the budget for FY 2016/17. The approval also required the Ministry of Public Service to accord priority to these recruitments. These new positions in the Directorate however, have not been filled.

13. Within the MoFPED, there also appears to be some rethinking on the role and independence of the DDCP. A unified Directorate underlines the need for coordination between the debt and cash management functions and the MoFPED intention of moving away from a cash rationing system. It aims in principle to integrate the previous fragmented debt management practices handled by the different departments of MoFPED and BoU. It brings together the front, middle, and back office functions of debt management in the new department.2 However, the mission was informed that some stakeholders in MoFPED have questioned the need for a separate Directorate and argued for moving the cash management, and debt middle and back office functions under the Accountant General.

14. The PFM legal framework and the approved guidelines, however, provide a basis for operationalizing the new institutional structure. The PS-ST, amongst other things, has overall responsibility for the cash and debt management function. The approved guidelines provide for a high-level CMC that is headed by PS-ST and will decide on cash prioritization and allocation. It shall advise the Minister on the processes for optimum management of cash including the modalities for meeting short-term mismatches between cash inflows and outflows. The CMC should have the necessary authority to address the uncertainty around the establishment of the DDCP and carry the reforms forward. It is important that this institutional issue is clarified as soon as possible in line with the Cabinet approval and as envisaged in the guidelines.

15. The establishment of a unified operational cash and debt management institution meets international recommended good practice. A fully integrated debt management function with front, middle, and back offices under a single umbrella allows for the efficient development and implementation of a comprehensive medium-term debt management strategy. A fully inclusive debt database is required to be complete and updated in order, not only for efficient settlement procedures, but also for ongoing analysis, review, and reporting of the strategy and its implementation. Separating the database operation from the non-settlement functions of a public debt office risks inefficient updating of front and back office activities, uncoordinated analysis and implementation of debt strategies, and difficulties in ongoing market development. A small part of the public debt management function relates to payment instructions for debt servicing operations. These can readily be performed directly using instructions from the debt database to inform IFMIS payments in the same way that all other central government payments are effected. Also, effective debt reporting is more comprehensively performed through the debt database than the IFMIS which is designed as an accounting platform rather than a debt management tool.

16. The cash management function in many developing countries starts out as a division of the Accountant General Office. Once it is fully established and operates effectively by ensuring, through active cash management transactions, that cash resources are forecast accurately and made available to pay committed expenditures, it is good practice to move the cash management function to the debt management office. This allows considerable efficiencies to be gained through the daily coordination of activities between cash and debt management. The initial formation of an integrated cash and debt management directorate has been seen as a progressive step forward for the government of Uganda and a leading reform in the region. Pulling the institution apart at such an early stage would be considered to be a significantly retrograde move.

17. It is important that the established positions are filled soon in order for the DDCP to become functional immediately following publication of the Guidelines. Even though all positions are not required to be filled immediately, many of the existing positions have been staffed by officials seconded from the AGO and they are currently being called back there. There is some confusion regarding the skills required of officials in both cash and debt management functions. Many of the positions are regarded as requiring accountancy qualifications, while most of the skills necessary for good government cash and debt management relate to relationship management, data handling and forecasting, and financial markets. Clarity on the role of the DDCP and the cash management roles will reinforce its authority and assist in the implementation of the guidelines that require substantial capacity building amongst relevant stakeholders including the MDAs as described below.


  • Issue the Cash Management Guidelines as soon as possible and organize sensitization workshops for MoFPED officials and MDAs (Jan-April 2017).

  • Establish the high-level CMC as provided in the guidelines (February 2017).

  • Operationalize the new Directorate of Debt and Cash Policy as approved by Cabinet (February 2017).

  • Continue to implement the Action Plan (Appendix 1) for strengthening cash management (ongoing).

III. Cash Flow Forecasting and Cash Management

18. The imminent publication and dissemination of the cash management guidelines is a very important step forward in commencing the cash management function. The Guidelines will provide the CPD with the authority and institutional standing to enable it to perform its cash flow forecasting function through building necessary relationships with stakeholders and counterparties within the MoFPED and across government. Once forecasting of government cash resources is sufficiently improved, relationships with the CMC, the debt managers, and the BoU will provide the CPD with capability to perform money market transactions aimed at smoothing available government cash balances alleviating mismatches in revenues and expenditures.

19. Previous TA mission reports, including the AFE July 2016 mission, have described in detail the steps required to commence cash flow forecasting. The action plan to achieve this has started through the approval of the Guidelines and the remainder of this plan requires implementation as soon as possible (Appendix I). As discussed in earlier reports, it is vital to start this process since it is inevitable that in the existing circumstances of strict cash rationing, the initial forecasts will be highly inaccurate. However, once analysis of the forecast errors and capacity building at the MDAs commences, experience shows that forecast accuracy can improve very quickly. By definition, until this process starts, no progress can be made and resistance to the reforms can grow.

Definition and coverage of the cash resource “pool”

20. The immediate priority for the CPD is to define precisely the cash resources which it plans to forecast over the fiscal year. At present, the CPD has chosen a selection of bank accounts which roughly comprise the existing TSA structure of accounts. These include the core accounts—the Uganda Consolidated Fund account (UCF); the TSA Holding account; and the LG TSA Holding account—and also include several other autonomous and semi-autonomous agency accounts many of which are connected to LG payments systems.

21. At this early stage, it is very important that the principle of simplicity is used throughout the forecasting function. This applies also to the coverage of the forecasts. The available cash resource pool should be defined as succinctly as possible in order to cover all budget cash movements but without necessarily including every detailed transaction. Many government entities receive direct transfers from the budget and these transfers can be used in forecasting the outflows instead of forecasting the detailed expenditures made by the entities themselves. This simplification allows forecasting and its subsequent error analysis to be manageable without loss of budget comprehensiveness.

22. It is recommended that the initial cash pool be defined as the UCF plus the TSA Holding and LG TSA Holding accounts. The LG TSA Holding account operates like the central TSA Holding account for LGs which are incorporated into the IFMIS ledger—see Section IV below. All other agency bank accounts should be excluded from the pool definition and the forecasting model at this early stage. Other significant bank accounts of central and LG agencies should be monitored since it will be necessary to aggregate these into the cash forecasts if they build up large surplus balances. At a more advanced stage it may be necessary and important to include in the pool definition those accounts which maintain large balances like the donor accounts and accounts of the EBU funds in order to analyze their operations more closely.

23. Figure 1 shows a schematic outline of a simple model and a more complex model for defining the government cash pool. It is important to understand that both models include all budget cash movements. All budget revenues and deficit financings flow into the pool and all expenditures and transfers flow out. The difference between them is that the simple model does not incorporate other bank accounts. Therefore, the only flows associated with these accounts that are relevant to the forecasting are transfers from the pool to the external accounts.

Figure 1.
Figure 1.

Uganda: Alternative Models for Government Cash Pool Definition

Citation: IMF Staff Country Reports 2017, 272; 10.5089/9781484317822.002.A001

24. In the more complex model, however, the other bank accounts are included in the defined pool. As such, their individual outflows (or expenditures) must be forecast in detail in order to ascertain the projected size of the pool balance. In this case, the transfers to these accounts are internal to the pool and, as with all internal cash flows, do not require forecasting since they do not affect the overall balance of the pool.

25. Defining the precise components of the government cash pool ensures that all budget cash flows are included in the forecasting of the pool balance across the fiscal year. Starting simply, as described, reduces the amount of work necessary to forecast individual bank account transactions but the pool can be expanded to include other bank accounts as necessary. However, a precise definition of the cash resource pool of the government is necessary elsewhere also. Current misunderstandings caused by a lack of sufficient awareness of the coverage of the government cash resources has recently led to difficulties and confusion. This is discussed in greater detail in Box 2 below.

Uganda: A Discussion on the Government Cash Position

There have recently been discussions between the IMF, the BoU, and the MoFPED (including the AGO) relating to the calculation and movements of the “Government Position” (GP). This statistic is meant to represent the total cash resources of the government at a particular point in time. However, there is confusion regarding the coverage of the calculation and the meaning of movements in the overall balance.

Briefly, the discussion has centered on the meaning of an annual drawdown in the GP— i.e., where the balance at the end of the fiscal year is lower than the balance at the start of the year. Confusion has arisen regarding the meaning of such a drawdown and its relationship to the execution of the approved budget—both from a fiscal and a monetary perspective. This confusion has been described as one between the economist viewpoint and that of the accountant. But, in reality, it is due to the misunderstanding of the precise definition of the GP.

When considering the GP as a measure of the cash resources available to the government, the AGO takes the most comprehensive view by assuming that every bank account should be aggregated and that all flows should be considered as government flows. In that case, a drawdown represents a net outflow over the fiscal year from all funds/accounts whether or not they comprise budget flows. The aggregation within this definition of the GP not only comprises the UCF and the TSA Holding accounts but also EBFs, non-IFMIS LG accounts, donor project accounts, LC deposits, and special funds. During the year, this measure falls but normally remains an overall positive balance. The AGO therefore argues that the drawdown does not represent an “overdraft” or “advance” and that it does not need to be financed separately. In this regard, the AGO is considering the GP to be a pool similar to that defined as the “complex model” in Figure 1.

The IMF, however, under the PSI program conditions, looks at a narrower definition of cash movements. Its pool definition “excludes deposits in administered accounts (including the energy and petroleum funds), project accounts with the central bank, and net recapitalization securities” and therefore concentrates on budget execution more exclusively. In this regard, the IMF is considering the GP to be a pool similar to that defined as the “simple model” in Figure 1. A credible budget, by definition, should balance—which means that, during the full fiscal year, total budget expenditures equal budget revenues plus budget deficit financing. If this is the case, and budget deficit financing is achieved through non-central bank borrowing, no advances from the BoU are necessary.

When a drawdown occurs in the size of the broader pool as defined by the AGO, two scenarios could have occurred. If the budget subset of the pool has been credibly executed, then the other funds must have incurred greater outflows than inflows during the year. For EBFs and donor fund accounts which do not require a legally balancing budget, such a scenario is perfectly justified and no advance is necessary. Otherwise, however, the budget subset of the pool has not been executed credibly and therefore total budget expenditures have exceeded budget revenues plus budget deficit financing. This situation would lead to a negative end-year balance in the IMF defined pool and therefore implies that budget execution has required greater financing than planned.

Confusion surrounding the interpretation of the drawdown is caused by lack of proper analysis and clear understanding on the movements of cash, and interaction between monetary and fiscal policy. Clearly, under the IMF defined pool, a drawdown means there has been excessive budget expenditure. The BoU claims that this should have been financed by market issuance of government securities since, otherwise, it means that the BoU has advanced the excess as monetary financing which is not permitted.

Only restricting analysis to the narrower pool definition, as performed by the IMF, implies that the BoU is correct. However, if the AGO defined pool is considered, there are again two scenarios if a drawdown has happened. If the extra funds/accounts included have maintained their balances during the year, then BoU advances must have occurred. Or, if the drawdown on the pool has been equal to the reduction over the year in the other funds, any excess budget expenditure has been financed by borrowings from these other funds. Or, a mixture of the two.

It therefore cannot be automatically concluded that excess budget expenditures have been financed monetarily. Only further analysis of the use of the other funds’ cash can determine the extent of this activity, if any. When such analysis is performed, it will reveal the extent of borrowing from the other funds. This borrowing should be made completely transparent to avoid future confusion and to ensure that the other funds are being properly managed and remunerated. It is suggested that, if such borrowing is happening and it is legitimate, the government should issue government securities to the funds to make these transactions transparent and efficient.

26. Although the simple structure is tailored to budget transactions, in Uganda certain inflows and outflows of the UCF are not budget-based. This fact needs to be understood when defining the pool and these non-budget transactions noted in order to compensate for movements in the size of the pool as they occur. By creating the simple pool, most non-budget transactions are excluded because their movements are through EBF accounts outside the definition. However, those transactions which do move through the UCF must be carefully monitored and compensated for when comparing the actual size of the pool cash resources against forecasts.

Further development of the cash flow forecasting model

27. At present, the cash flow forecasting model is based on a spreadsheet template which aggregates projected expenditures, revenues and deficit financing to produce a forecast TSA (pool) balance profile. The revenue data is obtained from the tax policy department and debt data from the debt management office. The CPD is using basic seasonal forecasting methods on historic IFMIS expenditure data to project expenditures.

28. Once the Guidelines have been issued, the expectation is that the expenditure forecasts will be sourced from the MDAs. For all, but the most consistent expenditures, MDA forecasts will be superior to historically-based projections. This is because the most volatile expenditures are usually capital projects which do not normally follow a seasonal trend. Additionally, where strict cash rationing systems have been operating, historic data is heavily distorted by this activity which controls the timing of expenditures by availability of cash resources rather than service delivery payment requirements.

29. The spreadsheet model must be configured to accept MDA forecasts of expenditures. Individual templates must be designed for each MDA in consideration of their size and capacities. It is very important for the CPD to recognize that only significant cash flow amounts are of concern in the forecasting analysis. Therefore, small MDAs will not need to break down their expenditure forecasts to a great degree of detail. Initially, it is often sufficient for small MDAs to provide forecasts only for major heads of classification. In many cases, only three lines will be necessary—personnel, recurrent, and development expenditures—by month (until the CPD feels comfortable forecasting at more frequent intervals).

30. Larger MDAs which have more significant expenditures will require a higher degree of granularity in their forecasts. Understanding the optimal granularity for each MDA will only come with time and experience. As the CPD starts to analyze the errors, or variances, of actual outcomes from the forecasts, it will become evident where significant volatility lies and these points will require the finest granularity in order to understand the errors and try to avoid them in the future. Since the model simply aggregates the total amounts expended in the time interval, it is relatively easy to alter the spreadsheet to adjust to the optimal degree of granularity.

31. The spreadsheet model still needs to be adjusted to handle error analysis. This is straightforward with columns being added to take the actual outturns for expenditures and revenues. The model will be able to determine the significant errors and highlight these on an exception basis.

32. Once the spreadsheet model is complete and operational, investigation into utilizing the IFMIS functionality should be undertaken. Although the IFMIS cash management module is not suitable for government cash flow forecasting (being primarily designed for corporate needs), it can be used to ensure that MDAs send through their forecasts in the format required by the CPD using IFMIS screens. This data can then quite simply be downloaded to the spreadsheet model.

Capacity building by the CPD at MDAs

33. The Guidelines will be issued to all MDAs and these state the main objectives, benefits, and requirements of government cash forecasting and management. They must be backed up, however, by a detailed capacity building program undertaken by CPD staff. It is vital that MDAs appreciate the objectives of cash management and that the main benefit of good cash flow forecasting will be to enable them to make budget expenditures as necessary rather than under cash rationing control. Only once this is understood and they have confidence that it will happen, they will make greater efforts to provide accurate and updated projections.

34. The CPD will need to discuss their data needs for each MDA as discussed above. The degree of granularity required in the forecasts should be decided with a view to supplying the required level of significance with the least effort. It should be explained that cash flow forecasting is not an accounting exercise. Only where it is expected that variances from forecasts will be sufficiently material to the aggregated forecast will it be necessary to provide detailed projections. This understanding will minimize the workloads of both the MDAs and the CPD and should encourage greater compliance.

35. It needs to be explained to the MDAs that the forecasts required by the CPD are quite different from those already given to the budget directorate. It should be made clear that existing work, cash, and procurement plans tend to be distorted due to the need for MDAs to ensure the best claim to scarce resources under the cash rationing system. This means that MDAs tend to front-load their projections unrealistically in order to obtain higher proportions of cash allocations before they are reduced later in the fiscal year by budget adjustments. The forecasts required for cash management, however, need to be more realistic and apply to the needs for service delivery rather than cash availability. Only if this is the case can the CPD in future ensure that cash will be available to meet all budget expenditures as and when they are committed.

36. MDAs must appreciate that their forecasts will be scrutinized by the CPD and significant avoidable errors will be reported to the CMC. Rather than at present, the forecasts will be monitored against actual outturns to determine whether the MDAs have provided realistic projections. The need for greater accuracy in the forecasts must be explained since it will only be when the forecasts of cash flows across government are sufficiently trustworthy that active cash management operations can be used to change from cash rationing to allow expenditures to be made when necessary.

The Cash Management Committee

37. The terms of reference and the composition of the CMC have been included in the Guidelines. Initially, the CMC must provide the CPD with the necessary authority to obtain accurate cash flow forecasts and maintain effective relationships with all stakeholders and counterparties.

38. Crucially, the CMC must also provide the leadership and champion the change management required for the transition from cash rationing to active cash management. This is necessary since there is always significant resistance and misunderstanding to such a change from many different quarters. The transition entails relaxing cash limits or making them identical to the cash flow forecasts once sufficiently accurate. Often, there is concern that there will be a loss of control over budget execution from the macro-fiscal and budget functions. It should be clear that the aim of good government cash management is to ensure that expenditures are paid as and when they are due. There is no challenge to the capability for the budget function to control commitments through the IFMIS. However, the cash management function is responsible for ensuring that, once a commitment has been formally approved and made, the liability is paid as due.

39. Many budget execution problems should be alleviated if the CMC is performing well. Clearly, good cash management, in conjunction with effective commitment control, should prevent incurrence of expenditure arrears. Also, the cash flow forecasting function should monitor significant variances between forecasts and actual outturns. This monitoring function is aimed initially at persuading MDAs to improve their forecasting and updating of forecasts. As the forecasts become more accurate and the CMC permits the cash managers to perform active cash management activities, however, the monitoring highlights unexpected occurrences in expenditure and revenue patterns. This monitoring can point to many factors including inefficient reporting, early contingency warnings, and fraudulent activity all of which should be reported to the CMC for action.

40. The report which the CPD should provide to the CMC each month will change over time as forecasting becomes more accurate and cash management becomes more active. Initially, the report should concentrate on informing the CMC of particular instances where MDAs are not making the effort to provide realistic forecasts of expenditures. It will be natural for them to continue to front-load forecasts in the same manner as for cash rationing. It will be the task of the CPD, backed up by the CMC, to monitor such unrealistic projections through error analysis and press the MDAs to make their forecasts more consistent with actual payment needs.

41. The report to the CMC should evolve into a general cash flow forecast for information to senior management and the BoU. It should compare forecasts to actual outturns and provide general reasoning for significant variances in major heads of classification (as well as continuing to analyze detailed forecast errors). It should update forecasts of cash resources available to the government. This should indicate the forecast profile of the defined pool of accounts but, also, needs to monitor other accounts for any significant build-up of cash balances which are not analyzed in detail. As proficiency is built up through greater experience, the pool should be expanded to include accounts where transfers lead to large cash balances. These should then be included in the detailed forecasting of their cash movements and the associated variance analysis. An outline of the structure of the regular report from the CPD to the CMC is included in Box 3.

Uganda: Proposed Structure of a Cash Management Report to the Cash Management Committee

  • a) The actual cash balances in the defined pool and other accounts at the end of the period;

  • b) The updated projected government cash resource profile for the remainder of the FY held in the defined pool;

  • c) The actual outturn of expenditures and revenues—including deficit financing—for the latest period;

  • d) A general report on the accuracy of the most recent forecasts against the outturns and suggested reasons for material variances (probably against major heads), e.g., “recurrent expenditures were generally significantly lower than forecast, probably due to continued front-loading of forecasts by the Education and Health Ministries’ or ‘tax revenues were higher than expected due to a revision in collection dates”;

  • e) A detailed report on all significant errors in MDA-provided forecasts. These will each show the area of forecast (e.g. the MDA and sub-head classification), the degree of variance of actual from forecast, the extent of investigation by the CPD into the cause of the forecast error, whether the error was deemed to be avoidable or unavoidable, and, if avoidable, what history exists of similar errors and what action is recommended to be taken by the CMC and the CPD;

If active cash management operations are not yet being implemented:

  • f) A report suggesting the active cash management transactions which would be recommended for the most recent projected cash profile;

  • g) A report giving a simulation of what would have been the effect of the cash management transactions suggested at the previous CMC meeting;

  • h) A recommendation regarding when it will be likely that forecasts are sufficiently accurate to commence active cash management transactions and how the target cash buffer level will be calculated;

If active cash management operations are being implemented:

  • i) The calculation of the optimal target cash buffer level in the defined pool;

  • j) The market transactions undertaken in the most recent period and their success, or otherwise, of maintaining the target buffer level of cash resources in the pool accounts;

  • k) A request for approval of a list of recommended market transactions to be performed in the coming period in order to maintain the cash buffer level at the appropriate level;

  • l) A report on the ongoing relationship management of the CPD regarding cooperation and coordination with, in particular, the debt managers, the budget and economic affairs directorates, the AGO, the URA, and large MDAs.


  • CPD should design the CMC reporting structure and initiate CMC meetings when required (February 2017).

  • CPD must precisely define the pool of cash resources to be forecast using budget flows as the simplest model, carefully noting the existence of non-budget cash flows to and from the defined pool (March 2017);

  • MoFPED should analyze the cash resource pool used for the definition of the Government Position and determine exactly how it relates to budget cash flows and non-budget cash flows in order to understand where any budget advances have been made (March 2017);

  • CPD should investigate the possible use of the IFMIS system to interface MDA forecasts with the cash flow forecasting model (March 2017);

  • CPD should continue to develop the cash flow forecasting model (ongoing); and

  • CPD must design and implement a capacity building and training program for MDA forecasting functions—in coordination with the BD and the AGO (ongoing).

IV. Review of the Treasury Single Account Implementation

42. MoFPED has made significant progress with TSA implementation. This has been made possible by MoFPED demonstrating ownership and commitment to furthering these reforms. It has effectively managed stakeholder expectations and concerns by implementing the TSA structure in a relatively short timeframe and has been able to demonstrate tangible benefits for stakeholders. As a result, after the establishment of the core TSA structure covering budgetary central government, AGO has been looking to progressively extend its coverage.

43. The core central government TSA structure consists of a separate Uganda Consolidated Fund (UCF), a TSA holding account, and TSA sub-accounts for accounting entities. It covers the budgeted domestic revenues, expenditures and financing and concentrates on balances held at the BoU and managed through the IFMS. This enables cash balances for the TSA structure to be aggregated overnight and consists of the following:

  • The UCF that receives domestic revenues from the Uganda Revenue Authority and non-tax revenue holding accounts and receipts from domestic T-bills and T-bonds issued for fiscal purposes.

  • The TSA holding account receives funds through transfers from the UCF based on quarterly expenditure limits approved by MoFPED and entered into IFMIS against each vote. There are also separate holding accounts for transfers to missions, local government, hospitals and educational institutions.

  • TSA sub-accounts are operational accounts for each accounting entity through which payments for recurrent and development accounts are effected. These are ‘just-in-time’ funded from the TSA holding account up to three times a day to cover payments for authorized invoices in IFMS. Any unspent balances are automatically swept back overnight to the TSA holding accounts.

44. Operationally, this structure has benefitted from complementary changes introduced by MoFPED and the BoU. These have included the IFMS roll-out, establishment of a centralized payment system by Accountant General, and straight through processing of payments using electronic fund transfers. The BoU has also supported the process by extending the banking hours, reconfiguring their system, conducting daily sweeping of balances from the TSA sub-accounts, and linking the BoU IT systems with the IFMS.

45. Recent changes that have been introduced in the centralized payment system do not impact on the operations of the TSA. The earlier process involved straight through processing of the consolidated payment files generated on IFMS by the AGO through electronic fund transfers. This was applicable to both domestic and foreign currency payments. AGO has modified the system in view of some frauds reported in 2015 and these payments are now processed with a lag of two days. In the revised process, once the payment instructions are received from AGO, BoU deducts the amount from the respective TSA sub-account and sends an e-statement to AGO/MDAs. MDAs have thirty hours to challenge the correctness of payment after which all unchallenged payments are processed automatically.

46. MoFPED has been implementing a phased approach to further extending the coverage of the TSA structure. Previous FAD/AFE reports have provided guidance on the factors that should be considered while deciding on the scope and timing of extending the coverage.3 These include consideration of the purpose of bank accounts, relative size of balances of their balances, IFMS roll-out, and the capacity of the MoFPED to monitor and include them in the cash plan. In the initial phase of implementation, MoFPED has concentrated on the UCF and those central government entities that are covered by IFMS. This approach is appropriate and consistent with previous advice from FAD/AFE.

47. The inclusion of donor project accounts into the TSA arrangements continues to pose challenges. Despite the rationale for inclusion of these accounts in the TSA, there are considerable change management activities that would be required to convince the donors to support the TSA. These include providing them with assurance and confidence that the project funding would not be diverted for other purposes; that donor specific reporting requirements would be met; and individual project implementation would not be impacted under the new arrangements. Building sufficient donor confidence by establishing procedures and demonstrating efficacy of the in-country systems always takes time. As a result, MoFPED has not been able initiate the planned pilot for inclusion of sample project accounts within the TSA. The mission was informed that the African Development Bank and the International Fund for Agricultural Development (IFAD)4 have been receptive to the idea of the pilot and renewed efforts should be made to reach an early agreement on this in order to demonstrate its effectiveness.

48. In the meantime, quick progress has been made on incorporating many Local Government (LG) accounts into the TSA. This has been made possible because of the roll-out of IFMS to certain LGs; connectivity and technological advances that have obviated the need for LGs to maintain multiple bank accounts; and extension of the central government payment process to LGs. An earlier AFE mission had outlined an approach for the LGs arrangement for the TSA and this has been adopted by the AGO.5

49. The AGO has adopted a tiered approach in establishing the LG TSA sub-structure. The tiers relate to the nature of financial management systems being used by the LGs and these include: (i) Tier 1 that use the Oracle IFMS which is the same as the national government MDAs; (ii) Tier 2 using Microsoft Dynamics, which is a standalone IFMS that is not interfaced with the main IFMS; and (iii) Tier 3 that do not use IFMS and, except for payrolls, and process their payments manually. Salary for all LGs is paid centrally through the TSA. The implementation of LGs TSA started in July 2015 and since then eighty out of the one hundred fifty-two LGs have been included in the TSA arrangements. The LGs TSA is held at the BoU and modeled on the central TSA for MDAs. The overall status of the LGs in respect of their TSA banking arrangements is in Box 4.

Uganda: Status of Treasury Single Account Arrangements for Local Governments

  • Tier 1: 14 LGs are serviced through the LG TSA holding account that receives transfers from the UCF and the LG revenue collection account at BoU for their own source revenues. Transfers from the holding account are made just-in-time to the respective LG TSA sub-account and balances swept back to the LG TSA holding account.

  • Tier 2: 66 LGs are serviced through their own TSA sub-accounts at BoU that receive transfers from the UCF and own source revenue collections from their holding accounts. LGs have access to their TSA sub-accounts and can monitor the balances.

  • Tier 3: 78 LGs currently are outside of the TSA arrangements except for salary that is processed centrally through the IFMS. The rest of the transfers are processed through warrants in IFMS and paid to the General Fund Accounts held at the district level in the commercial banks. Further progress on the inclusion of these LGs will depend on extension of IFMS to cover their operations.

50. Incorporating the LGs accounts into the TSA provides AGO with an opportunity to rationalize the number of TSA holding accounts at BoU. The TSA structure has a number of TSA holding accounts at BoU for secondary and primary schools, primary health centers and town councils that receive funds from the UCF. Further transfer from these holding accounts is made to these entities through IFMS. AGO plans to close these TSA holding accounts and instead transfer funds from the UCF to the LG TSA holding account to enable LGs to make direct transfers to the beneficiary entities through the IFMS. These holding accounts are intended as a mechanism for identifying and accounting separately for transactions and balances of these accounts and managing budget releases. AGO should continue to review and rationalize the number of accounts in the TSA structure at BoU as the accountability purpose can be managed through the IFMS ledger structure without the need for separate bank accounts. A smaller number of accounts will also obviate the need to process a large number of inter-account transfers, corresponding accounting entries and associated bank reconciliations. This will also help in clearly defining the scope and coverage of the UCF/TSA for cash flow forecasting purposes as discussed in the previous section of this report.

51. Inclusion of special purpose funds, autonomous and semi-autonomous government agencies into the TSA arrangements will require careful consideration. The special funds relating to Road, Energy, Petroleum and Contingencies are held at BoU but have been established legally outside of the UCF. The TSA structure should allow for aggregation of balances while recognizing the legal separation of these funds and where agreed respecting the requirement for ring-fencing the balances. Their inclusion into the TSA arrangements may require redefining their relationship with the UCF in context of the Uganda TSA. The autonomous government and semi-autonomous government agencies that receive their primary funding from the national budget, are outside the IFMS and deliver services on behalf of the central government (e.g. referral hospitals, tertiary educational institutions, commissions etc.) should form part of the TSA. AGO should consider the nature of their operations and explore the feasibility of maintaining their balances in the BoU using TSA sub-accounts. However, government-controlled state owned enterprises that are profit oriented and government financial entities should be left outside of the TSA.

52. Further extension of the TSA and improvements in the banking arrangements would benefit from an early finalization of the MoU between MoFPED and BoU. Significant changes have been made to the business processes of both MoFPED and BoU including: establishment of a TSA structure of bank accounts; extension of the TSA arrangements to LGs; centralized payment system; and linking the IFMS with the banking applications. It is important to formalize appropriate protocols around these processes to ensure smooth and coordinated operations between MoFPED and BoU. This would also be consistent with the MoFPED ambition to manage the UCF/TSA balances through cash planning and active cash management.

53. The mission was informed about on-going work on revising the existing MoU (2008) that defines the relationship between MoFPED and BoU. An advanced draft has been circulated and commented upon by different MoFPED directorates and BoU but the work has not advanced as expected. The process has been deferred pending the proposed amendments to the BoU Act. The draft of the MoU/SLA was not shared with the mission. As advised in the previous AFE reports, the MoU/SLA should cover the management and operational roles around the banking arrangements; definition of the government cash position; treatment of and compensation for the UCF/TSA position including payment of interest on government deposits; and fees and charges to be paid to the BoU for the services provided. It should also take into account the new PFM Act that has rules and controls on intra-year use of balances of the UCF, and the East Africa Monetary Protocol that requires partner states to phase out any central bank lending to its government or public entities.


  • Continue to bring LGs within the TSA coverage; and rationalize the UCF/TSA bank account structure with extension of IFMIS implementation (ongoing).

  • AGO should continue to explore with receptive donors the options for piloting donor funds within the TSA arrangement (March 2017).

  • Finalize the MoU/SLA between MoFPED and BoU setting out the terms of engagement for banking arrangements including treatment of and compensation for the UCF/TSA balance and charges for banking services (March 2017).

Uganda: Technical Assistance Report-Strengthening Cash Management and Reviewing the Treasury Single Account
Author: International Monetary Fund. Fiscal Affairs Dept.