CASH MANAGEMENT: HOW DO COUNTRIES PERFORM SOUND PRACTICES?

I. Cash Flow Forecasting

Cash management is a process which aims at making the right amount of money available at the right time in the right place to execute the government’s payment obligations. This widely accepted definition seems simple but putting it into practice might be challenging. In case the full amount (right amount) is not transferred, at the required time (right time) to the beneficiary (right place) the government can face with payment arrears and default. This three-legged process can be called as “calibration of 3Rs” where each R represents a “right” for the amount, the place and the time. Finding the “right” balance of 3Rs is like three-ball juggling, and none of the balls should drop anytime. Therefore, it requires a strong cash flow forecasting capacity and the ability to implement cash flow projections.

The main objective of cash management is to minimize gaps between cash inflows and outflows at any point of time, ideally in the next 3 months. To this end, in order to determine the “right amount” for the required payments, cash managers need to project expected revenues and payments, and identify the gaps.

Governance and Institutional Framework of Cash Flow Forecasting

Cash flow forecasting is key to ensure that government’s payment obligations are fully met, and arrears are not accumulated. However, if cash management is not well-positioned in the government’s planning and resource management cycle and has weak linkages with debt management, a cash manager will not be able to produce accurate projections of cashflows and cannot provide reliable and timely information to decision-makers, spending agencies, Central Bank and markets. Therefore, as confirmed by sound practices, setting appropriate institutional arrangements and identifying the responsibilities across institutions are essential for effective cash management.

Cash management, as an element of strong public financial management, helps keeping payments within the budget limits, through authorized budget allocations. It can further improve realism of fiscal projections and assist decision-makers to identify corrective fiscal actions. However, cash management is not part of the budget cycle and not a tool to maintain fiscal control.

Some countries consider cash management as a part of budget management and position it either entirely (Chile) or only some of its functions (Ghana, Seychelles) within the Budget Department. Besides, some countries carry over cash management fully or partially in the Accountant General’s Office (Eswatini, Lesotho, Nigeria). There are also cases where cash flows are projected by other departments (e.g. Macroeconomic Policy Analysis Unit of The Gambia).

Figure 1
Principal entities in charge of cash management in workshops’ participating countries

As shown in Figure 2, in most of the assessed countries, cash management is a function of the Treasury and positioned either as a dedicated unit (South Africa, Kosovo, Rwanda) or dispersed across divisions. It is also widely reported that cash and debt management functions are integrated (Colombia, Kosovo, Peru, Romania, Slovenia, South Africa and Thailand). This is indeed parallel to the practices in advanced countries, as demonstrated by a recent survey that integrated cash and debt management is a common practice across the OECD countries (Cruz and Koc, 2018). On the other hand, there is no specific solution which fits all countries. Therefore, institutional arrangements may vary across countries as long as main features of effective cash management are secured. Clear definition of responsibilities and workflows within and among the entities in charge of cash management, well-organized internal controls and existence of analytical capacity indicate that “appropriate institutional arrangements and responsibilities” are in place as defined by Lienert (2009).

Regardless of the reporting lines, a Treasury Single Account (TSA), where all the government revenues are collected and payments are made, is the main source of actual data for cash managers. TSA is the common practice across countries. Parallel to the general practices, in all countries discussed in this paper, if the TSA is in place, it is in the Central Bank and the main account is in local currency, even when there are TSA subaccount in foreign currency. Countries such as Lesotho and Nigeria recognized the TSA’s pivotal role in cash management and recently established their TSA systems. There are only a few countries that do not have yet a system for a centralized bank account (e.g. Equatorial Guinea) due to, among other reasons, limited capillarity of central bank branches and insufficient payment and process systems. Gradual implementation of a TSA “system” by consolidating government funds in fewer accounts, respecting existing country limitation is advisable.

In a case where cash management functions are fragmented, coordination becomes more challenging and inefficiencies in organizational arrangements may lead to less accurate cash flow projections. In order to overcome the deficiencies in coordination, many countries have established high and/or technical level committees (Bolivia, Chile, Ghana, North Macedonia, Romania, Gambia and Rwanda).

A. Cash Flow Forecasting Practices

In the absence of cash flow forecasting and planning, governments mostly make the payments in line with the priorities of the political economy, if there is not a sufficiently large cash balance. In such cases, inefficient allocation of financial resources may result in over or under borrowing that can also lead to increased costs and risks. On the other hand, in cases where comprehensive and accurate cash flow forecasting is in place, cash management ensures required funds are transferred to the beneficiary’s account at the right time. Along with the methodology applied to forecasting cash flows, the coverage, time horizon and frequency of forecasts are also critical to find the right balance of 3Rs.

The Government’s budget is an output of broader national policy processes and includes directives and limits for the collection and allocation of public resources. The strategic budgeting phase involves government institutions such as finance and planning ministries, sector ministries, large spending agencies and other entities like the civil society groups. Based on the strategic directions set at the policy level, budget officials prepare the formal budget which includes estimated revenues and expenditures for each budget entity. Considering that the government budget provides information on the size of the revenues and expenditures and about the public entities entitled for these transactions, as the first step, a cash manager produces an annual cash flow plan based on the budget. This process is simply an application of a top-down approach: (i) review aggregate budget figures (ii) conversion of data from budget to cash-base, (iii) generation of monthly, weekly and daily cash flow forecast with application of analytical models on the historical TSA data, (iv) incorporation of possible changes in government policy and regulations which may affect the collection of revenues and use of public resources, (v) addition of contingencies (Figure 2).

Figure 2
Cash Flow Forecasting Approaches: Top-down vs. bottom-up

The main output of cash flow forecasting is accurate daily forecasts for at least one month ahead, preferably for the next three months. However, the granularity of the budget data and their frequency do not allow the cash manager to determine the exact time of cash inflows and required time of the payments. Therefore, in order to find the right balance of 3Rs and produce reliable forecasts, cash managers combine the top-down approach with a bottom-up approach.

Coverage of cash flow forecasts

Cash management is part of the public financial management process and directly affected by the effectiveness of budget planning and execution. In principle, expected revenues and expenditure ceilings of central government agencies set in the government budget are the main inputs of cash flow forecasting. On the other hand, it is a fact that in many countries, in particular in the developing and emerging countries, the coverage of state budget could be limited, because public funds are utilized through extrabudgetary funds (e.g. Vietnam¹) and/or earmarked accounts (e.g. Ghana, South Africa²). However, in case that these entities/accounts collect and spend public resources through Treasury bank accounts, they should be included in the cash flow forecasts even if they are not covered by the budget.

The scope of the TSA is critical to determine the coverage of cash flow forecasts and cash management, which is the central government in most countries. However, depending on the decentralization of government, and the compliance of accounting and control systems, the scope of cash management (and of the TSA) can be extended from central government to general government level (e.g. France).

Contingent liabilities, such as loan guarantees and commitments of public-private-partnership contracts, are classified as off-budget items. These liabilities are monitored by most countries, and paid by the government, when called. In order to avoid any default due to unanticipated payments, potential calls from the contingent liabilities should be included in the cash flow forecasts.

Horizon and frequency of forecasts

Multi-year budgets have become more common and may provide a longer-term horizon for cash flow projections. However, even for the countries with multi-year budgets, the annual budget that is formally approved by parliament, is still the key document. Accordingly, with some exceptions, the horizon of cash flow forecasts is not going beyond one year. On top of budget estimations, cash flow forecasting and cash management are affected by a number of factors such as the means of payments (e.g. cash, cheque, electronic payments), the payment agents (e.g. Central Bank, commercial banks), the cash management capacity of central government and local government institutions, and the depth and liquidity of government bond and money markets (Miller and Hadley, 2016). Therefore, both horizon and frequency of cash flow projections rely on different parameters and show significant variances across countries. There are countries (e.g. South Africa, Albania) preparing cash flow forecasts for 3-years, yearly, monthly, weekly and daily periods, but on the other side, there are cases where only monthly cash flow forecasts are in place (e.g. Angola, Nigeria). Most of the countries are preparing yearly cash flow projections with different frequencies such as monthly (Kosovo) weekly (Seychelles) and daily projections (Bolivia, Colombia, Honduras, North Macedonia, Uruguay). In parallel, the periodicity of the revisions of cash flow forecasts differs across countries. In most cases the projections are reviewed and revised weekly or daily. However, there are countries revising the forecasts just monthly.

Box 1. Extensive Cash Flow Forecasting: South Africa

Extended cash flow forecasting enables South Africa to ensure that government manages its cashflow to have sufficient cash to meet its obligations; to neutralize the impact of government’s cash flows in the liquidity of the financial system; and to maximize returns on surplus cash. One of the objectives of extended cash flow projections is to ensure there are no large and unpredictable changes in liquidity in the banking system and monetary policy is not undermined.

In the cash flow projections, National Treasury forecasts revenues, expenditures, domestic and external debt redemptions, financing of borrowing requirements and the resulting change in cash balances for 3 years with different frequencies,

The Debt Management Performance Assessment (DeMPA) tool of the World Bank evaluates the reliability, horizon and frequency of cash flow forecasting, through Debt Management Performance Indicator (DPI) 11.1³, which requires monthly forecasts for getting Score C, weekly breakdowns for score B and daily estimates for Score A, the highest. As of end-2019, DeMPA assessments have been implemented in 80 developing countries, and only 28 percent of them met the minimum requirement or above (Figure 4, blue bars).

Figure 3
Share of DeMPA assessed countries that reached minimum requirement (score C or above)

As prescribed by DeMPA methodology, countries should ideally prepare daily cash flow forecasts, but even when the breakdown of the estimates is monthly (minimum requirement), it is advised to cover the budget year (horizon). On the other hand, given the budget figures may not be available yet closer to the end of the fiscal year, cash managers should have the capacity to estimate the cash-flows for the initial months of the following budget year.

Revenue forecasting

It is the common practice to forecast revenues through the application of a top-down approach considering that (i) main tax items (e.g. income tax, corporate tax and V.A.T.) and customs’ tax and duties comprise the significant part of total collections, which for good estimates is provided by the annual budget’s tax projections, (ii) revenues have strong seasonality and it is practical to generate monthly and weekly breakdowns from annual budget figures by using historical data, and (iii) it captures necessary adjustments due to the changes in government policy and regulation, macroeconomic and financial developments. Accordingly, in most countries, cash managers rely on the projections of the tax administration. On the other hand, there are cases where revenues are projected by the cash management unit (Peru), the fiscal policy office (Thailand), or the macroeconomic forecasting unit (Seychelles).

For the revenues other than tax and customs, such as dividends, grants and receivables from on-lent loans, the best way to produce forecasts is to apply bottom-up approach which refers to the collection of information from relevant institutions as the granularity is not the same and individual flows can be followed.

Expenditure forecasting

Given that the government budget imposes expenditure ceilings through appropriations, and the nature of non-discretionary expenditures is quite predictable, cash managers are relatively more comfortable for expenditure forecast. However, considering the time lags between the release of appropriations and actual payments from the TSA, cash managers tend to rely more on the historical TSA data in the expenditure forecasts. Besides, historical data, and the laws and regulations may also provide useful information on the payment dates of regular and large transactions.

Discretionary expenditures (e.g. goods and services, capital expenditures, subsidies and transfers) are prone to government policy decisions that makes them harder to forecast. Therefore, it would be more efficient to concentrate on major projects and programs which require large payments, rather than aiming to forecast all expenditures. On the other hand, in most countries, linked to the budget appropriation’s release calendar and due to the budget’s spend-it-or-loose-it policy, discretionary expenditures, in particular the goods services and capital expenditures, show a strong seasonality due to the weak performance at the beginning of the year and, on contrary, high spending in the last few months.

Due to the fact that cash managers cannot know the physical conditions of the infrastructure investments and are not able to assess the best time the goods and services are needed, sector-ministries and spending agencies are the best resources to provide information for the projection of discretionary expenditures which refers to a bottom-up approach.

Although the quantitative analysis and spending agencies’ inputs are at the core of expenditure forecasting, qualitative judgments contain additional information that cannot be incorporated into the models. They allow cash managers to adjust forecasts considering the reliability of data collected from different sources and the counterparts’ forecasting performance.

Incorporating debt data

Debt interest and principal payments are significant elements of cash flow forecasts, along with the cash inflows from the issuance of new debt. It is a common practice that debt managers provide cash flow projections related to debt service and borrowing. In addition, cash managers commonly coordinate with debt management office to collect information on the contingent liabilities, in case they may materialize.

Cash flow forecasts associated to floating, inflation-linked, foreign-currency denominated or any other variable debt need to be estimated using macroeconomic scenario(s) that not necessarily (or commonly) are produced by debt managers (a third party usually generates the required estimates, e.g. Central Bank or macro unit at MoF). Shock scenarios can be useful when defining a liquidity buffer policy, as it is also the case on the design of a Medium-Term Debt Management Strategy (MTDS).

Accuracy of forecasts

Producing reliable and accurate cash flow forecasts is crucial for public financial management policy setting and implementation through (i) providing critical information to decision making to take corrective fiscal actions, (ii) avoiding accumulation of arrears and (iii) minimizing cost of carry.

There are factors that create deviations between the initial forecasts and actual figures, such as external events, change in government policy, or lack of information. In order to eliminate some of these factors and improve accuracy of cash flow forecasts, countries introduce monitoring mechanisms and preventive actions. Real time monitoring of TSA balance allows cash managers to assess the performance of cash flow forecasts, but strong electronic banking/payment infrastructure is a prerequisite. More realistic budget projections, sensitizing the line ministries on the cost of forecasting errors and mitigation measures such as developing more coercive framework to prevent unexpected large payments help improve accuracy of forecasts.

Forecasting deviations from actual cash flows should be frequently reviewed in order to improve the accuracy of the projections and help cash managers understanding the sources of errors. To this end, it is useful to keep a repository of historical forecast data that enables granular analysis of the deviations throughout different periods of the year and time lags between estimated and realized cash flows. Just 9 out of the 24 workshop participating countries presented numbers of the accuracy of their forecasts, indicating that thorough assessment of estimates performance has not been extensively done in most of these countries and pointing out to a possible area for improvement.

A liquidity buffer, discussed in detail in Section III, is a useful tool to deal with deviations between the projections and realizations.