For fiscal economists, the key issues on budget execution are always whether deficit targets are likely to be met, and whether any budget adjustments (both on the revenue and expenditure sides) agreed at the preparation stage (or in-year) are being implemented as planned. On the expenditure side of the budget, the key issues are whether the outturn is likely to be within the budget figure; whether any changes in expenditure priorities (as against past patterns) are being implemented in specific areas as planned; and whether any problems are being encountered in budget execution, such as the buildup of payment arrears.

Fiscal economists therefore need to fully understand any weaknesses in the country's budget execution process. Is it transparent? Are there clear lines of accountability? Is information on execution of the budget available on a timely, reliable, and accurate basis? Is it thus consistent with the principles of good governance? Based on this understanding, where are problems likely to arise, and how might they be avoided or overcome? In some instances, action may be needed through budget execution procedures to bring expenditures back on track to the budget provision; hold expenditures below budget, in response to below-target revenue developments; or bring irregularities to the attention of the decision makers.

Thus, for fiscal economists and general budget advisors, the key questions are:

• What are the different stages of the budget execution process?

• Who is responsible for budget execution?

• How can budget appropriations be revised during the year?

• How good is the information on outturn expenditure?

• What are the problems encountered in budget execution procedures and how can these be overcome?

• How can expenditures be adjusted in-year?

• How should "good governance" be pursued?

After the legislative appropriation of expenditures, there are usually six main stages in the spending process.

1.  The authorization stage

Once a budget is approved by the parliament, ministries are authorized to spend money, consistent with the legal appropriations for each line item. Where parliament has not yet approved the budget before the budget year starts, it is normal to allow governments to start spending on a "Vote on Account" basis--a temporary authorization, often restricted to one-twelfth per month of the previous year's expenditure. In the francophone, Latin American, transition, and many Commonwealth countries, once approved, parliamentary authorization is for one year. In some Commonwealth countries, however, the authorization period may be set monthly or quarterly by warrant.

In the majority of countries, unspent funds in one year cannot be carried forward (carryover) to be spent in the next. In some OECD countries, however, unspent operating funds can be carried forward, usually up to a specified small percentage of the total funds (e.g., Australia, Canada, most Scandinavian countries, and the United Kingdom); and in some countries cash to pay for obligations incurred in one fiscal year but falling due in the next can be carried over (e.g., Italy, Japan, New Zealand, and the United States). However, it is more common to allow the carry-forward of some element of capital appropriations (or in some cases program expenditures), to allow for changes in the phasing of projects compared with the original budget plans, while still maintaining the same total cost.²¹

In some OECD countries where the emphasis is on giving agencies more freedom to manage their resources within an overall agency-specific budget to improve efficiency, and where multiyear expenditure planning is well established, the trend has been toward a greater use of such carryovers. However in these countries, aggregate expenditure control is much less of a problem and the prime objective is ensuring the most efficient and effective use of government resources. These circumstances do not typically apply in non-OECD countries, where the use of carryovers should generally be discouraged in the interest of financial discipline.

2.  The commitment stage

This is the stage where a future obligation (liability) to pay is incurred. The precise definition of commitment varies not only from one system to another but even among those well-versed in public sector accounting. Broadly, a commitment arises when a purchase order is made or a contract is signed, which implies that goods will be delivered or services rendered, and that a bill will have to be paid later on. But, as noted below, there are shades of interpretation. Good budget systems maintain data on commitments that can be monitored, because these will (for the most part) ultimately be reflected in actual expenditure and because their profile, in terms of cash payments to be made, may have important financial programming implications. But there are complications to be aware of.

• The existence of a commitment does not ensure that the goods will actually be delivered or the service rendered because the relevant ministry or spending agency may change its mind or disagree with the supplier later. This is especially true in countries with poorly organized public expenditure management systems, not least because suppliers are not guaranteed payment.

• The nature of commitments varies by economic category of expenditure. A critical dimension is the lag between entering into a commitment and the associated cash payments: this is especially important for the purchase of capital goods and current nonwage goods or services. But a debt interest payment or the wage bill, both due monthly, are also types of commitment.

• A commitment does not mean that a payment will be made within the same fiscal year--the payment may be made the following year. This is especially true for investment expenditure.

• In many countries, exceptional procedures allow an expenditure to be made without a previous commitment.

• In some countries, what is interpreted as a "commitment" is at best a reservation; that is, the request from a spending unit to the budget authority to put aside an allotment for a future expenditure. This cannot be considered a commitment in accounting terms, because no contract is signed at this stage. Some officials in transition economies tend, erroneously, to equate commitments with budget appropriations.

• In francophone and some other countries, there is a dual control over commitments: administrative control via the line ministry or spending agency and financial control by the ministry of finance. The ministry of finance's financial control represents a kind of "preaudit" confirmation that a commitment can be entered into, consistent with the appropriation.

• In many systems, commitments are either not recorded at all or the accounting for commitments is not consolidated (e.g., the line ministries and spending agencies record these only internally). When there is no centralized accounting of commitments, there is a potential danger of accumulation of payment arrears because no one ensures, when commitments are incurred, that they are consistent with planned future cash availability.

This signifies that goods have been delivered fully or partially according to the contract, or the service has been rendered and the bill has been received. Physical delivery can precede verification by some period of time. The line ministry or spending agency making the purchase usually has the financial and the administrative responsibility to check the bill; that is, to verify that the supply has been received in full compliance with any terms or conditions. The bill at this stage is recognized as a liability of the public sector, in an accrual accounting sense, and is therefore an important stage of the expenditure process. Even though it represents an accrued liability, it may not yet represent a cash liability, however--for example, when a grace period of 30 or 60 days was included under the terms of the purchase order. Information on verifications within the central government sector, however, is not usually available on a centralized basis.

4.  Payment authorization or payment order stage

This stage may have a different significance in different systems.²² In the francophone system a guiding principle is that the person who orders the supply (engagement) has to be different from the one who authorizes the payment (ordonnancement). The payment officer is normally a public accountant who belongs to the Comptabilité Publique and has specific responsibilities in terms of the expenditure process for authorizing the payment of verified bills. After verification of the bill, the spending unit must then hand it on to this public accountant, and request that the bills be paid; payment orders are normally centralized at the ministry of finance. For expenditure management purposes, this procedural distinction is not of major significance, although it does imply a different institutional source of data on payment orders than under many commonwealth systems (see below).

In contrast, in some Latin American countries the function of postaudit and payment is undertaken by the same institution, a Contraloría General, which also exerts a preaudit function on commitments. In this case the source of data on different stages of spending is the same institution.

In Commonwealth systems the issue of payment orders is typically the responsibility of the financial officer with delegated responsibility for this function. Systems vary: the issue of payment orders and checks may be decentralized--with spending ministries carrying out these tasks and reporting back to the center--or centralized in a treasury department, typically called the accountant general's department within the ministry of finance, which acts both as paymaster and prepares the final accounts of the government.

In transition economies, the situation also varies, but most countries now have treasuries that are increasingly responsible for the issue of payment orders. Some so-called "power" ministries, like defense and internal security, often have (so far) retained separate systems. Other inherited elements of the previous system can also be very misleading: where there are different tiers of spending units (first, second, third, etc.), some ministries of finance regard expenditure as having taken place when money is transferred from ministry of finance bank accounts to the first-tier units. But, in unreformed systems, that money may take some time to be further transferred to subsidiary units and then constitute "final" expenditure on goods and services. It is therefore necessary to distinguish between such final payment orders and transfers that really represent payment authorization.

5.  Payment stage

At this stage, the bill is paid--by cash, check, or electronic transfer. In some systems, the payment is made through a single ministry of finance account in the central bank or in a designated bank. In others, the payment is undertaken through the commercial banking system via bank accounts held in the names of individual line ministries. (This latter approach can make it more difficult for the ministry of finance to reconcile its accounts with those of the banking sector.)

6.  Accounting stage

The cash transactions are recorded as complete in the books, which allows a reconciliation from the cash based "above-the-line" fiscal accounts with the financing of any deficit "below the line." Some countries are moving toward accrual accounting, whose differences with cash accounting are discussed in Box 5.

For fiscal monitoring purposes, countries with less well-developed systems often concentrate on the final encashment stage of payment, when checks from government accounts are cleared through the banking system--the cash accounting stage. This is usually the most reliable, with timely information available on a centralized basis, and it allows for direct reconciliation with the monetary accounts. Indeed, the payments accounting data from the banking sector act as a check to ensure that the "above-the-line" information from government accounts is comprehensive and accurate. But, in many countries, it is also necessary to estimate expenditure in accrual terms, or, as it is sometimes termed, on a commitment basis. In principle, this enables an assessment of the extent of payments that will fall due or the arrears being accrued. Such data are most readily available under the French system. In Commonwealth and transition countries, it may be difficult to obtain reliable information directly: in some instances, the information must be assembled as payments made, plus estimated arrears (see the subsection on problems in budget execution).

Thus, the fiscal economist should always be sure whether information conveyed about spending to date refers to commitments, verifications, payment orders issued, or payments encashed. All data received should be collected at the same stage of spending or adjusted to reflect any inconsistencies. This may not be an easy task when data are derived from more than one source.

Budget implementation, in the sense of delivering services by undertaking expenditures, is the responsibility of the line ministries and spending agencies, within regulatory controls set by the ministry of finance. While financial information on budget execution is usually available on a centralized basis in the ministry of finance, the data are collected either through the line ministries or through ministry of finance controls in the system. These information flows must be identified; and they may differ substantially depending on the systems.

In the francophone system, such information is, in principle, most easily accessible. Commitments may be authorized by the ministry of finance or by the financial comptrollers;²⁴ in both cases the ministry of finance should have a consolidated statement of commitments. A department in the ministry of finance may centralize all the payment orders or these may be authorized by the financial comptroller (comptabilité publique), so again this information should be centrally available. The ministry of finance should be able to produce regular statements--every day where the system is computerized--on cash payments.

In Commonwealth countries, by contrast, some fragmentation in the budget implementation process is typical. A treasury or finance department may be in charge of authorizing expenditures--perhaps by issuing annual, quarterly, or sometimes monthly warrants to line ministries, which in turn contract and undertake expenditures, maintaining their own books of accounts. The monitoring of the implementation of development and recurrent budgets, however, may be undertaken by different ministries (e.g., the ministry of planning and the ministry of finance, respectively). In many countries, however, the accounting for budget execution transactions is carried out by the chief accountant in the spending ministry, who is nominally on the staff of the accountant general's department, a statutory appointee. The accountant general's department itself is in charge of keeping accounts of the government, making reports to top ministry of finance management, and presenting the final accounts to the legislature.

Typically the accountant general's department also has paymaster functions, being responsible for the issue of checks and the distribution of cash to meet other types of payment. But,

• sometimes debt service payments, at least the preparation of payment orders, are handled separately by a department other than the ministry of finance; and foreign-financed projects often have different payments procedures;

• sometimes the issue of checks is decentralized and carried out by staff from the accountant general's department in line ministries;

• sometimes it is centralized and carried out by the accountant general's department. But the centralization of check writing can lead to delays in processing payments due.

Transition economies have been moving away from a system where the budget was a part of the state plan, and replacing it with a centralized treasury system. Under the prereform system, spending targets, which were set sectorally in conformity with the sectors of the state plan, formed the basis of the budget. The main features were as follows:

• Budget execution tended to be more decentralized. The ministry of finance distributed funds to the main ministries, which in turn distributed these funds to the bank accounts of a large number of ministries and from these to an even larger number of budget institutions--many of them being enterprises.

• Payments were executed through the banking system usually electronically, and through the previously monolithic state bank system, reports were generated on payments made by each institution. While timely, the information derived was on a highly aggregated functional basis, with little economic content.

• Line ministries and spending agencies did report quarterly, in great detail, on their progress in executing their budgets, though with a considerable lag.

• Thus, despite the preconception of being highly centralized, budget execution was rather decentralized. Usually, a department within the central ministry of finance compiled the data on budget execution and prepared the accounts to reconcile them with the records of the state bank.

• This centralized department within the ministry of finance has typically formed the basis of new treasury departments that have subsequently been developed in these countries and that usually have assumed responsibility for the budget execution process.

During budget implementation, many countries find that they wish, or need, to change the line item appropriations approved by parliament. Some such revisions are necessary and desirable, but excessive switching of budgetary provision between items of expenditure (virement) and excessive use of supplementary estimates cause difficulties, and usually indicates a lack of budget discipline.

Once budget appropriations are set in the budget on a line-item basis, there is often a need to shift budgetary provision from one line item to another; an example might be the delay in one capital project because of bad weather, and the need or opportunity to accelerate work on another such project elsewhere. In general rules are set out (e.g., in the Organic Budget Law or in Financial Regulations in Commonwealth systems) on who can undertake such switches--often called virement. Typically, within a single economic subcategory and program, the spending agency itself is permitted to make switches--for example, from the provision for one utility bill (say, water) which is below estimate to, say, the electricity bill, which is above. But switches between economic categories--for example, from subsidies and transfers to purchase of goods and services or from one program to another (even if the responsibility lies with the same line ministry) typically should require the approval of the ministry of finance. Some switches, for example, from any other economic category to wages and salaries are usually absolutely forbidden. In some countries, the authorization needed to make switches between line ministries may require approval by the president or parliament.

Although virement is an acceptable practice it should not be abused because:

1. Excessive virement is an indication that budget preparation has been too casual; such virement exercises are not costless (they absorb administrative resources) and they discourage effective expenditure planning.

2. The use of virement ostensibly for one purpose, which then "leaks" into another; for example, higher wages and salaries, is often a reflection of poor budget classification.

3. The switch may not be genuinely one for one--that is, the reduction taken on one line item is not sufficient to pay for the increased expenditure on another.

The principles to observe in the use of supplementaries are as follows.

1. It is better to acknowledge expenditures in supplementaries than resort to "off-budget" transactions, to using suspense accounts (see glossary), etc.; the basic concept that the parliament should approve all expenditures must be regarded as sacrosanct.

2. That said, it is the responsibility of ministries of finance to exhort line ministries and spending agencies to live within the budget resources allocated to them--the "hard" budget constraint; that means not giving supplementaries too readily but encouraging the switching of resources from lower-priority expenditures to provide for expenditures above budget provision elsewhere, wherever possible during the fiscal year.²⁵

3. Financial regulations, backed by actual practices, must stick to the basic concept that supplementaries cannot be assumed. Spending above provision without an agreed drawing on a contingency reserve, virement, or a specific supplementary for that item is an illegal act that should be subject to disciplinary action.

4. Supplementaries should be approved only at fixed times of the year. The best practice is once at the end of (or in some systems immediately after) the financial year in question, where the expenditures have been legally financed from a contingency reserve in the meantime. In other systems, twice-a-year use of supplementaries is followed; the more frequent use (more than twice a year) of supplementaries is again an indication of a poorly prepared budget and inadequate budget execution.

The targeted phasing of credit to government throughout the year is typically a crucial component of fiscal (and monetary) policy. To meet such credit targets, the authorities must be able to monitor expenditure aggregates. The ministry of finance therefore needs to ensure that (1) aggregate expenditure data are being obtained in a timely way, (2) the data are accurate and reliable; (3) the ministry of finance has sufficient capacity to analyze and monitor these data, and (4) some unit has been explicitly assigned the responsibility for this task. These conditions are not always satisfied. Thus, improving the quality of the government's information system on expenditures is often essential, so that the ministry of finance can develop the capacity to adjust spending in response to macroeconomic and fiscal developments, while maintaining expenditure control.

At the heart of any fiscal and financial information system is the accounting system, which maintains the basic records of government transactions and, thus, outturn expenditures. How well the accounting system is operating can be judged indirectly by asking the following questions:

• How quickly are final accounts prepared after the end of the fiscal year?

• What are the lags in the reconciliation of above-the-line budget accounts with the below-the-line accounts of the government in the banking system? How often is reconciliation done and at what level of aggregation?

• To what extent are data at different stages of the expenditure process measured and reconciled; that is, commitments and verifications; verifications and payment orders; payment orders prepared and payments issued; and payments issued and payments recorded by the banking system?

• How pervasive are "suspense accounts" and "below-the-line accounts" (see glossary for definition of terms).

• Is there a genuinely independent audit (i.e., an audit body reporting to the legislature rather than the executive); and how quickly does the audit of the annual accounts appear?

Ideally, the government accounting systems are integrated into a fully networked Government Financial Management Information System and can give the necessary information on expenditure developments on-line. But that capability is largely confined to industrial economies and a number of Latin American economies.²⁶ Existing accounting systems in many developing countries were set up to monitor and control the legal compliance of expenditures with authorized budget provisions. As such, they may be poorly suited to provide the financial information required for macroeconomic, fiscal, and financial management. More usually, therefore, the ministry of finance may need to devise fiscal reporting solutions, outside the regular budgetary process, to provide some critical monitoring information, in the form of "flash reports."

Such "flash reports" on the expenditure side are usually prepared by collating and consolidating monthly reports of (and in some systems from) each line ministry, setting out the monthly total payments made. Better developed systems can also report commitments entered into, likely profiles for the associated cash payments, and payment arrears. The basic solution, however, is a fundamental improvement in the budget system's regular reporting capability, typically by longer-run improvements to the accounting system, through the introduction of a double-entry general ledger system, and the development of a Government Financial Management Information System.

Whatever the shortcomings in the information system itself, the capacity of the ministry of finance to consolidate, analyze, and use this outturn expenditure information is also important. Many developing countries have the information to monitor and control expenditure, but fail to make good use of it, because of a lack of administrative capacity in budget analysis.

Some key questions include:

• Is the ministry of finance short of suitably qualified people capable of budget analysis?

• Can the ministry of finance set up and operate a fiscal monitoring unit that could create and manage a monthly reporting system?

• How well does the ministry of finance use the information, if analyses of budget developments are prepared? Is there any kind of feedback mechanism for adjusting policy when an expenditure report shows unexpected, and particularly unwelcome, developments?

Some typical problems identified at the execution stage are the following:

• the multiplication of exceptional procedures that bypass expenditure control arrangements;

• the difficulty in reconciling bank statements with budget accounts and thus in obtaining reliable and timely data on cash expenditures;

• the accumulation of payment arrears;

• the lack of fund consolidation; and

• difficulty in managing and accounting for aid flows.

How should the multiplication of exceptional procedures be dealt with?

In some countries, the fear of corruption, or an unwillingness to share information among the organizations involved in the expenditure management process, have prompted the authorities to add "special" stages of control. The expenditure process can then become so cumbersome that incentives are created to bypass it, through "exceptional procedures," which are supposed to speed up the process for "urgent" expenditures. Many ministries of finance have tolerated (or been forced to tolerate) special accelerated procedures for politically sensitive expenditures.

The use of exceptional procedures is, however, all too often abused and should generally be discouraged. Sometimes, such procedures are used by officials to order supplies, without making a formal commitment; only later do unexpected bills surface. Sometimes suppliers are prevailed upon to make additional supplies available on credit. It is therefore unclear whether the appropriation limit has been respected, or whether there will be cash available to pay for the expenditures. This contributes to the emergence of payment arrears. It can also be very difficult, after the fact, to identify the kind of expenditure on which public monies were spent because regularization--that is, the formal recording of the expenditure in the accounts--occurs much later, if at all. The multiplication of exceptional procedures also has long-term implications, creating incentives for spending agencies to go outside the budget system to avoid control altogether; that is, encourage extrabudgetary activity. Sometimes, indeed, all too often, such procedures are associated with corruption.

Moreover, such procedures are often linked to the misuse of supplementary appropriations. When ministries wish to undertake expenditures on a specific line item beyond the appropriations approved by the parliament (and there is either no scope to switch resources into that line item from another, or they have been denied permission by the ministry of finance to do so), they need to seek parliamentary approval for supplementary appropriations. As noted earlier, if these are issued regularly during that year or after the year has ended--with the expenditures in the meantime financed from, and "frozen" in, a suspense account--budget overspending is being routinely financed.

Excessive use of exceptional procedures and supplementary appropriations should be discouraged, but, as noted, supplementary appropriations are not always bad. First, they are preferable to the buildup of payment arrears, excessive use of suspense accounts, or other "creative accounting" measures to temporarily hide expenditures. Second, they may be necessary to authorize new in-year expenditures that are justified by the fiscal situation.

The fundamental point, however, is that transparency requires all government transactions to go through the budget. Contingency reserves are there for exceptional expenditures. There should never be recourse to "exceptional procedures" or off-budget accounts.

How can the reconciliation of budgetary and banking
data be handled?

Exceptional procedures that bypass standard controls are one reason why it may be difficult to reconcile bank statements on cash expenditures with the budget accounts held on payment orders issued. But there are many other reasons for gaps between the data obtained from the banking system and those reported by the ministry of finance:

• differences in coverage;

• deficiencies in government accounting;

• the problem of arrears; and

• the "float."

Differences in coverage

Often, central government financial statements do not cover special funds and other extrabudgetary operations, while the central bank may identify accounts as belonging to the government in a residual (and, hence global) way--that is, all nonpublic enterprise accounts and nonprivate sector accounts. Those responsible for the budget should ensure that the bank accounts, which the central bank considers to be part of government, cover the same entities that the ministry of finance reports in government operations. Unless the coverage matches, reconciliation is not possible.

Differences in government accounting

Apart from coverage, many countries have trouble maintaining their accounting systems and hiring qualified staff. Long delays and many errors in recording transactions and reporting them are common. As the accounting system is put under stress, accountants take various short-cuts, such as the extensive use of below-the-line accounts and suspense accounts, to record transactions on a temporary basis, before "posting" them to the correct line item in the government accounts or obtaining supplementary appropriations. Unfortunately, many developing countries experience large backlogs in regularizing these accounts, and they become open channels for abuse and malpractice. These practices undermine the reliability of the ministry of finance's accounting data; it is necessary to ascertain how widespread such practices are and discourage their use.

Arrears

Perhaps one of the most common sources of discrepancy between bank accounts and ministry of finance reports arises from misguided attempts to slow down the recording of expenditures at the final stages of the spending process. Faced with a monthly or quarterly financing constraint, countries are often tempted to slow down the payment process, either by delaying the issue of payment orders or the encashing of checks, in order to meet financing ceilings. In accrual accounting terms, once goods or services are verified as delivered, the government has incurred a liability; only in cash accounting terms do such practices have any purpose (albeit misguided). If the above practices are relied on for a prolonged period, the ministries' liabilities (and hence expenditure levels) are not correctly reflected in the bank accounts, owing to the existence of unpaid overdue bills, which represent expenditure arrears. The ministry of finance should therefore carefully compare its reports on bills received and payment orders issued with those on payment orders encashed, as recorded by the central bank or government payment agency.

The "float"

Another source of discrepancy between the ministry of finance reports and bank accounts is the "float" of uncashed checks. While the term "float" can be variously defined, the term is most frequently used to mean the check float; that is, the gap between the amount of checks or payment orders issued by the government and the amount encashed by recipients through the banking system. It is usually a small and predictable percentage of total expenditure transactions, but the "float" needs to be taken into account in reconciling the data from above and below the line.

How can the accumulation of arrears be measured--and avoided?

Payment arrears can arise on any expenditure item, including wages, transfers, and debt servicing. Thus, while they are most commonly found on payments due to the private sector for the supply of goods and services, that is by no means the only item affected. For such goods and services, except in those cases legally disputed, an arrear exists when a bill has been received for services verified as successfully delivered but not paid after what is considered the "acceptable grace period," normally 30­60 days.²⁷ A typical problem encountered in measuring arrears is that accounting systems are unable to determine which bills are already beyond their grace period--for example, when bills received are not carefully noted in either a general ledger or separate bookkeeping record. Often, the amount of arrears can only be defined through a proxy, which is the difference between the amount of bills received (if known) and the amount of bills paid.

In practice, as Box 6 shows, the extent of arrears can be very difficult to establish. The amount of arrears should never be defined as the difference between expenditures committed and payments made, or as was formerly the practice in some transition economies, the gap between appropriations and payments made. Either will lead to an overestimation of the figure. As previously mentioned, not all the committed expenditures lead to an actual delivery of services; and the interval between the commitment and the actual payment often varies considerably from one kind of expenditure to another. An estimate of arrears as the difference between payment orders prepared for issue and payment orders cashed (and an allowance for the "float" of checks issued but not yet cashed) can be realistic--but only where the spending ministries verify the bills and send the corresponding payment orders as soon as they receive those bills. If such discipline is not followed, it is necessary to add also any bills received and due for payment at line-ministry or spending-agency level, for which no payment order has been prepared.

In Latin American countries, arrears (atrasos) in the expenditure process can be estimated, when data are available, as the difference between the verification stage (devengado) and the payment stage (pago). In general, it is difficult to measure arrears owing to either the total lack, or the weak reliability, of this information. In this case, the second best alternative for calculating arrears is to use the payment order (órdenes de pago) prepared as a proxy variable for the verification stage, and compare this with the payment stage (pago).

The Commonwealth system does not offer an easy indicator of arrears. Payment orders are usually prepared and issued by the receiving ministry, after certification that delivery has been made. Spending agencies are supposed to maintain commitment ledgers and to record their requests to the treasury for payment or for payment orders to be issued, against authorized appropriations (sometimes called "vote books"). But the vote books are not always completed reliably, or even at all, as regards the column for entering commitments. A comparison of actual deliveries with payment orders issued is likely to be carried out only in a piecemeal fashion by spending departments. As noted earlier, checks or payment orders may be issued centrally or by the spending ministry to the suppliers. Only in a central system is it relatively straightforward to get a governmentwide comparison of payment orders issued against checks issued, and then only after some delay. To estimate the size of arrears under a decentralized system, reliance has sometimes been placed on data generated at a late stage in the spending process, on the gap between payments authorized and checks cashed. As Box 6 indicates, such a measure can often be inaccurate.

When tackling arrears, the sources of problems need to be established. An accumulation of arrears may arise because:

• the budget provision is unrealistic and line ministries are allowed to commit expenditure within that appropriation (i.e., budget provision), even though there is no cash available to liquidate the expenditure;

• the budget figures are realistic, but the cash plan (and monthly cash limits) associated with the budget are not, or there is no in-year guidance on when expenditures can be committed;

• commitments are not recorded and therefore do not respect the budget ceilings or the timetable defined by the cash plan; or

• the spending ministries do not work efficiently and bills remain unprocessed.

The treatment of payment arrears in the fiscal accounts also poses a number of difficulties. (See Diamond and Schiller for a full explanation.)²⁸

The standard presentation of the deficit in the GFS Manual is, for the present, on a cash basis. This requires expenditure incurred in terms of payment arrears to be shown as a memorandum item. Adding the size of the "float" of uncashed checks and the memorandum item on arrears to the financing data from the monetary survey should facilitate reconciliation with the above-the-line deficit measured from the fiscal data.

But the GFS Manual also rightly states that expenditure is incurred by the government when goods and services are delivered and verified, and the payment is due. A better presentation in fiscal accounts (sometimes termed a modified cash basis) is to show expenditures on a commitment basis, with the difference between the above-the-line deficit total and cash financing identified as "forced financing" (i.e., by unpaid suppliers). The kind of data available on commitments from a French-style system can enable a full presentation. In other countries, such information may not be obtainable, except when a special survey of commitments/arrears has been undertaken.

Another form of presentation is to show the deficit measured on a commitment basis, the net change in arrears, and the corrected overall deficit on a cash basis. But, again, that requires data that are often not available or are very difficult to integrate. In practice, data constraints will often drive fiscal economists to the standard approach or to some broad form of showing expenditures on a commitment basis in preparing fiscal tables.

Besides taking measures to stop the accumulation (flow) of new arrears, those responsible for the budget should be familiar with the size of, and the strategy for resolving, any existing stock of arrears. This stock can be very difficult to estimate, as Box 6 implies. However, in many instances, the only reliable source of information would be an audit of the existing bills, making sure that the goods or services have actually been delivered. (Some suppliers may take advantage of accumulating arrears and the disorganization of the expenditure process by sending bills for goods that were never provided or that have, in fact, already been paid for.) Legitimately disputed arrears should not be included in the stock of outstanding arrears for clearance.

Once the stock is known, it is advisable for some central agency in the ministry of finance, say the treasury, to take over the responsibility for the payment of the stock of past arrears, so that this can be programmed as an identifiable component of the current fiscal year's expenditure. But liquidating a stock of arrears raises a number of issues:

1. If provision is made by line ministries from within their appropriations (typically at a late stage in budget preparation), it leads to reductions in the financial provision for other policies and programs. When paying off the old stock of arrears, it is important to avoid the accumulation of new arrears. Moreover, some line ministries or spending agencies may choose to incur new commitments with their resources rather than meet past liabilities.

2. If provision is made for repayment arrears by the ministry of finance, on the other hand--for example, from within its contingency reserve without reducing the line ministry appropriations--this may ensure the arrears are liquidated. But it generates a "moral hazard" problem. The line ministries may conclude that if they incur arrears in the future they will again be bailed out by the ministry of finance.

3. A third option is to finance the accrued liability by "securitization," a form of government borrowing. There is a range of options, including promissory notes that are discountable by the commercial banks (but should not be rediscounted by the central bank, as that would lead to monetization of the deficit); the issue of treasury bills; or the issue of longer-term bonds. This approach has the advantage of not disrupting the budget provision and may enable the private sector to be repaid more quickly, but it adds to government debt and to the government's interest bill; and it again raises "moral hazard" issues. Moreover, whereas the issue of treasury bills and bonds is often associated with full payment of arrears, promissory notes can be used to prioritize (by varying the terms of the promissory notes) and can only be cashed at a discount. Thus, the creditors face a loss relative to cash payment. Also, the issue of promissory notes can be nontransparent, being open to favoritism and corruption. In no cases should the arrears be directly offset against tax liabilities. This would only undermine tax compliance and encourage the future accumulation of arrears.

4. While option (2) and/or (3) may often be preferred, it should always be accompanied by a genuine strengthening of expenditure controls to avoid future arrears, and thus reduce the moral hazard problem. Its acceptability will also depend on many other wider factors: the case for prompt payment versus deferral; the interaction with any arrears on tax payments;²⁹ the mode of financing; and the pursuit of monetary policy objectives.

In most countries, the government's funds--for example, tax and customs revenues, proceeds from internal and external borrowing, etc.--flow into a single bank account (often called the treasury account) held in the central bank. This is thought not only to facilitate budget execution, particularly cash and debt management, but also to be the safest depository for government funds. But there is no theoretical reason why this must be so; in a few industrial countries, the role of government banker has been contracted out to a single commercial bank. In the vast majority of countries, however, the current practice (or the aim) is for government resources to flow speedily and efficiently into a single account held in the central bank.

Government funds are, however, often fragmented in developing countries. In some countries, the main treasury account coexists with other public accounts, such as the account of the amortization funds in francophone countries, or public entities are allowed to open accounts in private banks. In such cases, government funds may be available as deposits in bank accounts, even while the government is borrowing in the market and/or payment arrears are accumulating in the private sector, because the main treasury account has no cash available. Thus some fiscal tables--which are consolidated reports--can show a buildup of idle cash balances and the accumulation of arrears.

This fragmentation of funds undermines budgetary operations and generally implies additional costs to the government. Often the government is borrowing at high interest rates, while unused funds accumulate in bank accounts and earn a low rate of interest, if any. Commercial banks use the government's own money to lend back to them at a profit.

Such practices should be changed. First, information on these accounts--at least on the aggregate flows and balances in all central and commercial banks belonging to the government--should be available to the ministry of finance.³⁰ (This is necessary to enable the ministry of finance to reconcile aggregate bank accounts with the government's fiscal tables.) The second step is to consolidate as many of these accounts as possible within a single treasury account at the central government. Any accounts that are designed to bypass normal budget procedures should be closed, and expenditure previously financed from these accounts would then become regular budgetary expenditure and follow the normal budget procedures.

How efficient is foreign aid management?

For many developing countries, the degree of their reliance on external foreign assistance has increased over the years. Unfortunately, the growth in aid flows has often not been matched by the administrative capacity to manage them well. This can create several problems for fiscal economists.

Many of the problems concern the timing of aid flows. Commodity aid that enters a country may be recorded at different times by the donor (on shipment), by the central bank (on clearing the port), and by the ministry of finance (when the commodity is sold and money flows into the government bank account). When the aid involves basic commodities that have to be processed, distributed, sold, and the profits remitted back to the government, the delays can be quite substantial. Often the receipts from such commodity aid do not enter the government's main account but go into special "counterpart funds," which are administered jointly by the government and the donor. Sometimes these funds can be released only for donor-approved purposes. This source of financing should be regarded as tied aid rather than as general budget support. In countries dependent on such aid, it is important to understand the timing of these flows; distinguish financing from public expenditure transactions; and record such flows appropriately in government accounts.

For project aid, the reimbursement principle is increasingly employed, whereby the government first spends funds linked to a donor-supported project (expenditure), then claims reimbursement from the donor (financing). There are often large administrative delays in processing these claims, so that spending initially has to be financed from domestic sources. The extent of these delays should be investigated, and, if they are substantial, the executing ministries, in collaboration with the donors, should be encouraged to implement measures to speed up the processing of claims.

Another common practice is for the donor to set up special project accounts, sometimes in the central bank but often in a commercial bank, to which all project disbursements will be channeled, and from which all payments for the project will be made. This makes it administratively easier for the donor to monitor implementation, allows some control over the use of funds, and can be used as a means to regulate the speed of project implementation. However, since these accounts may be kept outside the regular government accounts, and so bypass normal accounting procedures, there are often delays in reporting and problems in consolidating these accounts with the government's main account.

Even more difficult for the accounting system is the practice of donors paying foreign suppliers directly on behalf of the recipient country. In this, and the case of special project accounts, the government is often dependent on information supplied by the donors to monitor the implementation and financing of projects. In many cases the flow of information breaks down or is available only after considerable lags.

Further domestic problems arise because of uncertainty about the timing of receipts of aid flows. On the one hand, there can be the need to make provision for counterpart funds within the budget--which are not used if the aid is not disbursed as planned. On the other, foreign aid can arrive unexpectedly, requiring the authorities to direct money into counterpart expenditures. Often, there is a lag between when foreign resources arrive at the central bank and their crediting to the government's accounts.

For so long as donors require the separate accounting and banking of their grants and loans, there can be no generalized accounting (or financing) solutions. In countries highly dependent on foreign aid, the fiscal economist should make a special effort to understand how commodity and project aid is handled on a case-by-case basis, so that consolidation of all government expenditures can be achieved, so far as possible. It may even be desirable to visit local agencies of the principal donors and discuss with them their main implementation and reporting problems; and, as necessary, to discuss with the authorities how procedures can be improved. The fiscal economist should encourage the government to maintain an up-to-date record both of its external liabilities and the timing of foreign inflows due.

Those executing the budget may be faced with requests to consider additional expenditure in-year; or they may be asked to look at the scope for cutting back spending in-year--for example, when revenue inflows have proved disappointing.

The most important point is that, as noted in the preceding section, only changes in expenditure policies can deliver sustained changes in expenditure levels. Yet, some country authorities can be tempted to seek ways to avoid expenditures being recorded as the first means of "cutting" spending.

First, fiscal economists and budget advisors need to keep in mind the difference between an accrued and a cash liability for public expenditures. As noted above, once goods are delivered and the bill is received and verified, an accrued liability has arisen. Moreover, in economic terms expenditure should be recorded as having taken place. But all too often governments, aware of the paramount importance of the monetary variables in macroeconomic adjustment programs, seek to avoid recording a cash liability. Limiting the amount of cash liability allows the government to hold down (below target, for example) borrowing in cash terms.

Such schemes for "cutting" expenditures should be eschewed as they typically involve:

• building up payment arrears;

• recording expenditures in suspense accounts;

• reductions in cash availability not matched by reduced appropriations and appropriate policy changes to reduce commitments; or

• delayed submission of bills (particularly at the end of the year).

Second, the basic principle must be that, in seeking to introduce expenditure reductions in-year, any attempt to reduce or slow down the rate of spending must be directed toward the commitment stage--that is, before a liability is incurred. It follows that the scope and targeting for such measures is restricted both by timing (for some expenditures such as debt servicing, the commitment and its timing are or should be immutable) and by the critical need to alter expenditure policy on discretionary expenditures, so that a commitment can be avoided. For example, in-year, it is easier to hold up a capital project than to reduce employment quickly so as to reduce the wage bill. In general, the earlier action is taken, the greater the flexibility that exists.

Thus, there are very real constraints on the adjustments to the planned budget that can be made through action on budget execution during the year. Large components of spending--usually termed nondiscretionary--are generally fixed in the short term. These nondiscretionary expenditures include commitments made in prior years that cannot be changed, such as debt service;³¹ spending mandated by other legislation, such as indexation laws, pension and wage legislation, or other social benefits or entitlements; wages and salaries for existing employees (in the short run); and many transfers, such as those to other tiers of government.

Although governments may often claim that all (or virtually all) expenditures are nondiscretionary, in practice, much of the spending on purchases of goods and services is, to some degree, discretionary--and even the size of the wage bill can often be reduced in-year. The degree of budget lock-in--that is, how much expenditure is genuinely nondiscretionary--needs to be assessed to give a reasonable picture of the maximum degree of expenditure reduction that can be imposed in any fiscal year.

Thus, any proposed actions need to be carefully crafted.

1. Changing the level of new commitments must be the target. This requires a conscious decision to change some expenditure policy, even if it is only advancing the start date of a capital project (to raise spending) or postponing a salary increase (to hold down spending).

2. Even if action can be taken, for example, to delay the start of a capital project, there may be economic costs (project delays), financial costs (penalty clauses), and wider repercussions of loss of goodwill when donor-financed projects are scaled back.

3. Within commitments, attention often focuses on capital items as the most obvious targets.

4. One other category--current purchase of goods and services--may be targeted for reductions, but it is often tricky. If a common percentage reduction is proposed, line ministries may leave important bills unpaid (e.g., utilities) or make unrealistic reductions in sensitive items, hoping to embarrass the ministry of finance later into releasing more resources.

5. Also to be eschewed are reductions in intergovernmental transfers. Such action represents passing the problem from one tier of government to another, with aggregate general government borrowing often unaffected.

A fundamental role for those advising on budgetary matters is to support public sector reforms and policies that promote the efficient use of resources and sustainable economic growth. In recent years, a greater awareness of governance issues has focused attention on fostering public sector efficiency, transparency, and accountability.³² A well-functioning budgetary system ensures accountability, in the sense that:

• every action is transparent;³³

• every participant is held accountable;

• every action is properly documented and reported; and

• every action can be subject to independent, professional, and unbiased audit and review.

Define the government sector clearly and comprehensively. This means not only that central government, state/local governments, and government financial and nonfinancial operations should be separately identifiable, but also that the respective roles of the executive, legislative, and judiciary should be clearly defined and widely understood. Also, the true extent of government quasi-fiscal activities and the contingent liabilities of government should be identified and monitored by governments.

View the budget as a complete process. Interference in any one element of the budget process may have repercussions on the system as a whole. Too often, when formulating fiscal adjustment strategies, attention on the expenditure side is focused on the last stage--actual cash payments--where the impact of government spending is manifested in the monetary sector accounts. As noted above, expenditure control needs to take place at a much earlier stage, long before payment orders are prepared and processed.

Adjust spending at the earliest stage possible. Indeed, the aim should be to have an input at the policymaking stage of budget preparation, when existing expenditure policies are confirmed or new policies adopted.

Minimize disruptions to the expenditure process. The budget system fulfills more functions than that of helping ensure macroeconomic stability. Its original function was to ensure compliance with the budget appropriation law, the basis of sound governance. The budget system should also promote other objectives, an efficient allocation of resources among programs, effectiveness of government operations, and efficient financial management of government resources. Care must be taken, therefore, that stabilization objectives are not pursued at the expense of these other objectives. This is often not easy to accomplish. Cash controls, though attractive for securing aggregate control and compliance with the monetary dimension of an adjustment program, tend to bear more heavily on easily controlled elements like operations and maintenance expenditures; they may distort, over time, the patterns of government expenditures and be damaging to longer-run government efficiency. Widespread arrears, often associated with overreliance on cash controls, may lead to dislocations in the private sector (see Section 5).

Respect the budget system's internal and external controls. At each stage of the expenditure process there are controls in place. In periods of fiscal stress there is a tendency to bypass regular budgetary procedures and circumvent controls. Often the result is an increase in corruption. The extent of opportunistic corruption, whereby individuals take advantage of a poor control system, varies but is all too common. Systemic corruption, though perhaps more rare, involves using political power to subvert the control systems. This can have a macroeconomic impact and may be indicative of the authorities' limited commitment to adjustment and reform. But it is not susceptible to any direct or quick solution; and only a strong political commitment will bring needed improvements.

Limit exceptional procedures. In some countries, the expenditure process has become so cumbersome that there are substantial incentives to bypass it, through "exceptional procedures." Procedures should be defined in laws and regulations; inspections and audits should be timely and comprehensive; infractions should be dealt with, and seen to be dealt with through appropriate disciplinary action.

Poor governance can undermine fiscal adjustment. Governance issues affect the ability to deliver fiscal objectives. Countries successful in fiscal reform either started with a sound public expenditure management system or began their effort improving a deficient one.³⁴ Weak governance should be addressed early in the reform effort.