A. The System in France

Public expenditure management obeys numerous rules articulated from two main texts:

• the 1959 Organic Ordinance on Public Finance Acts spells out the rules regarding the elaboration, approval, and execution of the annual budget; and

• the 1962 Decree on the General Regulation of Public Accounting details the rules applicable to the public accounting framework.

Both texts have been supplemented by lower regulations and by court decisions.¹⁰ While parts of the new laws already existed in previous regulations, the legal framework of the early sixties was better adapted to the requirements of a modern democratic environment where fiscal policies play a key role in the economy.¹¹ The main elements of the expenditure process are described in Figure 1.

View Full Size

The Public Expenditure Process for Goods and Services

Citation: IMF Working Papers 2002, 058; 10.5089/9781451847871.001.A001

• Download Figure
• Download figure as PowerPoint slide

The distinction between the person who orders the payment, and the person who handles it is key in the French public finance system. In the French public expenditure system, each line ministry initiates its own expenditure on the basis of its budget allocation. To ensure transparency and avoid abuses and errors, the person who initiates the decision to spend (l’ordonnateur) must, however, be dissociated from the accountant who effectively processes the payment (le comptable). The accountant is protected by a specific statute that ensures his/her independence from the ordonnateur (see Box 1). The split extends beyond the expenditure process and also applies to tax collection: the French system distinguishes between the person who establishes the tax base (la liquidation) and the person who actually collects the money on behalf of the state (le recouvrement). In practice, there are, however, significant nuances:

• on the expenditure side, some substantial categories of expenditure (les dépenses obligatoires) excape the two-phase process and are paid rather with an automatic payment-order system. This is the case, for instance, for pensions and wages. The rationale is that there is no discretionary power to spend or avoid spending and controlling the regularity of payments is sufficient. In addition, expenditure made through special procedures can also escape the two-phase process;

• on the revenue side, the distinction is slightly less significant: the two functions are performed by agents belonging to the same ministry (the ministry of finance), and some key taxes are directly collected on the basis of the taxpayer’s own declaration.

The separation of the administrative and an accounting phase implies distinct functions and agents. The administrative phase, handled by the ordonnateur comprises three steps: the commitment (l’engagement), the verification (la liquidation), and the payment order (l’ordonnancement or le mandatement). The accounting phase, handled by the comptable, is the payment stage, including controls over the regularity of the payment order and recording of the transaction (see Box 2).

Specific feedback is associated with each phase of the expenditure process. The execution of the administrative phase is reported according to the budget classification and will eventually result in a compte administratif while the accounting phase is reported according to the chart of accounts and will eventually result in a compte de gestion. In principle, the information coming from one source checks against the other (this issue is discussed further below in Box 3).

Because of its inherent rigidity, the system allows for alternative procedures, usually when the amounts involved are small. Thus, there are petty-cash procedures (régies d’avances) for which a régisseur assumes both functions at the same time. Other notable exceptions include the so-called special and secret fund procedures occasionally used by the prime minister and certain other ministers. This last procedure allows ministers, on the basis of the initial warrant of appropriation, to execute these outlays outside of normal control procedures (and especially outside the scrutiny of a public accountant).

The French system relies on a twin set of controls. Two control points parallel the two phases in the expenditure process. At the commitment stage, the financial comptroller, an agent of the ministry of finance located in the line ministries, has two roles: (1) checking the regularity of the commitment before the creation of rights and obligations (thereby effectively controlling the regularity of the commitment toward budget appropriation); and (2) keeping a record of the commitments made. The approval of the financial comptroller is required for the ordonnateur to go ahead with an expenditure at the commitment stage. A second approval of the financial comptroller is also required at the payment order stage although the approval is then limited to the previous commitment and the subsequent recording of the discharge. At the payment stage, the accountant exerts an ex ante control over the payment order before proceeding to the payment.

In principle, the twin set of controls guarantees the regularity of the operation. To facilitate reconciliation between commitments and payments, these controls are supplemented by a number of accounting rules: (1) commitments are closed before the end of the year and an additional two months is allowed after the end of the year to process and record cash transactions; and (2) for investment operations and large procurements covering more than a year, commitments are controlled and recorded according to specific autorisations de programme, appropriated for in the budget, even if they do not involve any cash transaction during the year.¹⁴

The presence several control points is also key in terms of tracking and of the division of power within the structure:

• early controls make it possible to have a full record of commitments,¹⁵ and subsequently a comparison to commitments discharged and payments, to have a full assessment of the stock of outstanding commitments and arrears; and

• these controls clearly reinforce the role of the ministry of Finance in the expenditure process through the financial comptroller and the accountant. However, none of these controls is, in principle, a control of opportunity.¹⁶

B. The System in Francophone Africa

The main features of the French system as described in Section A apply largely to francophone African systems. However, while the French treasury can operate efficiently within the complex structure as described, deficiencies in the less developed countries of francophone Africa have created severe bottlenecks. The accumulation of substantial arrears is a symptom of dysfunction.¹⁷ Comparisons in this area are always difficult to interpret because arrears do not result from a single factor. Although at first glance the existence of a twin set of controls does not prevent the accumulation of arrears, the reality may be less gloomy. We do not want to convey the impression that the system in francophone Africa is worse than that prevailing in Anglophone Africa. There are at least two reasons why comparisons are difficult and may be misleading:

• the comparison with systems where commitments are not recorded may be unfair as francophone African accounting systems can provide more reliable and up-to-date information on arrears;¹⁸

• the absence of autorisations de programme makes it difficult to distinguish outstanding commitments that should have been discharged, and outstanding commitments still pending because they relate to multiyear operations. Besides, some countries have difficulty closing operations within the authorized time after the end of the year because it is often shorter than their customary processing time. This results either in outstanding commitments never reconciled with actual payments, or a longer delay in closing the accounts exceeding those legally authorized.

Nevertheless, general audits have revealed acute dysfunction in some countries.¹⁹ Gradual, ad hoc adjustments have eroded the distinction between the administrative and accounting phases, at the expense of budget cross-verification. These adjustments became necessary because the economic environment changed, and trained manpower became scarce. The flow of information was often interrupted as data could not be gathered and compiled in a timely manner. Ad hoc solutions allowed the system to continue to operate. Gradually, these changes were embodied into the system and undermined

• the split between ordonnateurs and comptables;

• the set of controls built into the budget execution process; and

• the accounting records at the level of the treasury.

Some countries have created a centralized payment order scheme, usually taking the form of a specific division within the ministry of finance: the direction des ordonnancements. Although line ministries technically keep the initiative over expenditure, they no longer request the accountant to process and pay the suppliers’ bills. Generally, the directions des ordonnancements have been ill equipped to process payment orders in a timely manner (especially when they centralize all payment orders from the territory, or when they have to regularize delegated authorizations to deconcentrated authorities). This has resulted in delays closing the accounts and inadequate monitoring of the fiscal situation.

The emphasis on cash management in the Fund’s programs has aggravated the situation. The necessity of monitoring closely the impact of cash payments on bank financing and easily reconciling above-and below-the-line data has demanded the creation of instruments such as the tableau des opérations financières de l’état (TOFE) (fiscal tables) and (sometimes) a cash plan. However, substantial delays occur between commitments and actual payments, and some commitments remain outstanding. Therefore, the reconciliation of above-the-line data on a commitment basis with financing data is not straightforward (this is not the same issue as the scope of the TOFE discussed in Appendix I). One must extract commitments for multiyear operations with no good or service delivery during the fiscal year, and split the rest between (a) regular nonbank short-term supplier’s credit; and (b) payment arrears (both as nonbank-fmancing items).²⁰ These additional requirements have contributed to the relative lack of interest in commitment controls at a time when such controls should have been strengthened to enforce hard budget constraints and limit circumvention.

In order to produce the information required for verification and cash-management monitoring, special funds and petty-cash procedures have been generalized. In France, these procedures were meant to provide flexibility²¹ in very specific circumstances, under appropriate controls.²² Nevertheless, their existence should always be perceived as a quasi-legal way to circumvent the normal expenditure control mechanisms. There is a legitimate need for these procedures in francophone African countries: some areas are remote, for instance, and communication is inefficient, justifying the use of petty-cash procedures and “advance mechanisms”²³ (regies d’avance) to increase government efficiency for certain specific expenditures. Unfortunately, these procedures have instead been used by central public agencies with the political clout to impose their own expenditure priorities and timing:

• some countries have used petty-cash procedures as the modus operandi for a line ministry—or for all investment expenditure. In the absence of regularization, these operations end up being accounted for, in the best of circumstances, in suspense accounts (to be discussed later). The capacity to track the expenditure is lost, because it may well have been executed without any budget appropriation, as no commitment or payment order has formally been issued (see also Box 4);

• special funds have sometimes been used either without requesting prior payment order (in which case the special fund is equivalent to an unlimited capacity to draw on general resources, with no controls), or by creating special accounts which usually benefit from earmarked resources (this reduces accountability in the use of resources); and

• similar to the special funds, the ordres de paiement (OP) (see Box 4) are payments requested of the accountant in the normal way, except that they are not accompanied by vouchers. Emergency needs initially justified such orders, but they soon became common in several countries. Gradually, as these OP were not matched with initial appropriations by the financial controller or by the relevant control in the line ministries, the normal expenditure process and the need for budget authorizations were totally circumvented.

These circumventions have weakened any incentive towards efficiency, obscured responsibility, and ultimately undermined monitoring:

• The rupture of the link between the person who orders the goods and services from a supplier, and the one who orders the payment, has led to an avoidance of responsibility within line ministries, an indifference to expenditure prioritization and, conversely, an incapacity in the ministry of finance to properly prioritize payments. These factors have led to wrong incentives, impeding the reconciliation of commitment with payment data, and the proof of effective delivery.

• At the same time, the financial comptroller function has been weakened to the point that sometimes commitments being discharged are no longer recorded. In some cases, this has led to the creation of additional administrative steps by the financial comptroller, usually relying on manual transmission of information (as he is generally located in line ministries). This additional control, or the one performed by the above-mentioned direction des ordonnancements can only serve as an information device since it occurs too late in the expenditure process to block an unwarranted expenditure. Nevertheless, the existence of such payment orders within the ministry of finance itself has overshadowed other existing controls and become the only focus of attention.

Such evolutions, in particular the indiscriminate use of extraordinary procedures, weaken the capacity to maintain reliable records at any stage of the expenditure process. But the situation is particularly critical at the payment stage, that is, at the level of the trésor in francophone African countries. The lack of reliable information also reflects the general weakness of accounting records, a weakness which stems from another set of factors:

• Although insufficient human capacity in accounting²⁴ is not specific to francophone Africa, its impact there has arguably been more severe than elsewhere. The rather sophisticated and detailed French expenditure process with different layers of control is costly in terms of human resources and does not allow for staff fungibles. This trait has sometimes been viewed as corporatist behavior by the treasury officials to prevent other agents from becoming public accountants even if such exclusivity creates redundancies with agents from the tax or customs administration, for instance, and further delays fiscal reporting.

• As treasury balances are not available, monitoring data are computed on the basis of flows rather than changes in stocks. Delays in closing the accounts prevent the compilation of entry balances for the current year since closing balances from the previous year are not available. The general ledger tends to become obscure, and a cumulative process is initiated that blurs any subsequent effort. This also drastically limits the capacity to interpret accounts payable.

• The overdevelopment of suspense accounts has also played a crucial role in weakening the accounting records. Suspense accounts were initially designed to ensure that every transaction got recorded, even if temporarily, until comprehensive information might be available for proper records. However, delays in the circulation of information, the absence of integration between the budget execution phase and accounting records, and the overcentralization of the accounting process, together with the abuse of the specific procedures mentioned above, have substantially increased the size of suspense accounts. The size of these accounts is especially difficult to keep track of since, in the absence of effective closing at the end of the year, suspense accounts are never purged.

The weaknesses of the system have undermined efficient prioritization of expenditures in cash-trapped economies for two main reasons:

• When a budget is cash-constrained, the treasury naturally becomes a prominent center of power although scope for raising financial resources is much more limited than in the French system.²⁵ This situation can cause a difficult balance of power, because the director general of the trésor often chooses who he/she sides with. For example, the director general may be a close relation to the president, who has ultimate control over which checks are effectively paid. By contrast, the minister of finance must then rely on the budget department to monitor the fiscal situation. The twin set of controls described above becomes a power struggle where the treasury usually has the final say. The power struggle may also result in the trésor on the side of the prime minister or the minister of finance, leading to a different, but equally damaging impact on the public expenditure process. In any case, these practices undermine both the budget process and the credibility of the government. They also explain the insurmountable difficulties in creating cash-planning committees, linking prioritization decisions over expenditures to commitments made, and streamlining controls even if they are redundant.

• Officials have often applied circumventing practices to avoid hard-budget constraints, even after the crisis has been resolved. For example, if the actual control point is at the payment order stage (as often in francophone Africa when a direction des ordonnancements has been set up), commitments are made but payment orders never issued (for lack of transmission of the relevant vouchers to the ministry of finance or because the ministry itself prefers holding it back). This gives rise to the DENOs (for example) that plague many francophone African countries (see Box 4). If there remains a control point at the commitment stage, commitments may be entered without recording them properly, giving rise to off-budget expenditures, another common feature of an unraveling public expenditure management system.

The absence of sound accounting records at the treasury means that there is no reconciliation between the two phases of the public expenditure process, the administrative and the accounting phases, and the two corresponding records, the compte administratif (the budget execution record) and the compte de gestion (the cash implication of the budget). The loi de réglement is therefore no longer submitted to parliament (at least within a reasonable time) in many francophone African countries (see Box 3). This document presents the final record of the budget authorization during the year and normally includes as appendices the corresponding compte administratif and compte de gestion. In cases where it is submitted, it is often several years late and the corresponding compte de gestion is missing. An occasional bulk submission of lois de réglement in an effort to catch up on the delays makes effective verification virtually impossible.

• The absence of a loi de réglement itself, whose production is sometimes targeted as a conditionality in Fund programs, is only the tip of the iceberg, and its absence a symptom of the poor recording system rather than a factor of deficient reporting. The incapacity to reconcile the two sets of data further fosters wrong incentives:

• budgetary procedures are circumvented because expenditures made outside the budget, say through the circuit of the treasury, cannot be verified; and

• arrears accumulation cannot be accounted for, because the ultimate measure of arrears (including DENOs and all others) can only be made by comparing the used appropriations of the executed budget (compte administratif) to actual payments (compte de gestion). In fact, all other ways to measure arrears (including those in Fund programs) reflect the absence of such reconciliation and should in principle correspond to a short-term concern.

All these problems may have been aggravated by the monetary arrangement in these countries. Most francophone African countries have no autonomous monetary policy, as the monetary stance is, in most cases, determined by the Banque centrale des Etats d’Afrique de l’Ouest (BCEAO) or the Banque des Etats d’Afrique centrale (BEAC) under an agreement with the French treasury. While preventing inflation through monetization of the fiscal deficit, this arrangement has tightened the cash constraint for the treasury and increased the importance of accurate budget forecasts and public expenditure management control mechanisms. Slippages in the system usually lead to large, inexplicable movements in arrears, in the balance of special or correspondents’ accounts (see next section), and more generally in nonbank financing.

Finally, the legal system does not provide the right incentives. Legal provisions are not always in place in francophone Africa and when they are, the judicial authorities are weak or their enforcement is limited. In France, a strong system of sanctions over both the accountants and the ordonnateurs is included in the legal framework, and is effectively applied. The mise en débet for the accountants, and the gestion de fait for the ordonnateurs are effectively enforced by a specific jurisdiction (the account jurisdiction).²⁶ Sanctions are less effective in francophone African countries. In France, when an accountant is sentenced, he is personally considered to be financially responsible.²⁷ In Africa, his financial responsibility is meaningless as the personal wealth and income of public servants is unlikely to cover the amounts involved in the mishandling of public funds.²⁸ Besides, no insurance mechanism could protect these officials effectively.

A. The System in France

Another salient feature of the French system is the obligation of financial entities related to—or in—the public sector (les correspondants du trésor) to deposit some or all of their money at the trésor²⁹ This systematic centralization of funds has several implications; the main one is that the fiscal deficit can (to some extent) be offset by a change in deposits by public entities. Figure 2 presents a graph summarizing the main elements of the centralization of deposits.³⁰ As a result, the link between the deficit and recourse to financing weakens (Bloch Laine, 1955, was among the first to draw attention to this phenomenon).

View Full Size

The Circuit du Trésor

Citation: IMF Working Papers 2002, 058; 10.5089/9781451847871.001.A001

1/ Spending and deposits outside the circuit du trésor are leaks in the circuit and deplete its resources. On the other hand, other resources collected by the correspondants du trésor outside the circuit du trésor add to the liquidity of what is sometimes called treasury money.

• Download Figure
• Download figure as PowerPoint slide

For example, formerly in France, public servants were paid through a deposit on their postal accounts, and the postal bank, in turn, was forced to keep their deposits at the trésor³¹ Thus, when salaries were paid to civil servants, the whole operation would essentially boil down to a transfer from one treasury account to another. The operation had little impact, if any, on the need for liquidity, in spite of the increase in the fiscal deficit. Today, in France, civil servants are paid directly into their private bank accounts, but the earlier system still operates in most of francophone Africa.

The circuit du trésor has gradually narrowed, but it remains an important feature of the French system. This source of financing was important in the sixties and seventies, and represented a substantial source of autonomous nonbank financing.³² Although centralized deposits shrunk as a source of financing during the eighties and nineties along with financial deregulation and increased provision of banking services, the stock of deposits by correspondants still amounted to a hefty FF 405 billion in 1998 (FF 181 billion for the postal bank) for a total budget of FF 1669 billion voted by the French parliament.

This centralization of funds is in a sense equivalent to a single treasury account extending to the public sector at large. As a result of the involvement of the state in the economy, especially through the publicly-owned financial institutions, postal bank and national savings institutions (caisses d’épargne), assets of private entities and individuals ended up being deposited in the circuit du tresor and automatically contributed to the financing of the government deficit. This extensiveness, however, differentiates the French circuit du trésor from the general concept of a single treasury account, as entities whose day-to-day operations are not under the direct control of the central government participate in the general cash pooling and may, at times, act contrary to the best interest of the trésor in terms of cash management and planning.³³

The main problems with such an extensive centralization are the deceptive ease in financing a deficit, distortions in the economy, and a general a lack of accountability and transparency. France has suffered at times from some of these weaknesses, but not overwhelmingly so. The benefits include easier and cheaper liquidity management and a lower recourse to domestic financing than would otherwise be required. However, the reverse applies when a public company withdraws its deposits from its account at the trésor, and the state requires additional financing for an action that is not fully under its control. We shall show in the next section that this has often happened in francophone Africa and has led to numerous difficulties. Overall, the system presents a number of benefits, which may well outweigh its costs, as long as it is well organized and run according to strict accounting rules. This is considered the case in France.

B. The System in Francophone Africa

In francophone Africa, the circuit du trésor operated well up to the oil crisis of 1975. Up to the early 1970s, the budgetary systems were relatively small in francophone Africa, as they had limited need for financing: deposits corresponded to their intended purposes, that is, to stabilize financial requirements, and to force public entities to operate within a common cash-management framework.

With the oil crisis, countries that were not oil producers, started using these deposits to absorb the (structural) shock. De facto, the treasury system had gone bankrupt by 1975. Instead of acting as a tool for improving cash management and easing short-term financing constraints, the circuit du trésor became a vector by which cash tensions and illiquidity were permanently transferred from the central government to other public entities. This resulted in severe imbalances and all over the region; large public financial establishments went bankrupt, including postal banks and national savings banks.

The crisis forced numerous adjustments in treasury systems after 1975; they formally maintained the old system while undermining its foundations and involuntarily creating incentives for mismanagement. Three major changes emerged at that time:

• debt management was partially delegated to separate entities, the caisses autonomes d’amortissement (see Box 5);

  Box 5.

  The “Caisses autonomes d’amortissement”³⁴

  The caisses autonomes d ‘amortissement (CAAs) have existed in the French system since 1926. They were designed to bolster public bonds owners’ confidence by earmarking specific resources, e.g. tobacco or oil product duties originating from public monopolies to debt service. They gradually died out, but the concept was revived in 1986 with the creation of a new type of caisse d’amortissement where privatization proceeds were deposited.

  CAAs have also been a common in francophone Africa. Based on the principle that earmarking increases confidence, their use has changed with time:

  • they have sometimes been the recipients of foreign grants and loans (and not only of domestic taxes);

  • they have tended to operate as a separate financial entity from the treasury, and have sometimes totally taken over responsibility for public debt management.

  In general, CAAs were part of the circuit du trésor and their funds were therefore deposited at the trésor.³⁵ However, they gradually evolved into a full banking function, becoming progressively involved in the financing of the economy, but outside of the general budget management framework. Not surprisingly in a context of weak transparency and accountability, CAAs have been hit hard by economic difficulties and, rather than providing a source of financing, have become a source of imbalance in the economy as well as a liability for the central government.

• new budget management tools were developed to mitigate the illiquidity of the trésor. In particular, large numbers of special accounts and special funds were created, often linked to extrabudgetary activities and not necessarily located within the trésor; and

• more generally, short-term cash management of very scarce resources became a primary focus of financial policies and this aggravated the pressure on ill-equipped administrations.

Special accounts have developed in the context of cash tensions and relative illiquidity. Generally, rules pertaining to special accounts were designed to circumvent those of basic budget management, such as requiring an annual timeframe or avoiding earmarking. At best, special accounts were seen as a way of protecting specific policy objectives from the general illiquidity of the central government (by earmarking specific resources). At worst, they helped vested interests to finance their own purposes and protect themselves from the general illiquidity.

The cash crisis at the trésor was exacerbated by the fact that correspondants were allowed to maintain their own bank accounts in parallel to the account at the trésor. In the French model, correspondants are not administrative entities. They do not have specific bank accounts, and are integrated in data for financing the central government budget even if they can be identified separately, In francophone Africa, correspondants were allowed to circumvent the cash constraint created by the illiquidity of the trésor by owning their own separate bank accounts (an illustration of the accounting treatment of such movements in the TOFE is provided in Appendix I). Other public entities, such as universities and hospitals, also opened bank accounts to limit the risk of a sudden cash shortage. This exacerbated the cash constraint:

• even in cases where these bank accounts were limited, there was a leak in the circuit at a time of cash tensions in the public sector;

• this leak created a threat that massive transfers to privately owned banks and other financial institutions outside the circuit could suddenly deplete the circuit. Not only would such movements of funds be difficult to monitor or predict, but there was also a risk that these funds would be used for purposes that should not receive a high priority during a cash crisis;³⁶ and

• the existence of many bank accounts outside the control of the trésor fostered malpractices and fraud.

The need to address the crisis, often in the context of Fund programs, created a narrow focus on short-term cash management which was ill-suited to the characteristics of the francophone African system. First, the focus on short-term cash management distracted attention from structural problems whose eventual impact on the treasury position was sometimes quite substantial. Second, while negative positions of the correspondants of the trésor were prohibited by the law, the absence of supervision, the poor state of accounting records, and a generally weak enforcement of legal requirements led to automatic subsidies, through treasury advances, to many public institutions.³⁷ These implicit subsidies were not necessarily reflected in the fiscal reporting tables of the central government and, in any case, would not have appeared as expenditure items, undermining cash-management monitoring. As the issue of correspondants was rarely addressed at first in treasury planning, they started contributing to the deficit and/or massive errors and omissions appeared.

This problem was exacerbated by the focus on net credit to government (theposition nette du gouvernement or PNG) in the Fund’s programs. The focus on bank financing in Fund programs stems from the risk of macroeconomic instability, in particular of high inflation. Generally, central bank data are readily available and account for a major part of bank financing. Commercial bank data are also available as there are often only a few major commercial banks operating domestically. For all the weaknesses they may have, bank-financing data are more precise than the rest of the fiscal table, making them a privileged item for monitoring purposes. In countries where there is a regional arrangement, like the WAEMU, the strict limits imposed on recourse that can be made to the central bank constitute another advantage, as they cannot easily be circumvented. In fact, the central bank has the ability to capture incoming money if the position of the trésor is not satisfactory.³⁸ The focus on the PNG can, however, have undesirable implications (such as the ballooning of nonbank financing) and generates a lot of debate with regard to its scope (especially because consolidation above-and below-the-line can be different—see Appendix I).

A major variable in the fiscal table in some francophone African countries is nonbank financing, which includes the counterpart of hard-to-monitor treasury operations that have often ballooned considerably, sometimes to the point of generating instability.³⁹ Overall, bank financing is only one of the financing items and should not be overemphasized in the francophone system.

While clearly necessary, the remedies to the crisis have often aggravated the cash situation at the trésor in the short term. Efforts made to restore budgetary management orthodoxy through adjustment programs have led to the privatization of the most valuable correspondants, and to deregulation, especially in the financial sector. This has reduced the number of public entities able to play a substantial role as correspondants du trésor. The remaining ones are inactive or in financial disarray.⁴⁰ Thus, remaining correspondants tended to contribute to the financing need, implying additional cash tensions at the trésor without room for dilution in the rest of the public sector. Moreover, the trésor has usually been required to pay an interest on deposits kept by the correspondants that kept positive balances. Although this necessary consolidation phase is the direct consequence of previous errancy, the cash burden it creates, together with the weak capacity of the trésor to monitor and record financial transactions, has contributed to the emergence and development of payment arrears and still affects the overall financial situation in many countries.

The existence of a circuit du trésor also complicates the relations between the central and local governments, and alternative financial arrangements need to be implemented prior to further decentralization. Although most francophone African countries are very centralized, local governments have always been an institutional feature, even if their financial autonomy is limited. Local governments are usually correspondants du trésor. While they should in principle never run a negative balance, experience shows that this rule has often been overlooked, and local authorities (especially the major cities) have been mainly responsible for the deficit of correspondant and special accounts. The move toward a more decentralized environment has consequences in terms of financial organization that have rarely been addressed:

• The gradual decline in the size and scope of the circuit du trésor could reverse as local entities are asked to manage additional resources and deposit them in their respective correspondants’ accounts. Unless there is a strong improvement in local management records, decentralization could even lead to more deficits and increased cash pressure on the trésor.

• Increased autonomy and better financial management capacity may encourage local authorities, especially the largest entities, to deposit their funds elsewhere. On one hand, this could lead to an improved allocation of resources but, on the other hand, it will hamper monitoring at the central level while the trésor has to make significant transfers to the local entities with a positive balance. This could, at least in the short-term, have a significant impact on the central government fiscal balance.

• In the context of the HIPC Initiative, because many local governments will become service providers during the fight against poverty, the impact on the balance of treasury accounts is uncertain: an increased demand for accountability may strengthen local governments’ financial autonomy, while at the same time the need to track poverty-related spending, and the creation of poverty funds and HIPC accounts could limit local governments’ autonomy.