Mali: Technical Assistance Report - Implementing Fiscal Decentralization

This Technical Assistance Report discusses key recommendations for implementing fiscal decentralization in Mali. It proposes to set an optimal administrative organization of the territory, which should aim at building consistent and sustainable local governments that are best suited to address specific local needs. In this context, the regions deemed to be best suited to handle funds are envisaged to be decentralized and with large-scale development actions. The local government revenue should be optimized through local taxation, and an efficient system of transfers from the central government should be introduced. The report also proposes establishing good financial governance in Mali.


This Technical Assistance Report discusses key recommendations for implementing fiscal decentralization in Mali. It proposes to set an optimal administrative organization of the territory, which should aim at building consistent and sustainable local governments that are best suited to address specific local needs. In this context, the regions deemed to be best suited to handle funds are envisaged to be decentralized and with large-scale development actions. The local government revenue should be optimized through local taxation, and an efficient system of transfers from the central government should be introduced. The report also proposes establishing good financial governance in Mali.

I. Review of Decentralization in Mali

1. Decentralization is a method of administrative and institutional organization under which the government transfers some of its powers to local authorities which autonomy is granted by law. It promotes economic development and social progress, and generally pursues several political, economic and social objectives: (i) to protect unity and diversity; (ii) to optimize economic development in sectors where management at the community level is a source of efficiency; and (iii) to bring government closer to taxpayers so as to manage communities in the best interests of individuals. Decentralization is characterized by autonomy in three areas:

  • Material: the community manages its own assets and dealings.

  • Organizational autonomy: the affairs of the community are managed by its own bodies, which are accountable for the actions they take.

  • Functional autonomy: the community freely manages its own affairs in the context of the law and under the oversight of the central government.

A. Institutional Anchoring of Decentralization in Mali

2. Decentralization is anchored in the legal framework. The commitment to decentralization, which was expressed from the time of independence, was enshrined in the 1992 Constitution, which established freedom of administration of local governments as a principle.3 Since then, decentralization has been considered a safeguard for national unity and territorial integrity. The institutional framework is thus enshrined in the legislation and has been updated from time to time (Box 1). The 2014 administrative framework4 comprises of: 703 communes (e.g. villages in rural areas and townships in urban areas), 49 counties (cerdes), 8 regions (régions) and the District of Bamako, which is made up of 6 communes. All of these territorial units are set up as local governments (collectivités territoriales). Communal councilors, from among which the mayor is appointed, are directly elected by universal suffrage, and regional and county councilors are elected by electoral colleges from the levels of government immediately below (municipalities elect county councils, which in turn elect regional councils).

The Legal Framework for Decentralization

  • - Law No. 93-008 of February 11, 1993 establishing the Conditions for the Freedom of Administration of the Local Governments, as amended by Law No. 2012-005 of January 23, 2012;

  • - Law No. 2012-006 of January 23, 2012 establishing the Fundamental Principles for the Administrative Organization of the Territory (to be applied gradually through to 2018);

  • - Law No. 95-022 of March 20, 1995 establishing the Status of Local Government Officials;

  • - Law No. 95-034 of April 12, 1996 establishing the Local Governments Code, repealed and replaced by Law No. 2012-007 of February 7, 2012;

  • - Law No. 2011-036 of July 15, 2011 on the Tax Resources of the Communes, Counties and Regions;

  • - Law No. 96-058 of October 16, 1996 establishing the Tax Resources of the District of Bamako and its Constituent Communes;

  • - Law No. 2011-034 of July 15, 2011 establishing the Property Tax;

  • - Decree No. 95-2010/P-RM of May 30, 1995 establishing the Conditions for the Appointment and Powers of Representatives of the Central Government at the Local Government Level;

  • - Decree No. 02-313 P/RM of April 4, 2002, Decree No. 02-314 P/RM of April 4, 2002, and Decree No. 02-315 P/RM of April 4, 2002 establishing the Details of the Responsibilities Transferred by the Central Government to the Local Governments in the areas of Education, Health and Rural Water Management;

  • - Decree No. 2012-062 [sic]/P-RM of February 2012 establishing the Details of the Responsibilities Transferred by the Central Government to the Local Governments in the areas of Social Development, Social Protection and Economic Solidarity [sic];

  • - Decree No. 2014-572/P-RM of July 22, 2014 establishing the Details of the Responsibilities Transferred by the Central Government to the Local Governments in the areas of Sanitation and Pollution and Nuisance Control;

  • - Decree No. 2014-0791/P-RM of October 14, 2014 establishing the Details of the Responsibilities Transferred by the Central Government to the Local Governments in the area of Commerce;

  • - Decree No. 2012-082 establishing the Details of the Responsibilities Transferred by the Central Government to the Local Governments in the areas of Social Development, Social Protection and Economic Solidarity.

Chart 1:
Chart 1:

Administrative Map of Mali

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

3. The local governments participate in the governmental and legislative processes on a consultative basis. The 1992 Consultation established a unicameral system (one National Assembly), but also established a Local Government High Council made up of representatives of the local governments. Officially established in 2002, the High Council is consulted by the government on all issues relating to local and regional development policy, protection of the environment, and improvement of the quality of life of citizens in local communities. Although its opinion is only advisory, the High Council can make proposals to the government on its own initiative on any issue concerning the protection of the environment and improvement of the quality of life of citizens in local communities. The government is then required to present a draft law to the office of the National Assembly within 15 days of this referral by the High Council. Moreover, local elected officials are grouped into associations, whose purpose is to defend the interests of the local communities (associations of mayors, counties and regions of Mali, respectively).

4. Decentralization is a lever of local democracy, development and poverty reduction. Historically, decentralization has had two objectives in Mali: to expand the process of democratization and participation of the population in the management of public affairs and to promote local development by favoring the initiatives of the various civil society players and the creation of a framework controlled by the population organized where it lives. Thus, in February 2005, Mali adopted a National Decentralization Policy Paper, which defines the decentralization challenges for the period 2005-2014 and establishes a close link between decentralization and poverty reduction. The two central mechanisms for implementation of the policy paper are the National Local Government Support Program (PNACT) and the Institutional Development Program (PDI), which deal with decentralization and deconcentration5, respectively. More recently, in 2013, the poverty reduction strategy paper, which was re-updated by the transition government, affirms the principle of the expansion of decentralization and the transfer of additional resources to the local governments. On the basis of all of these strategy papers, operational objectives, which are the guidelines for the decentralization policy in Mali, have been developed: (i) national unity and territorial integrity; (ii) the freedom of administration of local governments; (iii) the respect of local characteristics; (iv) democratic management at the local level; (v) subsidiarity;6 and (vi) strengthening of contract management by the local governments.

B. Tight Control over Local Governments’ Freedom of Administration

5. The local governments in Mali are entitled to across-the-board general and specific responsibilities. The law7 assigns them a preeminent across-the-board role in economic and social development and the management of local public services. Their specific powers are established by the Local Government Code (Table 1).

Table 1:

Responsibilities of the Local Governments

article image
Source: Law No. 2012–007 of February 7, 2012 establishing the Local Governments Code, Article 22 (communes), Article 97 (counties), and Article 164 (regions).

6. The draft Agreement for Peace and Reconciliation in Mali resulting from the Algiers process (February 25, 2015 version) protects the preeminent role of the regions in the areas of economic, social and cultural development by giving them a new planning role. It also grants them broader powers in the areas of local development, agriculture, livestock, fisheries, forestry management, transport, commerce, industry, crafts, tourism and interregional transport, as well as in the creation and management of basic local equipment and social services (basic education and vocational training, health, environment, culture, road and communications infrastructure at the regional level, energy, water and sanitation). Finally it recognizes their capacity to introduce and apply their own taxes and revenues, within limitations set by the central government.

7. These powers are exercised under the oversight of the central government. Oversight of the local governments is provided by the central government and its representatives at the local level8 (Table 2). It consists of a twofold function of assistance/advice and legality check (e.g. compliance of local governments’ acts with the legislation). In principle, the assistance/advice is provided at the request of a local government, but it can in practice be provided on the initiative of the oversight authority.

Table 2:

Organization of the Oversight of the Local Governments

article image
Source: Ministry of Territorial Administration and Local Governments, General Directorate of Local Governments, “50 questions et réponses sur la décentralisation au Mali” [50 Questions and Answers on Decentralization in Mali], April 2010.

8. Technical support structures for the local governments are proliferating. At the central level, units responsible for the design, coordination and implementation of decentralization have been set up under the coordination of the Ministry of Territorial Administration and Local Governments. Also at the central level, decentralization and deconcentration support units (CADD) were established in 2008 in each ministerial department The achievements of these support units are slight with the notable exception of health and education, where, under their leadership, significant transfers of personnel and the related salaries have been made from the central government to the local governments. The representative of the central government at the local level provides a support and advisory function and the deconcentrated units (services déconcentrés - central government units located in regions, counties and communes) provide technical support to the local governments. Moreover, liaison and coordination officials have been put in place to monitor and assess the technical support and also to coordinate the development actions at the various levels: the National Local Government Technical Support Steering Committee (CNO); the Regional Development Action Steering, Coordination and Monitoring Committees (CROCSADs); the Local Development Action Steering, Coordination and Monitoring Committees (CLOCSADs); and the Communal Development Action Steering, Coordination [and Monitoring] Committees (CCO[C]SADs). A permanent mechanism for financing technical support has been established via the technical support allocation (DAT) in the context of the National Local Government Support Fund (FNACT) implemented by the National Local Government Investment Agency (ANICT).

9. A number of factors limit the implementation of local governments’ policies. These limitations have been observed in the field and recognized by local and central governments’ representatives. Limitations are mainly legal and financial/technical in nature.

  • Legal limitations. The existing legal framework that should spell out the details on the responsibilities and resources transferred to the local governments is incomplete. The only sectors covered by the framework are: education, health and water management in 2002, social development and solidarity in 2012, gender equality policy in 2014 (as of yet the latter two executive orders have not been implemented), and social development, sanitation and commerce in the summer of 2014. As well, regulations ruling the establishment and management of local government owned estate have not been prepared. Finally, the central government continues to implement externally financed projects (in the areas of health and water management in particular).

  • Technical/financial limitations. Central government transfers still constitute the main funding for local governments in Mali, while local governments’ own revenue is weak compared with their financing needs to cover current and capital expenditures. Technical assets and human resources have been limitedly transferred to local governments; and the support they receive from the central government is hampered by the insufficient degree of deconcentration of central agencies.

10. Optimization of local resources is limited. Local governments’ policies are sometimes ill-suited to the local and cultural needs-which can lead to tax avoidance, and further reduce the local governments’ own resources, while central government transfers remain limited and sometimes unpredictable. As a result, resource scarcity is further amplified and the local budgets scarcely cover the needs of local communities.

C. Insufficient Decentralized Resources

11. Incomplete fiscal decentralization drives to poor local budget structures. The mission compiled the financial data available at the National Treasury department into a local government fiscal statistics table (Table 3). The purpose of this table is to assess the structure of revenue, expenditure and financing for all local governments combined. The main lessons to be learned from this table are analyzed below with regards to resources, expenditures and net borrowing.

Table 3:

Table of Local Government Fiscal Operations

article image

Limited resources and a high degree of dependence on the central government

12. Local government resources are in principle diversified. By law9 they consist of: tax resources (central government taxes transferred to the local governments, direct and indirect local government taxes and levies); various types of revenue (agricultural products, financial revenues, income from government lands, fees); fiscal resources, consisting of allocations and special subsidies from the central government to the local governments; authorized loans, which are intended exclusively for the financing of investment; and other resources, particularly grants from external partners.

13. In practice, central government transfers make up the largest proportion of resources. As illustrated by Chart 2, the share of own resources in the operating and capital budgets of the local governments is very limited. Central government budgetary allocations still make up the largest proportion of resources. Own resources of the local governments, although trending slightly upward, remain very meager. Since 2007, local government own revenues (particularly nontax revenues) have increased significantly but remain limited and insufficient to cover the operating expenses of local governments.

Chart 2:
Chart 2:

Operating and Capital Revenues of the Local Governments

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Source: DNTCP, year-end treasury accounts of the local governments aggregated at the regional level for fiscal year 2012 (fiscal year 2010 for the northern regions).

14. Local government own resources are limited. Their own resources are broken down into extremely diverse categories of revenue (tax and nontax). However, the corollary of that diversity is that revenue levels are very low. On the tax revenue side, there are no less than 21 different taxes and levies (Chart 3), of which 3 (the business license tax, the local development tax, and the tax on wages and salaries) represent 77.3 percent of the total revenues collected in 2012, while 12 others10 represent just 3.4 percent of this total. The findings are similar for nontax revenues (revenues from land services and miscellaneous sales). More than 80 percent of nontax revenues collected by the local governments related to 13 categories of revenue in 2012, while 30 other categories represented 1 percent or less of total nontax revenues. This fragmentation is a source of complexity and bureaucracy in the collection of revenues and helps to explain the low productivity of local governments’ own resources.

Chart 3:
Chart 3:

Own Resources of the Local Governments

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Source: DNTCP, year-end treasury accounts of the local governments aggregated at the regional level for 2012 (2011 and 2012 amounts do not include the Northern regions).

15. The local taxation system is cumbersome. There is some disorganization in the process for the assessment and collection of local taxes and levies. The weaknesses mentioned most often by the staff interviewed are poor management of the tax rolls, insufficient coordination between units, insufficient material, human and financial resources in the assessment and collection units, contradictory role11 of elected officials in the collection of local taxes, insufficient motivation of officials, and tax avoidance.

16. The performance of the local taxation system is weak. Total revenues (own resources and transfers from the central government) of the local governments were estimated at 12 percent of the central government budget (excluding grants) and just 2 percent of GDP during the period 2011–13 (Chart 4). This revenue structure reveals two major weaknesses: a high level of dependence of the local governments on transfers from the central government and the limited credibility of revenue projections.

  • A high level of dependence on transfers from the central government: The revenues of the local governments consist largely of transfers and grants from the central government (1.3 percent and 7.2 percent of GDP and the central government budget, respectively).

  • The limited credibility of revenue projections: Local government budget revenue, and consequently expenditure execution, averaged only 48.2 percent during the period 2011–13 (Table 4).

Chart 4.
Chart 4.

Share of Local Resources (% GDP and % Budget12)

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Table 4:

Budget Execution Rates 2011–13

article image
Source: DNTCP, year-end treasury accounts of the local governments aggregated at the regional level for fiscal year 2012 (2011 and 2012 amounts do not include the Northern regions).

Expenditure with a limited focus on local development.

17. The local governments spend current expenditures mainly. In fiscal year 2012, the expenditures of the local governments totaled approximately XOF 95 billion (excluding the Northern regions, for which data are not available), or 1.8 percent of GDP. As illustrated by Chart 5, expenditure executed by the local governments comprises mainly current expenditures, among which payroll expenditures represent between 60 and 80 percent of the total, depending on the region. By comparison, capital expenditures by the local governments are small (see paragraph 18).

Chart 5:
Chart 5:

Current Revenue and Expenditures of the Local Governments

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Source: DNTCP, year-end treasury accounts aggregated at the regional level for fiscal year 2012 (fiscal year 2010 for the Northern Gao, Kidal and Timbuktu regions).

18. The central government remains predominant in the execution of expenditures at the local level. On the one hand, the central government finances most of the expenditures of the local governments by means of transfers and budgetary allocations paid in compensation for the transferred responsibilities. On the other, the central government itself executes a significant proportion of expenditures at the local level through deconcentrated units and central government policies (particularly multiregional projects).

19. Local budgets for investment are extremely small. Capital expenditures totaled a little over XOF 17 billion in 2012 (excluding the Northern regions), which is consistent with the very limited resources available. With considerable variation by region, capital expenditures executed by the local governments in 2012 (2010 for the Northern regions) essentially consisted of construction, infrastructure works, and equipment purchases (Chart 6). This emphasis on operating expenditure to the detriment of capital expenditure may be the cause but also the consequence of underdeveloped local taxation – the low level of local tax resources leaves little latitude for investment and the very low level of investments hinders the promotion of regional development, which would generate additional resources.

Chart 6:
Chart 6:

Relative Shares of Current and Capital Expenditures

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Source: DNTCP, year-end treasury accounts of the local governments.

20. Local governments are at risks with, yet to be quantified, financial debt. Treasury accountants do not report local government debt, although this possibility does exist. The accounts of the local governments (Table 3) record recurring deficits, and local governments can occasionally borrow (against investment projects), build up payments or wage arrears (which is accounted as a debt), and even execute extra-budgetary expenditures against informal loans, outside the purview of the National Treasury.

II. Appropriate Scope and Objective of Financial Decentralization

21. Since 2012, the topic of decentralization has reemerged in the political dialogue. In addition to the institutional, political and social crisis, 2012 was a year of armed secessionist movements and bloody conflicts in the North. There are at least two analyses of the causes of and solutions to the crisis in Northern Mali. One, an external analysis, considers the escalation in hostilities to be the result of the destabilization of the Sahel zone by armed groups and/or organized criminals and the fragility of the administrative system in Northern Mali. The other, a domestic analysis, considers lingering underdevelopment of Mali’s hinterland as fertile ground for identity-based claims and material demands, with these demands taking a violent turn in the North. In both cases, the expansion of decentralization appears to be a decisive factor in re-stabilization through addressing local characteristics, introducing local democratic processes, and relaunching a development dynamic. In 2013–14, an Inter-Malian Inclusive Dialogue (held in Algiers) has started with the aim of restoring peace in Mali’s North. Expanded decentralization and an increase in local resources have been included as essential elements in the talks under way.

22. This new step in the decentralization process builds on the regions while boosting the communes. Historically, decentralization began with an initial phase of “communalization,” accompanied by economic and social development focusing primarily on the communes. The counties and regions, the councils of which are elected by communal electoral colleges, remain characterized by a preponderance of deconcentrated units (services déconcentrés - central government units that are physically located at the local level), as opposed to decentralized units (services décentralisés - units that handle functions devolved to local authorities. The new phases announced by the government (Box 2) aim at strengthening the regions and the consolidation of the communes. The number of regions will thereby increase while councils elected via universal suffrage will take over management. This approach may possibly go to the detriment of the counties, which may disappear as local levels of government13 This decision by the government to concentrate the new phase of decentralization makes sense for the following reasons:

  • The regions are in a better position to handle large-scale projects in the area of enhanced local development and the construction of transportation and water networks and infrastructure.

  • The region is the best level for seeking economies of scale and organizing better revenue collection.

  • The region covers a socioeconomic and cultural framework consistent with the needs of the population.

Regionalization: a New Phase in Decentralization in Mali

Mali, May 2013: Poverty Reduction and Growth Strategy Paper

“A significant and gradual increase in the share of public resources managed by local governments (proposed at 30 percent of resources […] compared to 3 percent today) is a major medium-term challenge. […] This increase will require improving local taxation (broadening the tax base, transferring new taxes and improving the operation of the tax process), increasing central government allocations (especially through a systematic transfer of a share of VAT collected, the application of provisions relating to the offsetting of tax expenditures and losses, value-for-performance contracts, etc.), and strengthening of the capacities of local governments to mobilize external financial resources.”

Mali, October 2013, Participatory Forum on Decentralization (extracts)

The promotion of regional and local economic development involves increasing “the targeting of public policies to local characteristics and land use planning, based on regionalization, […] so that it plays its role in aligning local and regional development programs with national strategies”; “development centers that harness key resources” will be created and “central government-region value-for-performance contracts will be implemented”; finally, there are plans to “expand the tax base of the local governments and improve the collection of taxes and levies; remove all obstacles to the effective transfer of the financial resources linked to the responsibilities transferred from the central government to the local governments; significantly increase the budgetary allocations to local governments; and create an emergency fund for the benefit of local governments located in areas affected by a long absence of the central government.”

Mali, November 2014, Inter-Malian Inclusive Dialogue, progress report

The government undertakes to establish a mechanism for the transfer of 33 percent of central government fiscal revenues to the local governments by 2017 on the basis of an equalization system, with particular attention to the Northern regions, using criteria to be determined.

Mali, February 2015, Agreement for Peace and Reconciliation in Mali resulting from the Algiers process

“The definitive settlement of the conflict requires governance that takes account of local characteristics and involves the establishment of an institutional architecture based on local governments provided with bodies elected by universal suffrage and expanded powers, and management, by the population of the regions concerned, of their own affairs on the basis of the principle of freedom of administration.” The draft agreement calls for the regions to be provided with a Regional Assembly elected by direct universal suffrage, benefiting from a very broad transfer of responsibilities and resources and having appropriate administrative and financial powers. In addition to the setting of rates for local taxes, fees and levies, each region may introduce taxes appropriate to its own economic structure and development objectives in the context of the law. The central government also undertakes to establish a mechanism by 2018 for the transfer of 30 percent of central government fiscal revenues to the local governments on the basis of an equalization system, with particular attention to the Northern regions. Finally it is planned that the central government will transfer to local governments concerned a percentage of the revenues from the exploitation of natural resources, particularly minerals, on their territory, in accordance with criteria to be defined jointly.

Sources: government, press, documents sent to IMF staff.

23. Decentralizing one third of fiscal revenue is a proactive, yet ambiguous, objective. Based on the various statements (Box 2), it is still difficult to determine the exact scope of the planned transfers: 30 or 33 percent of central government revenue and grants, or one third of budgetary revenue (including budgetary support or not)? Government officials interviewed by the mission were not able to clarify the exact objective pursued by the government, or its rationale. The above-mentioned aggregates and their ambiguity apparently stem more from negotiation talks than from a financial prior evaluation.

24. The following subsections will assess the fiscal decentralization goal using three quantitative approaches (Table 5) based on the following:

  • A. Local government financial needs to fulfill their responsibilities: This involves identifying the costs currently lying on the central government to fulfill local policies local governments have yet to take over.

  • B. Local government capacity to generate revenue: This approach’s rationale is that local governments should benefit from, and further develop, resources available on their territory, to trigger a positive economic dynamic.

  • C. Local development financial threshold: This involves using a simple macroeconomic model to identify the threshold as of which fiscal decentralization begins to have positive outcomes.

Table 5:

Quantitative Targets of Fiscal Decentralization

article image
Sources: IMF staff, Article IV, Mali Treasury data, and mission calculations.

A. Local Government Financial Needs to Fulfill Their Responsibilities

25. Fiscal decentralization implies transferring responsibilities and funds accordingly. The law establishes the principle of a financial transfer that is proportional to and simultaneous with the transfer of responsibilities.14 In other words, the central government should simultaneously transfer responsibilities and funds in an amount equivalent to what it is actually spending on those responsibilities.

26. One can identify costs of local policies currently lying on the central government and local governments and break them down by major area of responsibility. There is no precise deconcentration and decentralization plan with costs broken down by ministry or/and central government expenditures. However, existing budget execution data provide a rough assessment of the cost of local policies and, consequently, the financial burden likely to be decentralized. The central government budget is based on elementary cost centers called “functional units”. Each functional unit bears a code number that makes it possible to identify the nature; the sector; and the level of execution of the expenditure (see Annex 1 - Central Government Budget Analysis by Functional Unit). The central government expenditures for fiscal year 2014 were thus distributed as follows (Table 6):

Table 6:

2014 Central Government Budget, Expenditures by Intervention Sector

article image
Source: DG Budget, 2014 budget execution data, and IMF staff calculations.

27. Transfers to the local governments already amount to 11.6 percent of the central government budget (Table 7). These transfers, which total XOF 146.7 billion, concern the education and health sectors and, to a lesser extent, general administration. Added to these amounts is an allocation to the National Local Government Support Fund (FNACT), which is primarily earmarked to local investments and related operating expenditures. These transfers aim essentially at covering payroll expenditures related to education and health (87.1 percent of the total amounts transferred) and other current expenditures (Chart 7).

Table 7:

Financial Transfers from the Central Government to the Local Governments

article image
Source: DGB, budget execution data for 2014, reprocessing by IMF staff.
Chart 7:
Chart 7:

Contribution of the Regions to the Central Government Budget15

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Sources: Treasury units, reprocessing by IMF staff.

28. Local expenditure (deconcentrated and decentralized units) represents one fourth of the central government budget. Overall, central government spending for local levels amounts to 24 percent of expenditures executed in 2014, or XOF 302.6 billion (Table 8 and Annex 2). This central government spending for the benefit of local governments includes direct and indirect transfers (via the FNACT). In addition to the latter spending: deconcentrated units execute a large share of central government expenditures in Regions or in Counties, including investment projects; Local state owned enterprises receive government transfers; and Cross-sector projects, funded by the central government, are implemented locally and accrued as central government funding for local development.

Table 8:

Central Government Expenditures for the Benefit of the Local Level

article image
Source: DGB, budget execution data for 2014, reprocessing by IMF staff.

29. In the long term, the cumulative resources of the local governments could amount to XOF 264 billions, half of which have already been transferred to them. If the central government disengages from policies overlapping with local governments’ and if it transfers all funds related to those policies, the cumulative amount of transfers to local governments would be close to the total amount of local expenditures (Table 9). Yet, there are two caveats in this reasoning:

  • Some expenditures do not seem suitable for decentralization and should be executed within the central government budget This is the case in particular for allocations to legal entities (such as the Office du Niger), sovereign expenditures, and spending on oversight and auditing activities carried out by central and deconcentrated units of the central government.

  • Additional expenditures could be decentralized along with the new responsibilities to be assigned to the local governments for the implementation of public policies in the areas of culture, sports and leisure, administration and infrastructure development (roads and rural electrification), and production and commerce.

Table 9:

Possible Transfers to the Local Level

article image
Source: DGB, budget execution data for 2014, reprocessed by IMF staff.

Approach Based on Responsibilities – Limitations and Caveat

The exact amount of the transfers must still be subject to an objective line-by-line review of functional units to be transferred to local governments. The above study was done on a broad brush basis. It should be complemented by case-by-case analyses, conducted with line ministries, in order to assess whether central government policies should be executed centrally, deconcentrated, or implemented by local governments.

This study has assumed that the productivity of existing units remains unchanged. It therefore does not take into account the necessary reform of the central and deconcentrated units that must accompany decentralization so as to streamline and optimize public management.

Finally, the study is based on the budget executed in fiscal year 2014 (invoices) and should thus be supported by an analysis of other fiscal years.

This approach and its methodology are available to the Malian budget units. The mission carried out this study using only data available in the General Directorate of the Budget. The study does have some methodological limitations, but its interest lies in its simplicity. Budget units can easily reproduce and update the study each year. Annex 1 to this report provides a guide on the use of this budget model.

Source: Mission.

B. Revenues Available in the Regions

30. The revenue-based approach aims at assessing the sustainability of decentralization. It involves providing the regions with funds equivalent to the revenues that they can generate. This approach may trigger a virtuous circle in which local economy correlates with local revenue, which in turn stimulates local development. The logic of this approach advocates that the amount of resources transferred (whatever the terms and conditions) be equivalent to the revenues (local and national taxes, fees and other revenues) raised in local governments constituencies.

31. Revenue collected in regions and sent to the central government budget vary considerably. Chart 7 shows the tax and nontax revenue contribution of each region to the central government budget. During the period 2007-13, the District of Bamako contributed 83 percent (tax and nontax revenues) of the central government budget on average, despite a slight decline in recent years. Less than 1 percent of central government budget revenues come from the Northern regions (Gao, Kidal and Timbuktu). The Kayes, Koulikoro and Sikasso regions contribute 6, 3 and 5 percent of central government revenues, respectively.

32. Although it varies significantly across the country, the revenue collected locally amounts to one fifth of Mali domestic revenue. The structure of local and national revenues is highly asymmetric, with the bulk of collections being concentrated in Bamako (Table 10). Nonetheless, all of the revenues collected by the national Treasury outside Bamako represented approximately 19 percent on average during the period 2011–13. These revenues come from taxation (national and local taxes and levies collected in the regions) and other revenues (fees for services provided, other nontax fees, revenues from government lands and public assets).

Table 10:

Structure of Local and National Revenues

article image
Sources: Treasury staff, reprocessing by mission.

33. Own revenues of the District of Bamako are comparatively low and could improve. Government officials mention numerous difficulties in levying taxes within the district. Collection of per capita taxes (such as the regional and local development tax) is specifically difficult given taxpayer mobility. The ratio comparing the amount of local taxation to national taxation is significantly lower in the district (Table 10). This suggests there is room for improvement. Ceteris paribus, taxation in the district could improve by a factor of 1 to 7 to reach the (admittedly low) level achieved in the regions.

34. Under this approach, tax collected locally amounts some XOF 210 billion. Would the government redistribute local revenue collected in regions across the country some XOF 210 billion would be transferred to the local governments. This amount comprises of: (i) XOF 30 billion for the district, provided it improve its tax collection performance through productivity gains or a change in tax bases and rates; and (ii) XOF 177 billion, or the equivalent of all taxes collected in the regions.

Revenue-Based Approach – Limitations and Caveat

The proposed empiric approach is not based on a study of tax potential. It is limited to identifying existing tax revenues without considering possible improvements except for the District of Bamako, where the objective proposed is to increase own resources to bring them into line with those of the regions.

The coverage of the revenues identified by the Treasury staff may not be complete. Some revenues collected by state owned enterprises, local governments or customs in the stead of the Treasury, may be recorded at a different level from the one at which they are collected.

The instability in the Northern regions leads to a bias in the tax surveys carried out by the Treasury staff. The calculation proposed in this report should therefore be redone when the situation has normalized.

The approach based on tax revenues does not make assumptions regarding the conditions for the redistribution of resources to the local governments. This report naturally advocates optimizing the own resources of the local governments (see below), but the central government may choose the option of transfers by allocation, as long as the allocations remain in proportion to the taxes generated at the local level.

The details of this approach and its methodology are available to the Malian Treasury staff. The mission conducted this study using only available data in the National Treasury and Public Accounting Directorate. The study does have some methodological limitations, but it is interesting in its simplicity. It can easily be reproduced by the Treasury staff and updated each year as needed.

C. Threshold for Triggering a Development Dynamic

35. The development dynamic in the regions has stagnated over the past decade. A number of local development indicators show that the positive effects of decentralization have been mitigated or have been in relative stagnation during the past decade:

  • Poverty (Chart 8) - Three times as many individuals live below the poverty threshold in rural areas and their number is declining much less rapidly than in urban areas (between 2001 and 2010, poverty declined by 40 percent in cities and 19 percent in rural areas).

  • Human development (Chart 9) – The improvement curve of the human development index, most of the components of which depend on the delivery of local services,16 has stagnated over the past nine years. Between 2007 and 2012, Mali’s ranking deteriorated from 175th to 182nd in the world.

  • User satisfaction with communal services (Table 11) – According to a 2014 survey conducted by the United Nations Development Program (UNDP), respondents expressed excellent feedback about the construction of primary schools but a much more mixed sense of satisfaction for other local services.

Chart 8:
Chart 8:


Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Source: 2014 World Bank Report, cartography of poverty in Mali
Chart 9:
Chart 9:

Human Development Index

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Source: UNDP/ World Human Development Report (WHDR).
Table 11:

Degree of Satisfaction with the Quality of the Decentralized Services

article image
Source: UNDP, 2014 National Human Development Report, decentralization study, current situation and outlook.

36. There is a threshold as of which fiscal decentralization brings favorable outcomes. A recent study17 based on international examples correlates development indicators with a decentralization coefficient (decentralized resources / central government tax revenues) and analyzes the effect of decentralization on the favorable outcomes expected. These outcomes are measured in terms of statistics (schooling, access to facilities, connectivity, etc.). This study shows that below a certain threshold, decentralization does not have favorable outcomes and can even negatively affect the level of development.

37. The mission applied this method to Mali using some available indicators, such as basic education (repeater rate and food supply at school) and income poverty. This correlation model yields the following findings:

  • During the period 2001-09, decentralized resources represented 3.8 percent of the tax revenues of the central government on average, but they vary by region.

  • The effects of decentralization seem still to be mixed in terms of the quality of local school services - the repeater rate has been steadily declining over the period but food supply has not improved.

  • The lower the level of decentralized resources, the lower the level of income poverty.

38. This situation varies by region. The levels of decentralization and local development are extremely varied, depending on the region, and the relative attractiveness of the regions also varies considerably. The demographic attractiveness of the Southern regions and the District of Bamako is significantly higher, and the output of these regions is also greater:18 Timbuktu, Gao and Kidal generate just 5 percent of national GDP, the other regions 52 percent, and the capital city the remaining 43 percent. Nonetheless, in addition to the efforts made vis-à-vis the local governments, poverty is still highly dependent on the social way of life, the agrarian or industrial production models, the average level of education, and security.19

39. To trigger a positive dynamic, decentralized allocations must be increased to approximately XOF 180 billion. Projections done using a simple econometric model (Annex 2) at an aggregated level show that to generate favorable outcomes, decentralization requires an increase in transfers of financial resources to the local governments. Poverty reduction actions alone require a slightly higher increase in the resources of the local governments.

Decentralization Threshold – Limitations and Caveat

The approach proposed depends on the quality of the available statistics. It is based on the data held by the Malian Statistics Institute, the availability and historical continuity of which vary.

Factors exogenous to decentralization are also relevant. Housing, ways of life, social structures and political stability have a significant impact on local development.

The action of the deconcentrated units of the central government and legal entities located in the regions are also a factor in development and poverty reduction, even if, in principle, these units do not fall under the local governments. Owing to the local deployment of these units, joint actions have been undertaken and positive synergies have appeared.

III. Appropriate Approach and Levels of Financial Decentralization

40. Financial decentralization is a credible alternative for local development policies, which have been suboptimal until now. The process of decentralization aimed at allocating an essential role to local governments in the area of financial and administrative management helps to remedy certain central government deficiencies, particularly in the adequate supply of public goods. This process thus constitutes a way of improving the efficiency of public interventions in favor of development. By ensuring closer proximity of the authorities, and thus a better understanding of public demand, fiscal decentralization aims to increase the efficiency of the public administration so as to reduce rural poverty and improve access to basic public services.

41. Conversely, poorly implemented decentralization can have undesirable outcomes. The approach bears risks, however. Poor implementation of fiscal decentralization and ill-estimated transfers may lead to unstable and/or inappropriate allocations (in nature and volume). In this case, local governments would implement uneven and ineffective local policies with little economic and social impact, or even build up deficits and debt. If the transfer mechanism is not suited to the needs, it could increase local inequalities, impede the country’s growth, and impoverish the central government without triggering a virtuous circle of growth in which local governments reap the benefits of their development efforts.

42. The following sections review the best ways for a successful fiscal decentralization. The objectives are to stimulate a local dynamic for local development and inclusive and equitable growth, while preserving the stability of the macro fiscal framework and consequently the economy of the country. To this end, the authorities should gradually implement transfers in line with the needs of the local governments, and review the structure of local revenue to decrease dependence on grants and allocations from the central government. This requires optimization of the local governments’ own resources and alignment of transfers considering local governments’ needs and central government’s available resources.

A. Protecting the Stability of the Macro fiscal Framework

43. Malian government self-assigned objective is to successfully complete fiscal decentralization within a few years. As indicated above, the government’s objectives are part of both the Inter-Malian Inclusive Dialogue and the growth and poverty reduction strategy. The timetable set for achieving these objectives is ambitious:

  • Transfer of 30 percent of “central government revenues”20 by 2018;

  • Overhaul of local government map by 2017;

  • Transfer of staff from deconcentrated units to local governments.

44. It is possible, based on these hypotheses, to develop an indicative financial trajectory. To do this, the mission used the assumptions included in Box 6.

Assumptions Underlying the Calculation of the Cost of Financial Decentralization

General assumptions:

Target level of decentralization = 30 percent of fiscal revenues (including grants) by 2018;

Average level of transfers (2009-14 period) = XOF 105.8 billion (10 percent of the central government budget, excluding grants);

GDP increase of 6.7 percent per year through 2018, corresponding to average GDP growth during the period 2009-14;

Increase in tax revenues of 7.9 percent per year;

Decentralization implementation costs (increase of 8 percent per year, including inflation);

Recruitment of additional personnel, to promote a subnational civil service;

- Rehabilitation or construction of infrastructure for the new local governments.

- Assumptions underlying the scenarios:

- Scenario 1 includes the gradual entry into operation of 11 new regions at a rate of 4 in 2016, 4 in 2017, and 3 in 2018.

- Scenario 2 consists of scenario 1 plus an increase of 5 percent in payroll expenditures for the two civil services (subnational and central government). Under this assumption, the central government is not able to complete all personnel transfers planned and must support the extra cost of duplications.

- Scenario 3 is equivalent to scenario 2 but worse (10 percent in extra personnel costs).

45. Fiscal decentralization involves costs that must be kept under control. The objective of decentralizing 30 percent of fiscal revenues by 2018 implies an average annual increase of transfers of XOF 76 billion over the period 2016-18. This increase in budgetary allocations to the decentralized levels of government reflects a 38 percent increase in transfers on average through to 2018. Table 12 shows a projection of these costs under the following three scenarios:

In a “sustainable scenario” (i.e., costs that are sustainable for the central government budget), the central government transfers exactly the financial and human resources corresponding to the transfers of responsibilities. Under this hypothesis, transfers of expenditures (toward the local governments) are accompanied by savings (for the central government) of an almost equivalent amount and some slight additional costs resulting from the creation of the new local governments (11 regional councils, or an increase of approximately 2 percent in the operating expenditures, including payroll, of the local governments). These additional costs, totaling some 3 percent of the central government budget, should be absorbed by the natural progression of central government revenues.

In a “tighter scenario” (i.e., costs that can be absorbed by significant gains in the productivity of local taxation), the government is not able to make a one-on-one transfer of the necessary human resources to the subnational government staff for the operation of the local governments. It must therefore retain excess civil servants,21 while granting the local governments the necessary appropriations for implementing decentralization. These additional costs can amount to 5 percent of the central government budget and can be covered only by a significant increase in central and local revenues.

The “pessimistic scenario” (unsustainable costs) postulates that the government will keep more than a quarter22 of its deconcentrated officials while continuing to increase transfers in order to achieve the target (30 percent of revenues). In this case, the annual extra cost of decentralization will amount to more than 12 percent of the central government budget.

Table 12:

Additional Costs of Decentralization23

article image
Source: Central government fiscal reporting table (TOFE), account of the regional assemblies, and mission calculations.

B. Optimizing Own Resources

46. Optimization of own resources is a significant challenge for the local governments. Own resources (made up of local taxation and other revenues from government lands or fees for services provided) have a number of advantages:

They are predictable and form part of the freedom of administration of the local governments.

They can adapt to specific characteristics and encompass community disparities better than central taxation.

They are linked to the activity, and the local government thus benefits from the improvements in local development, through the expansion of the base for these resources.

They are levers of public policy that can increase the attractiveness of the local community and finance local services.

The proximity of the local government to the taxpayers can reduce collection costs, unless cases of collusion appear.

Own resources can absorb at least a part of the additional costs generated by decentralization.

47. There is significant room for improvement in local taxation. In most cases, the local governments’ tax potential is under-exploited, leaving room for improvement without major innovations (for more details on the weaknesses of local taxation, see the first section of this report). The Malian government (see Box 7) is considering various areas for improvement.

Areas for Improvement of Local Taxation

Improve the tax cycle

Reduce tax avoidance by raising awareness on non-payment of local taxes and levies;

Enhance support by the central government tax administration to optimize local collection of taxes (e.g. those ruled by the General Tax Code) and relieve local elected officials of the responsibility for collections, which is deemed to be contradictory with their positions;

Harmonize the provisions of the Local Governments Code and the General Tax Code to streamline collections;

Assign a portion of local taxes to cover collection costs;

Include the rate of collection of local taxes in the performance criteria of the Tax and Treasury units;

Improve the financial controls applicable to the local sector.

Streamline taxation

Review numerous existing taxes against their revenues and costs, and eliminate unprofitable taxes in order to concentrate on profitable ones (taxes on: business license, local development, and salaries);

Expand the tax base of the local governments;

Identify and list new taxpayers;

Reconsider the base for the regional tax for local development (move from a per capita tax to a property tax in urban areas only);

Subject to the establishment of reliable property registries, create a property tax.

Improve the predictability of revenues

Better understand the tax potential when available statistics make it possible to assess it and better exploit it;

Identify tax expenditures and reduce pointless tax niches (costly and/or with no public policy affect);

Strengthen local capacities to improve revenue forecasting.

Sources: PAGAM GFP.January 2015, interview with the General Tax Directorate.December 2014, Presentation by Philippe Assezat, “Improve the tax for local development in the regions in urban areas.”August 2010, “Study on the revision of the local taxation system,” Stéphanie Flizot.

48. There is similar room for improvement of local nontax revenues. Local governments have numerous sources of nontax revenues: sales of harvests and forestry products, fees and revenues from the use of government lands, provision of services, financial products, works on behalf of third parties, and other current administrative revenues. The optimization of these revenues also involves reviewing current taxes by considering: their collection costs, their acceptability to users, any exemptions that may exist, and the effectiveness of audits.

C. Ensuring a Just and Transparent Allocation of Transfers to Local Governments

49. The system of allocations and grants must be reviewed against best practices. Even optimized, own resources cannot cover all local government needs (Box 8), which is why it is necessary to continue the transfer of allocations and grants. These transfers should have a number of essential qualities:

They should cover the needs of local governments, bearing in mind their specific economic, geographic and social characteristics.

They should encourage local governments to boost their own resources.

They should be predictable and anchored to national revenues so as not to destabilize the macroeconomic framework, particularly when the central government is encountering fiscal pressures.

They should not create significant asymmetries between poor local governments and those benefiting from windfall resources.

They should be assigned to certain categories of expenditures (i.e., investments) only in cases in which public policies or prudent management so require.

Optimization of Own Resources Will Not Cover Needs

To achieve the government’s decentralization objectives, the amount of own resources will need to be increased:

Fifteenfold under the low decentralization assumption (30 percent of central government revenues, excluding grants)

Twentyfold under the high decentralization assumption (30 percent of central government revenues, including budgetary support)

Source: Mission calculations.

50. Most decentralized countries have introduced systems for the equalization of transfers adapted to specific characteristics and local policy needs. While some countries introduce conditional transfers, targeting specific expenditures (Egypt, Ghana, Mauritania and Rwanda), others introduce a more general formula for the allocation of transfers. The current practice is to group these two types of transfers (Ghana, Kenya, Uganda and Tanzania) with a view to pursuing national priorities, while also giving flexibility to local executives in the pursuit of local development programs. Moreover, many countries adopt a budget allocation calculation base, i.e., the share of the central government budget intended for transfers. As an example, Morocco passes on 30 percent of total VAT proceeds, 15 percent of corporate income tax proceeds, and 1 percent of personal income tax proceeds in the form of transfers. Cameroon has established a surcharge (the “centime additionnel”), which is obtained by increasing the amount raised through central taxes such as the corporate income tax and personal income tax by 10 percent; 70 percent of this surcharge goes to the communes and the remainder supplies an investment fund. In the case of Ghana, transfers cannot be less than 5 percent of tax revenues. This anchoring of the main allocation to a national tax presents a real advantage of predictability and transparency of transfers at the central and local levels. Annex 3 presents the formulas used in countries comparable to Mali.

51. An adequate equalization formula combines transparency, simplicity and equity, but must avoid any disincentive effect. Transfers are also a tool for promoting good governance and protecting macroeconomic stability. The combination of these various factors guarantees better predictability of allocations, which benefits not only the central government but also the local governments as they prepare their budgets. Moreover, a less complex system avoids a lack of transparency of transfers and enables taxpayers to assess the effectiveness of local development policies. A basic characteristic of the resource distribution equations is that they must promote the optimization of local taxation. Although various formulas help to reduce excessive asymmetries between local governments, equalization equations do present a risk of being a disincentive to the optimization of local tax performance.

52. Mali is already using an equalization system for local investment. In 2007 Mali established a National Local Government Support Fund (FNACT), which is financed mainly by the technical and financial partners and in part by central government budget allocations and grants. The National Local Government Investment Agency (ANICT) manages this fund earmarked to investments essential for the improvement of public services. To allocate transfers, ANICT uses an interesting equalization formula (Box 9), which combines two series of criteria: (i) situation criteria (the needs of local governments based on their local context); and (ii) performance criteria (sound tax administration by the local governments).

ANICT Mechanism for the Allocation of Capital Grants

ANICT uses an equalization formula that distributes the rights to draw an initial allocation consisting of central government and donor contributions using a formula with two types of criteria:

1. Situation criteria: the number of inhabitants in the community, its distance from major commercial supply centers, the local poverty index;

2. Performance criteria: the capacity to collect tax resources based on the region’s collection rate for the local development tax.

Transfers from ANICT are threefold:

  • – Subsidies for investment and technical support. These two subsidies represent 90 percent of the total transfers from ANICT. They mainly serve to finance investments undertaken under the contracting authority of the local governments (80 percent) and technical capacity building (10 percent).

  • – Other subsidies support the operations of the local governments, particularly the financing of operating expenditures (support for local government operations) and the financing of activities conducted in the context of cooperation between local governments.

  • – Finally, a specific allocation guarantees the loans of local governments, in compliance with the conditions set out in Article 252 of the Local Governments Code.

Source: National Local Government Investment Agency (ANICT).

53. The government could use a similar but improved formula to calculate the transfers for operating local government current expenditure. The ANICT equalization formula has the advantage of simplicity and transparency, but it also has limitations:

The calculation basis, i.e., the initial allocation,24 is a variable that is somewhat too random. The initial allocation varies depending on the sources of financing and the central government budget equilibrium. This leaves discretion to the authorities responsible for deciding on the annual amount of allocations and makes it impossible for local governments to forecast the amount of allocations. To reduce this bias, the authorities could anchor the allocations to a more permanent base, such as to central government tax revenues or a particular tax category, in order to base the transfers on the national economy.

The equalization formula takes little account of geographic differences. The current formula uses the local population as the only criterion, which can disadvantage large, sparsely populated regions. It may be preferable to integrate size or population density in this formula, as other countries do (Kenya, Uganda and Rwanda).

The use of a single tax25 as a tool to measure tax collection capacity may be inadequate, given the extreme disparities in the revenues from this tax from one local government to another, particularly between urban areas and rural areas.

D. Regionalization: A Priority

54. The number of levels of local government is an essential component of the architecture of decentralization. This architecture is generally based on proximity (commune, urban community, neighborhood) for the implementation of basic services not requiring large economies of scale. Other broader levels (departments, regions, counties, cantons) are assigned responsibilities for economic development, investment, infrastructure, land use planning, development of human capital, and even police and security. For example, three types of architecture exist within the WAEMU (Box 10).

Levels of Decentralization in the WAEMU Countries

One level (the commune): Benin and Guinea-Bissau

Financial decentralization at one level is interesting in its simplicity and seems particularly suited to small countries. It is based on a single level of local government – the commune – administered by a mayor assisted by a municipal council. Administrative oversight, advice and assistance, and monitoring of the legality of the acts, is provided by the first deconcentrated level of the central government (the Prefect).

Two levels: Burkina Faso and Cote d’Ivoire.

Decentralization is based on two levels of local government – regions/provinces/districts and communes – with special status given to large cities. This organization is suited to the size of these countries and creates a clear separation of tasks between the two levels to avoid duplication. There is no hierarchy among the levels of local government, with each being freely administered within the limits established by the law and under the oversight of the central government.

Three levels: Mali, Niger, Senegal and Togo

This territorial organization is divided into three levels of local government – regions/districts, counties and communes. Its interest lies in the tighter administrative network it creates and the relative proximity of local executives. The sharing of responsibilities between the intermediate levels (counties or departments) and large levels, in contrast, is difficult, and there can be duplication or overlapping of responsibilities.

Source: WAEMU, December 2014, preliminary report for the white paper on decentralization.

55. The government has not yet established the local government architecture or the contours of the new administrative map of Mali. The law of February 2012 reforming the Local Governments Code introduced 11 new regions but its implementation has been postponed. In 2013, the government26 expressed concerns regarding the appropriateness of establishing these new regions (a number of which are deserts) and wondered about the possibility of eliminating the counties. Another envisaged option was grouping or merging small communes to streamline their management and adapt their contours to social characteristics based on traditions and customs. Under this latter approach, the number of communes would decrease from 703 today to about 600.

56. In the end, the choice of the most appropriate architecture should be based on a rational logic.27 Political choices or compromises resulting from the negotiations under way may well influence the choices and the number of levels ultimately selected, but important implications for public finance management must also be taken into consideration:

Each local government means an additional budgetary cost: for example, the cost of local elections, infrastructure (construction and maintenance), facilities for executives (remuneration, transportation and even housing), management of resources, and operational expenditures for implementing public policies.

The difficulties of budgetary coordination and monitoring increase as the number of local governments increase. The workload (and costs) handled by the various units - oversight, financial controls (internal and external), technical assistance for local management, fiscal planning, accounting, and tracking of financial assets and liabilities - increases owing to the number of actions taken by the local governments and, to a lesser extent, the volume of financing handled.

It is easier to take into account the specific characteristics of broad decentralized levels when they are less numerous and of a significant size. The regions and large urban communities are able to seek economies of scale and the best offer for services if they can look to large and competitive markets for their public policies. Conversely, small communities have higher management expenses for their procurement and must deal with imperfect supply structures (collusive oligopolies).

57. The above considerations support the government’s decision to increase regionalization and advocate in favor of gradual and differentiated implementation. Considering the magnitude of the fiscal decentralization (see section 2), which will see the local budgets triple, it is recommended to proceed cautiously with a few key phases:

  • Consolidation of the communes: This involves adjusting budgets to provide communes with means to assume their responsibilities (local services, including health and basic education).

  • Orderly dismantling of the counties, if that is the government’s decision: This involves organizing the elimination of the counties (some of which capital cities will become heads for new regions) in a rational manner and ensuring the transfer of their assets (infrastructure, human resources, knowledge) to the new regions.

  • Phasing the regionalization: Transfers of resources to the regions should be subject to a number of prerequisites (regional administration actually in place, establishment of an appropriate governance framework). Following this logic, the transfers would be adjusted according to: (i) the needs dictated by geography and circumstances (e.g., the reconstruction of the North, support for zones that are vulnerable to poverty or health risks, improving connectivity for zones with limited communications networks); (ii) the capacities of the executives and the structure of local markets: this involves a relevant sequencing to allow local executives to step-up their management, the central government to introduce appropriate controls, and the local markets to adjust supply to the needs of the executives.

E. Conclusions and Recommendations

article image

IV. Best Fiscal Decentralization Approach for Ensuring Local Good Governance

58. The decentralization process involves inherent risks. Decentralization consists of assigning responsibility for the management of substantial funds to local governments in the service of ambitious projects that are decisive for the country’s harmonious and balanced development. The stakes are thus very high, as are the risks. The risks relate in particular to the proliferation and dispersal of resources and players, making it more difficult to control and monitor public action. The local governments’ poor capacity to handle their new responsibilities could also hamper the overall development effort, if allocation of available resources are suboptimal and not in line with taxpayer’s interest or government policies. Mitigating these risks requires suitable resources, particularly, although not only, sufficient qualified staff. As well, the establishment and implementation of a robust and effective governance mechanism is a key factor in the success of decentralization

59. The local governance framework should rely on an organized, structured and competent administration. Governance is a set of clear, locally appropriate procedures governing public action, along with internal and external audit arrangements aiming at detecting, correcting and penalizing errors and irregularities. It covers all aspects of management, planning and reporting. It requires the identification of each player’s responsibilities of, documented operational procedures, and competent staff assigned throughout the territory where needs are greatest.

60. The aim of this section is to propose the establishment of a framework suited to the challenges of decentralization. It assesses the risks and challenges inherent in the expansion of decentralization, presents a number of approaches for strengthening human resources and establishing an effective governance, and proposes a relevant framework for control. These are essential to the sound use of local resources and the introduction of a development dynamic that supports the stability of the country’s macroeconomic framework.

A. Review of Local Governance in Mali and its Risks

61. Despite the progress made, the functioning of the local executives remains seriously flawed. The General Directorate of Local Governments tracks indicators on the implementation of decentralization and the respect of certain rules of good governance by local governments. These indicators cover, in particular, approval of initial budgets and budget execution accounts and the holding of regular sessions of the deliberating assemblies and the regional and local development action steering, coordination and monitoring committees. As shown in Table 13, the results observed across all regions of Mali indicate that these best practices for local governance are not yet fully entrenched in local government practices. Setting aside the difficulties encountered since 2012 in the Northern regions, performance in the other regions of the country varies considerably and is often weak. The case of the District of Bamako, which has recorded poor results for 4 indicators, is particularly worrisome in this regard.

Table 13:

Local Government Governance Indicators

article image
Source: General Directorate of Local Governments.

62. The audits conducted reveal recurrent difficulties, over time and in all local governments. These difficulties relate in particular to a failure to respect contracting and procurement procedures (failure to use competitive bidding, ineffective or noncompliant services and deliveries, terms of payment, etc.). However, they also concern the expenditure planning and forecasting phase and the limited capacity of local executives to prepare accurate and realistic budgets, taking into account all expenditures (chronic under-budgeting of infrastructure maintenance expenditures, for example) while at the same time not overestimating the resources actually available (impact of the difficulties in collecting local taxes). These weaknesses lead to budget execution and cash flow difficulties, which are reflected, for example, in delays in the payment of the wages of local government officials.

63. There are many weaknesses in the areas of investment as well, involving everything from the failure to respect procedures to poor performance of public spending. For example, audits conducted in 2012 on a sample of 116 works projects carried out in 2010/2011 by 19 counties and 43 local governments in the regions of Ségou, Koulikoro, Kayes and Sikasso28 showed that:

  • The respect of administrative and control procedures for procurement was insufficient for 66 percent of projects (based on the documentation presented).

  • The respect of public expenditure payment arrangements was insufficient for 47 percent of projects (share paid by the local governments, advances, contract holdback, etc.).

  • Administrative and financial control was insufficient for 77 percent of projects.

  • Verification and monitoring of completion were insufficient for 69 percent of projects.

  • 70 percent of works were defective and 17 percent of completed works were not functional or remain unused.

64. There is an execution and control procedure in place for local expenditures. The internal and external control mechanism for local public spending in Mali is robust and involves numerous participants at various stages (Table 15). It involves both the central and deconcentrated units of the central government responsible for oversight of the local governments and for the support/advice function, as well as the external auditing bodies, headed by the Office of the Auditor General and the Supreme Audit.

Table 14:

Participants in the Auditing of the Local Expenditure Process

article image
Source: Compilation from the Local Governments Code and various financial regulations.
Table 15:

Deployment of Units Responsible for Financial Oversight

article image

65. The coverage of the support/advice and audit mechanisms is limited. Both the capacity and available resources of the deconcentrated units of the central government and the units responsible for external auditing are limited. As a result, units providing oversight are varyingly deconcentrated, and not all are present below the regional level, which limits the capacity to monitor and guide the local governments (Table 15). To offset this weakness, delegation arrangements do exist29 but are deemed ineffective.

66. The units responsible for oversight and auditing are notoriously understaffed. By way of illustration, the National Financial Control Directorate has 96 officials in its regional offices and only 44 officials in its local and county offices (or less than one official per county). The National Treasury and Public Accounting Directorate has 11 senior accountant positions and a network of 91 receivers-collectors distributed throughout Mali. A single receiver-collector is thus responsible for monitoring and auditing the operations of 8 local governments on average, in addition to his or her activities in the deconcentrated units of the central government. Similarly, the National Development Planning Directorate has regional directorates and local planning offices in all counties, for a total of 189 officials (15 percent of which are Category A), while needs are estimated at around 1,400 officials.

67. The staff are also unevenly distributed across the country. Most of the deconcentrated staff are located in Bamako, while the regions are more sparsely staffed, particularly the North (Table 16). The proportion of senior officials (Category A) is also higher in Bamako than in the rest of the country, but remains below 20 percent on average.

Table 16:

Deconcentrated Staff of the Central Government in the Regions

article image
Source: Central Payroll Office (statistical databases of wages paid in December 2014).

68. The lack of resources involves not only staffing levels but also equipment. Premises in the regions are often unsuitable and dilapidated. Equipment is out of date or missing (e.g., the Gao and Kidal treasury offices have no vehicles), which limits the ability of the officials to perform their functions.

69. The capacity of the external audit entities to carry out their work is limited, also owing to the lack of resources. The Office of the Auditor General conducted some audits of deconcentrated units of the central government and a small number of local governments (Bamako and Kati) in 2013. The Account Section of the Supreme Court was just starting the verification of the local government accounts for fiscal year 2009 in 2014. As of yet, this work had covered 105 of the 703 communes in Mali, or less than 15 percent of the total. Finally the financial and budgetary discipline chamber of the Account Section is not operational and has thus far done no work. More generally, as underscored by an audit of the expenditure process conducted in January 2014,30 the lack of effective sanctions for the deficiencies identified during audits and verifications deprives them of any practical impact or dissuasive effect on inappropriate conduct.

70. Overall, the general context is not very favorable to local good governance. The structural weaknesses in the expenditure process and lack of resources and capacities are combined with insufficiencies in the internal and external audit arrangements in a context deeply affected by fraud, corruption, irregularities and management errors that are all too rarely penalized. Successful decentralization and achievement of the resulting development objectives requires that these dysfunctions be remedied, at the risk of a deterioration in the overall performance of public policies. Given the multiplicity and scope of the tasks to be performed, in a context in which resources are very scarce, the approach must be gradual, apply multiple levers and be based as a priority on organization, clarification of procedures and responsibilities, and capacity building.

B. Necessary Consolidation of the Local Civil Service

71. Strengthening of the local staff is one of the keys to the success of the decentralization process and control of local public management. No governance arrangement, no matter how complete or robust it may be, can guarantee the regularity and effectiveness of public spending without a sufficient number of competent men and women who are located where the needs lie and who are aware of the necessity of applying the appropriate rules and procedures. In order to make the local governance framework a reality, decentralization must thus go hand-in-hand with the consolidation of the subnational civil service.

72. The draft Agreement for Peace and Reconciliation in Mali31 lists strengthening the local authorities’ staff as an essential objective. To that end, it includes a number of commitments on the part of the central government including: the transfer of deconcentrated units to the local governments in their areas of responsibility; enhancement of the attractiveness of the local civil service, particularly in the Northern regions of the country; and recruitment to the local government staff, with most of positions being reserved for recruits from the North.

73. Despite a previous commitment to strengthening the local governments’ staff, it remains fragile. In 2009, the Strategy for Capacity-Building in the Regions noted the weakness of the staff and capacities in the local government units and developed an action plan to remedy this situation. Since that time, some progress has been made but not enough to cover the needs. The staff assigned to the local authorities has increased, particularly through the recruitment of teachers and health officials. It now stands at more than 47,000 officials, but remains very insufficient in certain areas (Table 17), particularly in the technical field, although it is called upon to play a key role in handling the development responsibilities assigned to the local governments.

Table 17:

Movements of Central Government Officials Between Regions

article image
Source: Central Payroll Office.

74. The local governments’ staff is based on part-time, very short-term jobs. The local governments are still dependent on central government officials made available to them in the context of ad hoc support missions or on a somewhat longer basis through secondments. These arrangements, which are harmful to the autonomy of the local governments, are both costly and inefficient in that they do not favor the transfer of capacities to the local governments. Moreover, contractual staff accounts for approximately one quarter of the staff of the local governments, which is a factor in their weakness and an additional cost.

75. Finally, the local governments’ staff is relatively poorly qualified (Chart 11). Supervision is weak, as indicated by the relatively small proportion of Category A staff in the general administrative framework, particularly in the technical field.

Figure 10:
Figure 10:

Local Government Staffing Levels by Field

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Figure 11:
Figure 11:

Local Government Staffing Levels by Category

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

Source: General Directorate of Local Governments.

76. Transfers of central government staff to the local governments are essential but difficult to carry out. Expanding decentralization assumes a consolidation of the capacity of the local governments’ staff to implement public policies, both in terms of staffing levels and competencies. Realistically and in line with the principle of subsidiarity, this consolidation needs a transfer of central government officials to local governments’ staff as responsibilities are transferred, not by massive recruitments. To date, no transfer plan identifying the functions and jobs concerned and defining the operational conditions for transfers from one civil service to another has been prepared. In any event, such transfers are difficult from a social and human resources management standpoint. Officials are unwilling to accept assignments in the regions owing to the difficult security and climatic conditions, the distance and isolation, and a level of development and equipment that is significantly inferior to that found in Bamako. The local civil service also suffers from an image problem that reinforces the reluctance of officials to join it.

77. The regions are generally not attractive for central government officials. An analysis of payment termination certificates (CCPs) for 2014, which can be used to identify the movements of civil servants and contractual officials of the deconcentrated units of the central government between regions, illustrates these difficulties and reveals sharply contrasting situations (Table 17). Some regions are attractive and record more arrivals than departures (particularly Bamako). Others, which are in fact the most vulnerable, suffer from staff attrition, with departures exceeding arrivals-particularly the regions of Kayes, Koulikoro, Timbuktu, and Kidal.

78. As well, deconcentrated units sometimes have difficulty filling positions in the regions. For example, the National Financial Control Directorate had an actual staff of 140 in January 2015 (96 in the regional directorates and 44 in the local offices in the counties), as compared to a theoretical staff of 187 (139 in the regional directorates and 48 in the local offices).32 The Budget Directorate has an actual number of staff for its deconcentrated units (outside Bamako) that exceeds the number established in the detailed organization chart33 (Table 18). But the situation varies by region: six regions have a surplus while the regions of Timbuktu and Kidal have an actual staffing level that is below the theoretical staffing level.

Table 18:

Actual and Theoretical Staffing Levels in the Deconcentrated Units of the Budget Directorate

article image
Source: General Directorate of the Budget – January 2015.

79. Incentives must therefore be considered to facilitate the transfer of central government staff toward the local governments. These measures can take the form of bonuses linked to jobs located in difficult zones.34 Above all, confidence in the local civil service must be restored by means of measures to guarantee local government officials careers that are no less favorable than those in the national civil service (particularly in terms of advancement and movement up the salary scale) and satisfactory working conditions and remuneration (with particular emphasis on the regular and timely payment of wages, see below).

80. Decentralization must be accompanied by a reform of the central government units. Decentralization cannot be limited to transfers of responsibilities and resources to the local governments. It encompasses a broader reform of the central government and requires adapting the central and deconcentrated units to a new mode of public action. In application of the principle of subsidiarity and with a view to respecting macro fiscal balances, the strengthening of the subnational government staff should be accompanied by a commensurate decrease in central government staffing levels (notwithstanding the limited resources currently available to some deconcentrated units of the central government). However, this decrease must go hand in hand with a change in mission, with the deconcentrated units refocusing on support/advice and auditing functions, which presumes a smaller number of particularly highly qualified staff. A capacity-building plan that identifies jobs and the corresponding qualifications must therefore be prepared for each unit, using an approach that is both complementary and parallel to the development of the personnel transfer plans.

81. The function of local elected official must also be professionalized. While the quality and number of local civil servants are decisive for the success of decentralization, the capacity of local executives to manage complex operations involving substantial volumes of financing is equally important. The function of local elected official must therefore become more professional. This particularly requires awareness-raising regarding the principles of local governance (including ethics and anti-corruption and conflict of interest efforts) and training in public management, which could be provided at the time newly elected officials take office and then on a continuous basis throughout their terms of office. Professionalization also requires fair remuneration for the function of local elected official to make the position more attractive and somewhat shelter elected officials from the temptations of fraud. However, the remuneration paid to mayors, deputy mayors, presidents and vice presidents of county councils, district councils and regional councils has not been upgraded since 2006.35 It ranges between XOF 25,000 per month for mayors of communes to XOF 175,000 per month for the mayor of the District of Bamako (XOF 120,000 per month for the presidents of regional councils).36

C. Establish a Framework Favorable to Good Fiscal and Financial Governance

82. The framework applicable to fiscal and financial management of the local governments must be improved. A body of laws and regulations37 sets out the principles for local government fiscal, financial and accounting management and provides a framework for their operations. The framework defined does not pose any particular problems in and of itself in that it reproduces generally accepted principles intended to control fiscal and financial risks. It must, however, be overhauled and modernized, in line with the provisions of Directive No. 01/2011/CM/WAEMU establishing the Financial Regime for Local Governments within the WAEMU (which should have been transposed into Malian legislation by December 31, 2012). In particular, beyond theoretical principles and as noted above, the fiscal and financial rules applicable to local governments are frequently poorly implemented, affecting the regularity and quality of public spending. Several areas can be considered for improvement with a view to enhancing the fiscal and financial governance of the local governments and controlling the risk of management errors.

83. The local governments are affected by the lack of predictability of the resources available to them. This makes it difficult to prepare and execute local budgets. The following factors come into play in this area:

  • The characteristics of local taxation and collection methods, with the result that revenues collected are often significantly lower than the forecast amounts or indeed the assessed amounts.

  • A significant dependence on allocations paid by the central government and donors. No multiyear budget programming exercise guarantees transparency regarding the amounts and frequency of central government disbursements, which tend to vary from year to year. The local governments are also highly dependent on external financing, which can be affected by all manner of factors (political, economic, etc.).

  • The lack of synchronization of budget timetables (of the central government and local governments): under the Local Governments Code, the local budgets must be prepared and approved by October 31, i.e., by a date on which the central government budget and therefore the amount of its allocations to the local governments are not yet known.

  • Erratic investment procedures. On the expenditure side, the time required for procurement procedures (pre-contract phase) has a negative impact on projections for the completion of investments and is often not compatible with the one-year budget rule.

  • Cash flow constraints. The lack of transparency regarding resources can result in cash flow difficulties at the execution stage, which hampers expenditure forecasting.

84. The oversight mechanisms in place only very partially guarantee the sustainability and accuracy of local budgets. The sustainability and accuracy of local budgets are two essential principles of good governance. However, they are eroded by the lack of transparency of resources mentioned above. Moreover, little or no consideration is given in the budgeting process to some risks that are likely to affect available resources. For example, the identification of liabilities and contingent liabilities does not appear to be fully understood by local governments or by the oversight authorities. This is the case in particular with borrowing. Although borrowing is authorized for the financing of investment, it is in principle strictly controlled to avoid any slippage.38 Nevertheless, even though the financial data available appear to indicate that local governments have very little recourse to borrowing, there are reports of highly indebted communes that are encountering difficulties with repayments, although it was not possible to confirm this. The mission of the public accountants is to monitor and assess the financial position of the local governments, and in particular to identify risks. However, financial analysis tools are deployed in only 160 communes and counties, and the training of collectors seems insufficient at this point to allow them to perform their functions.

85. In general, the monitoring of budgetary sustainability and accuracy is poor. Ex ante budgetary control, which is the responsibility of the oversight authorities, does not include such assessment criteria and remains largely formal and focused on verification of procedural and compliance requirements. Financial checks of commitments focus mainly on the availability of cash and thus constitute a verification of the financial sustainability of the expenditures undertaken. However, in practice, such controls are limited by the delay generally seen between expenditure commitment and validation and do not systematically prevent cash flow difficulties.

86. The procedures for the execution and monitoring of local budgets should improve. A previous report39 presented recommendations on increasing the reliability and strengthening the effectiveness of the public expenditure process. These recommendations fully apply to the local expenditure process. Implementation of these recommendations, which focus in particular on reducing the contracting delays and better targeting controls to make them more effective, implies depending on the existing monitoring and control arrangements, particularly the deconcentrated financial control and public accounting units. These units play a pivotal role throughout the budgeting, financial and accounting process and constitute important levers for improvement. They also play a much-needed advisory function for the local governments, through capacity-building and the dissemination of good practices. Some preconditions must, however, be met to enable them to fully perform their role in the modernization and improvement of local public management. In addition to increasing the staff located at the local level, which can be achieved only gradually, capacity-building through targeted training is an essential prerequisite. Computerization of the expenditure processes should also be continued along with the integration of accounting and budget information systems in order to facilitate and improve the tracking of local government financial operations and simplify the budget reporting and accounting systems. In line with current orientations in favor of concentrating fiscal decentralization on the regions, these changes must first focus the regions, before gradually being expanded to the other levels.

87. Introducing “value for performance” contracts between the central government and the regions would be a significant step forward. This will consolidate the local governments’ budgeting process by giving them more reliable and more realistic bases for their budgets, as long as commitments made are respected on both sides. Increasing the own financial resources of the local governments in the context of a comprehensive reform of local taxation would also help to stabilize revenues and reduce the dependence of the local governments on external sources of financing. Controlling resources and ensuring greater transparency regarding expected revenue levels are essential conditions for making budget forecasts more reliable and ensuring that execution is as close as possible to the expenditure authorizations given by the deliberating assemblies at the time of approval of the budget. An initial survey of requirements to be included in the value-for-performance contracts has already taken place (Box 11).

Central Government-Region Value-for-Performance Contracts

In 2014 Mali began to work toward the establishment of contractual relations between the central government and the regions (or district) in the area of development planning,40 and the signing of value-for-performance contracts between the central government and the regional governments (in at least two regions) is planned in 2015.

These value-for-performance contracts do not replace the existing arrangement for support for local government investment (ANICT / FNACT), but rather supplement it.

The value-for-performance contracts, which are concluded for a term of five years, cover the multiyear scheduling and financing of regional (or district) structural projects in the context of the implementation of the Economic, Social and Cultural Development Program (PDSEC) and give priority to “structural investments that create wealth and jobs.”

article image

The value-for-performance contracts are financed by the local governments (from their own resources), the central government and, possibly, contributions from the development partners, foreign local governments or the private sector. FNACT is responsible for the financial management of the contracts.

The list of projects included in the value-for-performance contracts is proposed by the regions and finalized by CROCSAD41 after the opinion of the sectoral ministries concerned has been obtained and the financial commitments expected from the central government have been approved by the Minister of Finance. Monitoring and evaluation of the value-for-performance contracts is provided by CROCSAD and the ad hoc monitoring committee.42 An initial list of the projects that could be included in the first generation of value-for-performance contracts was prepared in late 2014: 86 projects involving a total investment of XOF 1,276.4 billion were identified, with significant regional disparities both in terms of the number of projects to be included and the amounts to be financed. The 2015 central government budget allocates XOF 830 million for the financing of the value-for-performance contracts.

To accompany the regional development policies and the performance contact approach, there are plans to create regional development agencies in each of the regions and the District of Bamako. These agencies will be responsible for promoting regional and local development and for assisting the local governments in acting as contracting authority for regional and local development, covering all aspects (planning, scheduling and completion of development operations, management of public services, mobilization of resources).

The agencies, which will be placed under the oversight of the Minister for Local Governments, will be chaired by the president of the regional council, while the governor of the region will act as deputy chair.

At this stage, there are plans to provide each agency with 6 officials (including 4 professionals), financed by the central government. The financial impact of the creation of these agencies is estimated at XOF 1.2 billion for the first year and XOF 85 million per year thereafter for the 9 agencies.

Source: [General] Directorate of Local Governments.

D. Introducing an Accountability Framework

88. The accountability framework is based on three principles: transparency, control and responsibility. Good financial governance is possible only in a context of accountability that recognizes good management and penalizes errors and deficiencies. This accountability framework is based on transparency and public reporting on local management, which is a condition for the implementation of controls. These controls are implemented without infringing on the freedom of administration of local governments: although they aim to assess management performance and compliance, under no circumstances do they focus on the appropriateness of the decisions made by local governments. Finally accountability is not compatible with impunity. Local elected officials and civil servants must report on their decisions and actions and, where necessary, sanctions proportionate to the seriousness of the errors or deficiencies identified must be systematically applied.

89. There are weaknesses in the general financial accountability framework in Mali. Although the internal and external controls are structured and documented, covered by the information systems, and abundant, if not to say redundant, they are ineffective owing to a lack of graduated sanctions and proceedings against officials (public and private) involved in corruption.43

90. The accountability of local elected officials and civil servants first requires greater transparency of public management. Local elected officials and managers have a large number of management reporting obligations. The national decentralization policy paper establishes the principle of public reporting on local government management and the Local Governments Code requires a public discussion prior to the approval of the budget. This discussion must cover progress with the implementation of the Economic, Social and Cultural Development Program (PDSEC), the budget execution account for the year ended, the operations of the local government agencies and services, and the draft budget. As well, the Local Governments Code and the local government budgeting and accounting instruction establishes the obligations of local government budget managers and accountants to maintain and report on the accounts. The quality and availability of the fiscal, financial and accounting information are essential conditions for the implementation of internal and external controls. Beyond that, their dissemination and presentation at public accountability sessions are a way of making elected officials accountable to the citizens, who are also taxpayers and voters.

91. The development of citizen oversight thus constitutes a useful and relevant supplement to the institutional accountability mechanisms. Initiatives have been taken to promote citizen oversight, but they remain limited and need to be stepped up. The Ombudsman of the Republic each year organizes a national day devoted to the “Forum for Democratic Review” (espace d’interpellation démocratique). During this day, citizens address oral and written questions to the line-ministries,44 which are required to respond.45 Recommendations to the authorities, to civil society and to the political parties are made at the end of the discussions. This practice, which has the benefit of organizing the citizen review approach, could be developed and expanded to the local government level. More generally, the expansion of citizen oversight could rely on the National Civil Society Council, which is responsible for promoting good economic and financial management throughout the country. In October 2012 the council established a framework for cooperation with the Auditor General with the following objectives: sharing the audit reports with civil society organizations and, more generally, facilitating access to reports of the Office of the Auditor General and their dissemination; making citizens aware of the role and importance of the Office of the Auditor General; creating a forum for discussion between the two structures; and supporting the technical capacities of the National Civil Society Council. This initiative could provide an opportunity for raising public awareness of the challenges inherent in management and for developing and consolidating the role of civil society in the oversight process and increasing the accountability of elected officials and public managers.

92. Enhancing oversight of the local governments requires better coordination of the participants involved. As indicated above, ex post controls are frequent and they mobilize resources that are both insufficient and, more particularly, very scattered: Financial Control (in its new mission of evaluation of performance), ministerial inspections offices, the comptroller general of public services, ANICT, the Office of the Auditor General, the Accounts Section of the Supreme Court, and even the financial and technical partners, etc., participate to different degrees and at different stages in the oversight of local government actions and management and evaluate the regularity and quality of local public spending. Streamlining efforts through better coordination of the audit plans (while respecting the prerogatives and independence of each entity) would help to increase the scope and effectiveness of the audits and evaluations thus conducted. In particular, planning of interventions throughout the country could help expand the number of local government levels that are subject to some kind of audit each year.

93. Greater accountability of local participants requires application of the laws and regulations and the implementation of the sanctions provided by the law. Mali has a body of laws and regulations that establishes the accountability of local managers in cases of lapses. For example, Law No. 06–043 of August 18, 2006 establishing the Status of Elected Officials of Local Governments, which sets out the rights and obligations of elected officials, specifies in its Article 15: “elected officials of local governments are required to carry out their functions with diligence, integrity, objectivity and impartiality. In the exercise of their functions, they are formally prohibited from soliciting and receiving gifts, gratifications or advantages of any kind, either directly or through intermediaries.” Local budget managers can be pursued in the budgetary discipline chamber of the Account Section of the Supreme Court and as such are subject to fines ranging from XOF 50,000 to XOF 300,000 for the failure to respect the public revenue and expenditure procedures.46 Moreover, like all citizens, local civil servants and elected officials are subject to criminal sanctions if they are found guilty of crimes or offenses. Although Mali has lagged in implementing the WAEMU directive on the establishment of a fully functioning Court of Accounts, it has transposed Directive No. 01/2009/CM/WAEMU of March 27, 2009 establishing the Fiscal Transparency Code within the WAEMU into national law.47 Finally, a law for the prevention and suppression of illicit enrichment was adopted in May 2014. It is applicable to any person vested with public authority, including local elected officials. It sanctions substantial increases in the assets of public officials that cannot be justified on the basis of their legitimate earnings with severe penalties, including imprisonment in the most serious cases. Legal entities involved in illicit enrichment can be closed and excluded from public contracting.

94. The legal arsenal does indeed exist, but it is not used, which seriously affects the real impact of the accountability framework. The budgetary discipline chamber has never met, legal proceedings are rare, and sanctions are very weak. This almost total lack of accountability of local elected officials and managers in cases of irregularities, errors or dishonesty reinforces a sense of impunity and constitutes a major risk to making public spending more secure in a context of increased decentralization.

E. Conclusions and Recommendations

article image
article image

Annex 1: Central Government Budget Analysis by Functional Unit

The analysis of the budget is based on the breakdown of the functional unit (UF) code. This is a 17-digit code that is based on the central government budget nomenclature as established by Decree No. 03–163/P-RM of April 16, 2003 and containing all of the characteristics of each budget line item. A functional unit corresponds to each budget line item.


Breakdown of the Functional Unit Code

Citation: IMF Staff Country Reports 2015, 287; 10.5089/9781513515076.002.A001

*RNP = Répertoire National des Projets (National Inventory of Projects)The identification of the nature and level of execution of central government expenditures requires a combination of three sorting criteria: “sectors,” “geographic code” and “services and projects”

Central Government Budget Nomenclature

article image

In practice, the functional unit codes shown in the Excel spreadsheet covering fiscal year budget execution1 must first be disaggregated. This disaggregation is performed by using the “convert – fixed length” function (“data” tab) in Excel.

As soon as the sector codes (first digit of the UF), geographic codes (8th, 9th and 10th digits of the UF) and services and projects codes (second digit of the UF) are separated, it is possible to sort the data based on these three successive criteria and thus to group the corresponding budget line items and identify the associated amounts. To simplify the analysis, it can be useful to create tabs by sector and to use the “subtotal” function to determine the amounts by geographic code and by services and projects.

The deconcentrated units correspond to “services and projects” code [4], while transfers to local governments correspond to “services and projects” code [5]. Manual reprocessing is required, however:

  • - To identify (if necessary) certain financial transfers corresponding to payroll expenditures of the regional and District of Bamako governments for the “education, training and research” sector from among the budget line items corresponding to “services and projects” code [4]; these transfers are identified by the note “regional council” or “district council” in the UF description.

  • - To distinguish transfers to the local governments from other financial flows corresponding to “services and projects” code [5]; transfers to the local governments are identified by the note “regional council,” “county council,” “commune,” “district council” or “district mayor” in the UF description.

Annex 2: Model Used to Calculate the Threshold for Positive Outcomes from Decentralization

The optimal level of decentralization, measured in terms of the resources transferred to the local governments, is obtained using the following model:


Yit represents a socioeconomic indicator (level of education, number of school cafeterias, poverty threshold, etc.) and i and t represent the individual dimension (region) and temporal dimension, respectively. Transfit corresponds to the financial resources transferred each year to each decentralized local government (decentralization level proxy). This variable is also squared, which allows for an optimal transfer threshold that triggers positive outcomes. trend represents the “time” variable and makes it possible to remove any development indicator trends not induced by an increase in transfers. εit represents the stochastic error term.

Our sample consists of the eight regions of Mali, excluding Bamako,

1 over the period 2001-2011. The model is estimated using the fixed effects method in order to take account of regional disparities. For greater accuracy of the coefficients estimated, temporal fixed effects should ideally also be introduced. However, purely technical constraints, particularly the size of the sample and the loss of degrees of freedom, restrict the consideration of such effects.

Tableau 1:

Impact de la décentralisation sur l’éducation et la pauvre

article image
Note: (*), (**) et (+) indique la significativite aux niveau 10%, 5% et respectivement. Les écarts-types sont donnés entre parenthèses

Table 1 below presents the results of the estimation of the above model. Overall, the coefficients present the anticipated results. An increase in transferred resources reduces the repeater rate (-0.04) and improves the effectiveness of primary education (4.0). What is even more interesting is that column 3 shows that decentralization had a negative impact on the development of school cafeterias. However, the decentralization variable squared appears with a positive sign, showing that there is an optimal level of decentralization to be achieved to obtain a positive impact on the development of school cafeterias This level of optimal decentralization, measured in terms of transfers (Transfit*) is obtained by simple derivations:


The threshold of the positive impact of decentralization resulting from our sample is estimated at XOF 179.7 billion.2 It corresponds to the threshold as of which the three rates studied improved, although with trend differences, as poverty improves on condition of a slightly higher amount of transfers. These calculations reveal that an increase of approximately one quarter (23 percent) in the current budgets of the local governments would make it possible to reach the positive outcome threshold.

Annex 3: Some Examples of Transfer Formulas in Countries Comparable to Mali