66444 POVERTY THE WORLD BANK REDUCTION AND ECONOMIC MANAGEMENT NETWORK (PREM) Economic Premise JANUARY 2012 • Number 73 JUN 010 • Numbe 18 Enhancing the Capability of Central Finance Agencies Richard Allen and Francesco Grigoli A well-organized and effective finance ministry and its associated central finance agencies (CFAs) are essential to good fiscal outcomes. Recent studies of CFAs in selected low-income countries demonstrate that their organizational structure varies widely and that decision making on public finance is strongly influenced by political economy factors. This note presents some common themes arising from the case studies and provides some descriptive evidence about the organizational structure of the CFAs. This note also includes evidence of a “U-curve� pattern in the development of the organizational structure of CFAs over the long term. It confirms that political economy analysis, though difficult to apply in practice, is highly relevant in the development of reform strategies for strengthening CFAs and public financial management systems. CFAs are not a single organization or entity of government, but Table 1. Central Finance Functions a group of ministries and agencies1 of which the ministry of fi- Macroeconomic forecasting, analysis, and fiscal policy nance (MoF) is usually the most prominent, with collective re- Tax policy sponsibility for the design and execution of a country’s wide Budget preparation and analysis array of financial and fiscal functions.2 In this note, the term Public investment management “CFA� is used as shorthand for ministries and agencies that play a significant role in undertaking such functions. These func- Aid and debt management tions can be divided for convenience into the 16 categories Financial assets and liabilities shown in table 1. Intergovernmental fiscal relations The allocation of roles and responsibilities for central fi- Treasury and cash management nance functions among government agencies varies substan- Accounting and reporting tially from country to country and results in higher or lower Internal auditing degrees of fragmentation. There is no “best practice� model.3 In some countries, the MoF employs more than 100,000 staff, in Public procurement others, less than 100. Such huge variations are not only related Civil service pay to easily quantifiable factors such as population size and the Financial sector regulations size of the public sector, but also to a range of informal and non- Financial framework for managing state-owned enterprises technical factors—described in the present study as political Tax revenues and customs administration economy factors. Public financial management reform coordination This note addresses the important but little discussed issue Source: Authors’ compilation. of how the organization and management of CFAs could evolve 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise in light of political economy considerations.4 Starting with the chies, corporate planning, and information systems; and work of Dressel and Brumby (2009), which sets out a political strengthening the management of human resources and inter- economy framework for analyzing CFAs, this note highlights nal incentives. some common themes and features that were elaborated in a In practice, many countries have focused attention on set of case studies of 10 low-income countries (LICs).5 In addi- strengthening capacity, with less emphasis on strengthening tion, the note presents some descriptive findings based on a da- capability. The two concepts are typically linked: where capac- tabase of CFAs in 55 countries at various stages of develop- ity is low, capability is also likely to be limited. However, this ment, and concludes with a presentation of the case studies’ relationship does not hold in all cases: a weak configuration policy implications. and/or organization of inputs and a high-cost operating envi- ronment, perhaps also marked by institutional constraints Methodological Framework such as a finance minister who lacks a power base within the The framework summarized in figure 1 defines five key inter- government, may mean that even when capacity is high, capa- faces for CFAs—with formal political institutions, administra- bility may be low. In other countries, the reverse situation of tive institutions, civil society, external actors, and within the low capacity and high capability may arise: finance functions CFAs themselves. are well organized and professionally staffed and business pro- The framework makes an important distinction between cesses are efficient, but outcomes are constrained by limited the capacity and capability of CFAs. “Capacity� refers to the inputs. volume or scope of inputs such as human resources or ICT sys- It can be argued that conventional instruments for analyz- tems. “Capability� focuses on how such volumes can be con- ing public financial management (PFM) systems—notably the verted into better performance through mechanisms such as Public Expenditure and Financial Accountability (PEFA) As- clarifying roles and responsibilities in performing CFA func- sessment Framework and much technical assistance work car- tions; strengthening arrangements for coordination and infor- ried out by the international financial institutions (IFIs)—have mation sharing within and across CFAs; clarifying relations focused primarily on the formal, technical characteristics of with line ministries, civil society groups, development part- these systems and have paid too little attention to analyzing the ners, and other stakeholders; improving the management of important political economy factors noted above. However, internal business processes such as decision-making hierar- these factors often help explain why many new initiatives for Figure 1. The Political Economy Environment of CFAs sociostructural context formal political institutions s no rm (executive, rm no s parliament, SAI) incentives interests civil society incentives central �nance incentives administrative (interest groups, agency institutions media, advocacy interests (line ministries, interests roles functions groups) decentralized units) leadership incentives interests external actors (capital markets, s no m donors, international r rm no s �nance institutions, and others) Source: Dressel and Brumby 2009 (figure 1). 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise reforming PFM systems and CFAs fail to be implemented fully, CFAs. The internal capability of an MoF was broken down into and why de jure reforms based only on changes in formal legal the following six dimensions: (i) organizational structure; (ii) at- systems tend to be more successful than de facto reforms based traction and retention of staff; (iii) performance management; on actual changes in behavior, which lag behind (Andrews (iv) decentralized decision making; (v) change management; 2010). and (vi) information technology. Issues of interest that emerged The case studies conducted by the World Bank (2011) help from the case studies include inadequate staff records, poor sys- to fill a gap in the literature on strategies for reforming PFM, tems for recruiting and retaining staff, ineffective performance namely insufficient analysis of the institutional and political management, inefficient top management structures and over- economy foundations of such strategies. Such analysis is impor- hierarchical decision-making processes, and ineffective use of tant to ensure that initiatives to strengthen CFAs and PFM sys- information and communication technology (ICT) systems. tems are politically tractable as well as being technically sound. Fourth, the case studies confirm the view that changing the Reforms should be compatible with the incentives of the na- organization of finance ministries and CFAs is a difficult pro- tional authorities and donors to ensure that reform implemen- cess that is likely to meet opposition from both politicians and tation is not constrained and does not experience the setbacks staff. Moreover, there is an absence of international bench- that often occur in developing countries. marks, guidelines, and standards indicating the desired organi- zational structure or number of staff required to implement Evidence from the Case Studies key finance functions. In advanced countries, these arrange- World Bank (2011) includes case studies of 10 LICs, 6 of which ments vary widely from country to country depending on the were in the Africa region (Benin, Cameroon, Chad, Ghana, legal and administrative traditions and practices. Rwanda, and Sierra Leone), 2 in East Asia and the Pacific (Mon- Fifth, the studies tend to confirm results of other work: golia and Tonga), 1 in Latin America and the Caribbean (Nica- weak coordination among donors—and between donors and ragua), and 1 in the Middle East and North Africa (the Repub- their clients in finance and planning ministries—is a major lic of Yemen). This section summarizes some of the main problem in designing and implementing coherent strategies for lessons emerging from these studies. reforming CFAs and PFM systems (Allen 2008; de Renzio First, in several of the countries studied, the head of state 2011). There is a tendency for donors to bundle together a large plays an unpredictable and sometimes capricious role in mak- number of initiatives in a single PFM reform strategy or lending ing executive decisions on the budget and financial policy. operation,7 without giving due attention to the government’s While such powers are normally exercised within the frame- capability to implement the measures concerned; donors also work of the constitution and budget law, they are not always tend to be too optimistic about how long it will take to imple- compatible with the idea of an efficient and transparent budget ment reforms. In reality, it is the marginal improvements in process, as described in the literature on public finance. An im- PFM that are important rather than the overall plan, which is portant implication of this analysis is that large tracts of the fis- often too elaborate and overengineered. Moreover, securing cal process are likely to be outside the direct influence of the fi- large loans from donors may improve the recipient country’s nance minister and off limits from a reform perspective. status regionally and internationally, as well as providing a valu- Second, it was observed that, in some countries, the head of able new source of economic rents. state uses his political authority to divide the financial power Finally, the study confirms evidence set out in other stud- attributed to the minister of finance among several ministers. ies—for example, de Renzio (2011) and Allen (2008)—namely Such actions are often taken to avoid the accumulation of exces- that reforms of CFAs and PFM in many LICs are strongly influ- sive authority in the hands of one person, and to preserve the enced by donor requirements, with insufficient government discretionary powers of the president over spending and reve- “ownership.� Moreover, experts from advanced countries, nue collection. Another reason is that ministerial portfolios not hired to advise on the technical aspects of PFM reforms, may only provide access to rents, but also accommodate factions recommend finance ministers to proceed with reforms that are within the ruling coalition. Such tactics of “divide and rule� are insufficiently related to local needs and capabilities.8 not unknown in developed countries,6 and seem to be quite Analysis of the Organizational Structure commonplace in less-developed countries. In some LICs, espe- of CFAs cially early in the tenure of a new head of state, the reverse phe- nomenon can be observed: the head chooses to use the finance In addition to the case studies described above, data on a wider minister as an ally in building the economic and financial power sample of countries were collected on the allocation of central of the state, and gives the minister strong powers. Such an alli- finance functions and related activities among the MoF and ance can also give a strong momentum to the reform process. other CFAs, as well as on staffing issues and ICT usage.9 Third, several of the case studies involved functional and or- A core objective of the empirical analysis is to consider the ganizational reviews of the MoF and its relationship with other extent to which CFA activities are concentrated in the hands 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise of one agency—usually the MoF—and to what extent they are accounts within a treasury single account are other examples distributed among other agencies of government. For this of this centralization trend. This trend has both a technologi- purpose, two measures of fragmentation were used: a narrow cal aspect (to increase overall fiscal control) and political as- definition that measures the proportion of finance activities pects requiring greater accountability and transparency: the that are carried out by the central departments and units of emergence of a professional middle class holding the govern- the finance ministry and a broader definition that measures ment accountable for fiscal performance; a less dominant role the proportion of activities carried out by the finance minis- for rent seeking in driving budget allocations; the rise in power try and/or the subordinate agencies that report directly to it. of the finance ministry as an institution of government; the The analysis also draws a distinction between policy activities increased political importance of the annual budget; and in- (such as budget formulation and tax policy) and activities creasing reliance on international capital markets as a source that are largely operational (such as tax collection and procure- of finance. ment transactions).10 In high-middle-income and advanced countries, finance It can be argued that, as countries move from the lowest in- ministers are unlikely to relax their grip on central controls of come group to the highest, the fragmentation of CFAs will ex- the budget process unless political conditions require them to hibit a U-shaped pattern such as that shown in figure 2. do so. However, even in the budget area, the role of the finance At low-income levels, countries often have highly fragment- ministry may change from directly operating the systems con- ed CFAs in which control of public finances is divided accord- cerned to an oversight and monitoring function, while direct ing to powerful political groups and, as discussed, by heads of operations are transferred to line ministries and/or subordinate state who deliberately fragment the authority of the finance agencies responsible for treasury and procurement functions. minister to boost their own authority. Dispersed manual sys- In many operational areas, finance ministers may delegate op- tems of accounting, reporting, and financial control tend to re- erational responsibility for the activities concerned to subordi- inforce this phenomenon. nate agencies under their supervision. Fragmentation may As countries move to middle-income status, pressures to therefore increase again, as the “managerial� culture and decen- consolidate financial activities within the finance ministry in- tralization spread, new agencies are created, and further com- crease, often with the active encouragement of the IFIs. For puterization takes place. example, many countries in Central and Eastern Europe went This historical perspective of fragmentation has important through this process in the 1990s after the breakdown of the policy implications.11 LICs should be encouraged to centralize former Soviet Union. Soviet-style central planning bureaus their finance functions and be cautious about enacting poli- were dismantled, and central financial control systems rein- cies—for example, fiscal decentralization—that do the reverse. forced. The goal of joining the European Union (EU) was a Middle-income countries should be encouraged to gradually further incentive for these countries to centralize finance devolve authority for financial decision making to the agencies functions, because they had to comply with the strict financial responsible for specific functions and programs, but without conditions (on internal control, audit, and public procure- creating undue stress on their financial systems. For example, ment) imposed by EU rules. Policies such as integrating the too rapid growth of autonomous government agencies has been budget and planning processes, bringing extrabudgetary a feature of developments in many middle-income countries, funds within the budget, and consolidating government bank including in Eastern Europe and Central Asia, but without ad- equate safeguards and controls this growth can threaten the overall financial stability of these countries. Devolution should Figure 2. CFA Fragmentation across Income Levels go hand-in-hand with the building of institutions and the de- velopment of credible anticorruption policies. Some evidence in support of this interesting U-curve hy- pothesis is provided in panel A of figure 3, where both the CFAs fragmentation broad and the narrow index reflect the U-curve. As expected, panel B reveals that operational activities are in general more fragmented than policy activities across income groups. Panel B also shows that, while the fragmentation of operational activi- ties follows the U-curve, the curve for policy activities is flatter. These tentative results need to be confirmed with more com- prehensive data. low income lower-middle upper-middle high income Figure 4 indicates that the allocation of roles and responsi- income income bilities for central finance activities among the finance ministry Source: Authors’ illustration. itself and other government agencies varies substantially from 4 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise Figure 3. CFA Fragmentation across Income Groups 70 80 60 70 60 50 fragmentation index fragmentation index 50 40 40 30 30 20 20 10 10 0 0 low income lower-middle upper-middle high income low income lower-middle upper-middle high income income income income income fragmentation index—narrow fragmentation index—policy activities—narrow fragmentation index—broad fragmentation index—operational activities—narrow Source: Authors’ illustration. region to region.12 Of the six regions studied, the Middle East tially huge additional source of fiscal revenues—and fiscal and North Africa and Europe and Central Asia show the high- risk—that need to be managed, sometimes involving special est levels of fragmentation, and South Asian countries the low- instruments such as sovereign wealth funds. In the case of est. It should be remembered that the small sample size is a fragile states, the reverse is true. As discussed in a recent World limitation for such interregional comparisons. Bank study (Fritz, Hedger, and Fialho Lopes 2011), there is Finally, one would expect the fragmentation of CFA activi- often a unique opportunity in such countries to consolidate ties to be higher in resource-rich countries than in non-re- fiscal functions and activities within the finance ministry be- source-rich countries because resources constitute a poten- cause of the emergency status of public finances, the fragility Figure 4. CFA Fragmentation across Regions 70 60 50 fragmentation index 40 30 20 10 0 advanced East Asia & Europe & Central Latin America & Middle East & South Asia Sub-Saharan economies Paci�c Asia Caribbean North Africa Africa fragmentation index—narrow fragmentation index—broad Source: Authors’ illustration. 5 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise of political groups, and because consolidation is usually in- thors and should not be attributed to the IMF, the World Bank, sisted upon by donors and welcomed—or at least acquiesced their Executive Boards, or their management. to—by the government as a requirement of donor support. About the Authors Data analysis provides tentative support for these hypotheses (results not reported). Richard Allen is an economist and consultant with the Interna- tional Monetary Fund and the World Bank, was formerly a staff Policy Implications of the 10 CFA Case member of the Bank and the IMF’s Fiscal Affairs Department Studies (FAD), and was also with the Organization for Economic Co-oper- The World Bank (2011) synthesis report shows that political ation and Development. He is preparing, with Barry Potter and economy analysis of CFAs is highly relevant—indeed essential— Richard Hemming, the International Handbook of Public Finan- to the design and implementation of new initiatives to strength- cial Management, which will be published by Palgrave-Macmil- en CFAs. An important implication of this finding is that po- lan in late 2012. Francesco Grigoli is an economist at the IMF’s litical economy analysis is valuable in filtering out reform FAD, formerly a consultant of the World Bank’s Public Sector proposals for which political economy factors are favorable to Group. He works on public expenditure and fiscal policy–re- success from those where there is a high risk of failure. One po- lated issues. He holds a PhD in development economics from tential method of analyzing such factors may be through case the University of Insubria, Italy, and was a Visiting Scholar at Co- studies similar to those famously developed by the Harvard lumbia University for the last two years. Business School for analyzing how companies resolve structur- al, institutional, and human resource issues that constrain their Endnotes growth and productivity.13 1. These typically include, in addition to the MoF, certain Second, the case studies support work by Andrews (2010), departments of the central bank and the ministry of economy; de Renzio (2011), and others and demonstrates that “best line ministries; subordinate agencies of the MoF with responsi- fit,� “good enough governance� and second-best solutions, as bility for functions such as tax collection, customs administra- opposed to best practice solutions, provide a better approach tion, debt management, and public procurement; and organi- to reform and need to consider the institutional environment zations that are external to the executive branch such as the as well as technical conditions. A change in emphasis is re- legislature and the external audit authority. quired in cases where international finance institutions or 2. In the so called “limited access order� countries, the role bilateral donors are providing support for building stronger and responsibility of the finance ministry largely reflects the PFM systems in developing countries. Such a change in em- consensus of the political and business elite on how to use pub- phasis is the reverse of most current practice: technical solu- lic resources to help maintain the prevailing political order. In tions would be proposed only after a detailed review of insti- contrast, in “open access order� countries, the finance minis- tutions has been carried out and solutions tailored to the try’s role and responsibilities are largely defined by the goal of specific political and institutional environment of the coun- maximizing public resources for the public good. try concerned. Lending instruments could be developed 3. Allen and Kohnert (forthcoming) provide a discussion on (such as the Bank’s Performance for Results Framework) that comparative CFA structures in a range of advanced countries, facilitate such an approach. Reforms should only be proposed with special reference to France, Germany, and the United when they are compatible with existing institutions and in- Kingdom. centives.14 Third, the data provide some support for the U-curve hy- 4. The note is based on the findings of a study of CFAs in pothesis that historically CFAs may go through various stages of low-income countries that was financed jointly by the Bank- fragmentation in their evolution from low-income to high-in- Netherlands Partnership Program and the World Bank during come status. The database provides useful information on the 2008–11. The full results of the study are included in a synthe- comparative fragmentation of CFAs across regions and coun- sis report (World Bank 2011). tries, the number of staff working in central ministries of fi- 5. World Bank (2011) describes these case studies in detail. nance including a breakdown by skills and gender, and the use 6. For example, countries such as Australia, Brazil, Canada, of ICT systems. If further developed, the database would pro- France, the Republic of Korea, Turkey, and the United States all vide a valuable research tool. have divided finance ministries. In most of these countries, po- litical factors—notably an attempt to dilute the authority of the Acknowledgments finance minister—played an important role in the decision to The authors are grateful to Linda Van Gelder, James A. Brumby, split these functions (Allen and Kohnert forthcoming). and Nicholas C. Howard for their valuable comments and sug- 7. The use of “basket funding,� in which donors pool their gestions. The views expressed in this note are those of the au- resources into a single fund and slice up the reform program 6 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise into pieces for which each of them take responsibility, may ex- 13. For example, players representing different organiza- acerbate this tendency for reform programs to grow beyond the tions—the finance ministry, line ministries, parliament, civil capability of either the donors or the client to implement them society, and others—could be brought together for the imple- effectively. mentation of a specific policy initiative (for example, a new 8. A related issue is that countries may not have the human pension scheme or a proposal to reform local government fi- capacity to implement the reforms that are urged on them by nance) or a structural reform (such as establishing an inte- the donor community. As a result, it is commonly observed grated revenue authority) to review how objections to such a that CFAs supplement their regular staff positions with consul- reform, and constraints to implementation, can be resolved, tants and advisors financed by the donors. Such advisors often and how incentives to facilitate implementation can be aligned. take responsibility for line management positions and once 14. Hasnain (2011) provides a useful example from Mongo- hired are difficult to move out. More important, they may per- lia, where the parliament exerts a powerful influence on public manently weaken the capability of the CFAs to build local ca- finances and is able to block important reforms if they do not pacity to fill the positions concerned. meet the perceived needs of constituents. The paper argues 9. Of the 110 countries that were originally targeted, only that the parliament may be persuaded to support fiscally valu- 55 countries responded to the survey, while responses to ques- able reforms, for example, the development of an economically tions on staffing and ICT systems were 22 and 30, respectively. efficient system for allocating development grants, if their ac- 10. More formally: cess to constituency development funds—which has a low fiscal Fragmentation Indexi= , ivit carried cost—is also preserved. s activity carried out of acti it es is t nu MoF A is anthe total numberuniquely by the MoF, n is the where ac vi ies References ically carr activities, is ot ex t nd is the s bs total number ofd ut or po the number of activities that are g not systematically carried out w do ref exist, and i is the sub- Allen, R. 2008. “Reforming Fiscal Institutions: The Elusive Art of the Budget us gh r ind x va u or l not ect a m ra c Advisor.� OECD Journal of Budgeting 2008/3. script indentifying the index definition. Thus, a higher index fa ti n i s t a now Allen, R., and P. Kohnert. Forthcoming. “The Anatomy of the Ministry of Fi- value will reflect a more fragmented finance ministry. The r b b oF In Public nance.�with Financial Management in the 21st Century, ed. M. Can- broad fragmentation index is computed in the same way, but A giano and M. Lazare. IMF. M F S r d b e of activities that arel carried out primarily by a sub- now includes pa at l Andrews, M 2010. “Good Governance Means Different Things in Different o l agency ofsthe MoF, with or without the contribution ordinate Countries.� Governance 23(1). de Renzio, P. 2011. “Can Donors ‘Buy’ Better Governance? The Political Econ- of the MoF. Similar indices have been computed separately for omy of Budget Reforms in Mozambique.� Cadernos IESE No. 9. policy and operational activities. Dressel, B., and J. Brumby. 2009. “Enhancing the Capabilities of Central Fi- 11. Because the advocated reform approach is based on the nance Agencies: From Diagnosis to Action.� Framework paper, World Bank. notion of “best fit� solutions linked to specific country circum- Fritz, V., E. Hedger, and A. P. Fialho Lopes. 2011. “Strengthening Public Finan- stances, these policy implications need to be interpreted broad- cial Management in Postconflict Countries.� Economic Premise 54, World ly and flexibly. Bank. Hasnain, Z. 2011. “Incentive Compatible Reforms: The Political Economy of 12. Andrews (2010) shows that countries that exhibit good Public Investments in Mongolia.� Policy Research Working Paper 5667, outcomes can have very different governance structures, chal- World Bank, Washington, DC. lenging the idea of one best-fit model for public financial man- World Bank. 2011. Strengthening the Capability of Central Finance Agencies: Syn- agement systems and government structures in general. thesis Report. Washington, DC. The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. They are produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at: www.worldbank.org/economicpremise. 7 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise