Political Economy of Public Financial Management Reforms Experiences and Implications for Dialogue and Operational Engagement Verena Fritz Marijn Verhoeven Ambra Avenia The World BAnk  i Cover: (top) Philip Shuler/World Bank, 29 June 2013; (bottom) UN photo/Marco Dormino, 05 December 2014. Standard Disclaimer This report is the product of the staff of the World Bank Group. The findings, interpretations, and conclusions expressed herein are entirely those of the authors and should not be attributed to the World Bank Group or its affiliated organizations, or to members of the Board of Executive Directors of the WBG or the governments they represent. The WBG does not guarantee the accuracy of the data included in this work. Acknowledgements T he report was prepared by a team comprising Verena Fritz, Roland Lomme, Jens Kristensen, and Gloria Joseph Haji (Nigeria); Marijn Verhoeven, and Ambra Avenia, with contributions by Kai Kaiser and Saeeda Rashid (Philippines); and Chiara Bronchi, Stephanie Sweet. It is based on the inputs received from a Denis Biseko, and Mahjabeen Haji (Tanzania). Managerial guid- range of individuals and teams, to whom the team would like to ance was provided by Adrian Fozzard and Jim Brumby at the out- extend its profound appreciation. set, and Hisham Waly and Jim Brumby during finalization. Aart Individual case studies were developed by the following Kraay gave helpful guidance and advice for the quantitative analy- authors: Allan Gustafsson, Alexi Gugushvili, Verena Fritz, Liana sis. The team would also like to thank the formal peer reviewers Skhirtladze, and Heidi Tavakoli, with advice and comments Rachel Ort, Nicola Smithers, Manuel Vargas, and Tim Williamson provided by Aleksi Aleksishvili and Zaza Broladze of the Policy for their excellent comments and suggestions, and Richard Allen and Management Consulting Group (Georgia); Stephanie Sweet who provided earlier comments and feedback. Research assis- (Nepal); John Heilbrunn, Stephanie Sweet, Elijah Udo, Verena tance was provided by Matthew Collin and Maximilian Mareis. Fritz, as well as Stephen Jones, Gabriel Ojegbile, and Tochukwu The team wishes to thank all the government and nongov- Nwachukwu from Oxford Policy Management (Nigeria); Stephanie ernment counterparts met during the missions for their time and Sweet, Ronald Holmes, Marijn Verhoeven, and Allan Gustafsson generous sharing of information, experiences, and views, as well (Philippines); Helen Tilley, Revocatus Sangu, Kenny Manara, Sha- as numerous interlocutors in World Bank offices and from among kira Mustapha, and Stephanie Sweet (Tanzania). other development partners. Key collaborators from each of the country teams include: The team is grateful to the Governance Partnership Facility Mariam Dolidze, Inga Paichadze, and Mediha Agar (Georgia); and the Korean Development Institute Trust Funds for the funding Bigyan Pradhan and Sudyumna Dahal (Nepal); Ismaila Ceesay, provided. The World BAnk  iii Contents Acknowledgements iii Foreword vii Abbreviations ix Executive Summary xi Chapter 1: Rationale for the Report and Methodology 1 1.1 Approach and Methodology 3 Chapter 2: PFM Progress and Association with Country Characteristics—Key Findings from Analyzing Relevant Indicators 9 2.1 Variation in PFM Performance across Regions and Income Groups 9 2.2 Do Country Characteristics Drive PFM Performance? Key Findings from the Quantitative Analysis 11 2.3 Country Characteristics and PFM Performance Trends in the Five Case Study Countries 14 2.4 Chapter Summary 19 Chapter 3: Political Commitments and Fiscal Trends 21 3.1 Political Commitment to Reforms—What Would We Look For? 21 3.2 Overarching Policy Goals 25 3.3 Rents and Rent Distribution 26 3.4 Fiscal Trends—Limited Pressures to Reform 28 3.5 Chapter Summary 34 Chapter 4: Institutional Conditions and Dynamics for Advancing PFM Reforms 37 4.1 Institutional and Legal Arrangements 37 4.1.1 Public Finance Laws 37 4.1.2 Organization of ‘Central Finance Functions’ 38 4.1.3 Relationship between the Executive and the Legislative 40 4.1.4 The Importance of Intergovernmental Relations and of Effective PFM Systems across Levels of Government 43 4.2 The Political-Technical Interface for Driving Reforms 44 4.3 Demand Side—More a Potential than an Actual Driver for PFM Reform 46 4.4 The Role of Ideas and Ideology 48 4.5 Chapter Summary 49 Chapter 5: How Is Progress on PFM Being Made? 51 Reform Intentions, Steps Taken, and Results 51 5.1 Stand-alone Versus Embedded PFM Reforms 51 5.2 Reform Intentions—Similar Intentions Despite Diverging Problems 53 5.3 Partial Reforms as a Frequent Occurrence 55 5.4 Clarity of Expectations—What Are Specific PFM Reform Efforts Expected to Deliver? 56 5.5 Upstream Budget Planning: Getting the Basics Working, and Efforts at Introducing MTEFs and Program Budgets 57 The World BAnk  v 5.6 Progress and Challenges with Regard to Improving Budget Execution 60 5.6.1 Introducing Treasury Single Accounts 60 5.6.2 Introducing Financial Management Information Systems (FMISs) 64 5.6.3 Reforming Accounting and Fiscal Reporting Standards 68 5.7 Strengthening Public Procurement 74 5.8 Strengthening Internal Audit 77 5.9 Reforming and Strengthening the Role of External Audit 79 5.10 Chapter Summary 83 Chapter 6: Key Findings and Implications for What to Do Differently 85 References 97 Annex1: Data Source and Methodology 103 Annex 2: OLS and FD Robustness Checks 105 Annex 3: Framework Used for Developing the Country Analysis 107 vi  Political Economy of Public Financial Management Reforms Foreword E nsuring that public money is well managed has been a cen- financial management; and the role of citizens in demanding tral concern of citizens and the state since ancient times, reforms to ensure a better use of funds. The authors also exam- motivating an accumulation of technology and knowledge of ine the stated intentions of public financial management reforms how to arrive at well-functioning systems. Equally, it is incontro- across countries, and how these were pursued, or not, over a period vertible that public money plays a central role in the development of 10 to 15 years. Quite simply, what you see at the outset is rarely process, and that better management of public money is important what you get in the end. to achieving inclusive development. So, one could expect that bet- Over the past two decades, we have seen many approaches ter harnessing of public money is a shared objective across indi- to public financial management reform. For instance, reformers viduals, professions, and communities. have treated public financial management reform as theater, with Yet, in many contexts, the management of public money is an emphasis on shaping stakeholder perceptions and removing not driven by a concern for development impact and equity; but roadblocks through soft and harder methods. We have also seen rather, by concerns to provide some with preferential access to the reform process treated as an engineering project, with a tightly public money, while leaving less for the majority of citizens and defined critical path and intending to proceed along a sequence for achieving public policies. In other words, there are winners of pre-defined steps. This report draws conclusions about what and losers, and disputes over who takes the larger share of public is likely to work in different circumstances and why it is likely to money are common. work—based on close attention to how political, institutional, and While technology and knowledge about good management of fiscal drivers intersect. public money have come a long way, we do not always see these From a methodological point of view, the report provides a applied. Also, in many places we see slow adoption of better prac- strong empirical basis to our understanding through a combination tices, and repeated efforts, despite there being observable avail- of sophisticated econometric analysis with some of the nuance able capacity to do much better. and understanding that can only be gleaned from case studies. It In this report, rather than looking at the technical factors is a report that many have wanted to see for a long time; providing that make for better public financial management, the authors real insight and wisdom on this subject. Some of the results are look at the non-technical factors, that is the political economy of surprising and some are somewhat counter-intuitive, but they are the management of public money. Through country studies from all highly relevant for development practitioners and public finan- around the world, the report examines the level and durability of cial management reformers. political commitment to reform, and associated high-level policy Congratulations to the authors; I commend the work to you. goals, as well as the implications of underlying fiscal trends; insti- tutional and legal starting points and conditions, including the Jim Brumby relationship between the executive and the legislature in public Director Governance Global Practice The World BAnk  vii Abbreviations AFROSAI African Organization of Supreme Audit GDP Gross Domestic Product Institutions Ge-GP Georgian Electronic-Government Procurement APC All Progressives Congress GFS Government Finance Statistics BMIS Budget Management Information System GFSM Government Finance Statistics Manual BMPIU Budget Monitoring and Price Intelligence Unit GIFMIS Government Integrated Financial Management BoT Bank of Tanzania Information System BPP Bureau of Public Procurement Deutsche Gesellschaft für Internationale GIZ  BRN Big Results Now Zusammenarbeit (German Agency for International BTMS Budget and Treasury Management System Cooperation) BTr Bureau of Treasury GOCC Government-Owned and Controlled Corporations CAG Controller and Auditor General GPPB Government Procurement Policy Board CAO Central Accounting Office HIPC Heavily Indebted Poor Country CFA Central Finance Agency IAG Internal Auditor General CCM Chama Cha Mapinduzi (Party of the Revolution) IASB International Accounting Standards Board CIDA Canadian International Development Agency IFAC International Federation of Accountants CME Coordinating Minister for the Economy IFMIS Integrated Financial Management Information CMU Country Management Unit System COA Commission on Audit IMF International Monetary Fund COTS Commercial off-the-shelf INTOSAI The International Organisation of Supreme Audit CPAR Country Procurement Assessment Report Institutions CPIA Country Policy and Institutional Assessment IPSAS International Public Sector Accounting Standards CSO Civil Society Organization IT Information Technology DBM Department of Budget and Management LAC Latin America and the Caribbean DECS District Expenditure Control System LEPL Legal Entity of Public Law DoF Department of Finance LGA Local Government Authority DP Development Partner LGU Local Government Unit DTCO District Treasury Comptroller Office LIC Low-Income Country EAP East Asia and the Pacific LMBS Line Ministry Budget System EAMP East African Monetary Protocol LMIC Lower-Middle-Income Country ECA Europe and Central Asia Loi organique relative aux lois de finances LOLF  EFCC Economic and Financial Crimes Commission (Organic Law pertaining to Budget Acts) EU European Union MAPS Methodology for Assessing Procurement Systems FAAC Federation Account Allocation Committee MDAs Ministries Departments and Agencies FCGO Financial Comptroller General Office M&E Monitoring and Evaluation FCS Fragile and Conflict Situation MENA Middle East and North Africa FEC Federal Executive Council MIC Middle-Income Country FGN Federal Government of Nigeria MILF Moro Islamic Liberation Front FIRS Federal Inland Revenue Service MoF Ministry of Finance FMBNP Federal Ministry of Budget and National Planning MP Member of Parliament FMIS Financial Management Information System MTEF Medium-Term Expenditure Framework FRC Financial Reporting Council MTFF Medium-Term Fiscal Framework FTE Fiscal Transparency Evaluation NAO National Audit Office GAM Government Accounting Manual NCPP National Council on Public Procurement The World BAnk  ix NGAS New Government Accounting Standards PSASB Public Sector Accounting Standards Board NGO Nongovernmental Organization PWI Procurement Watch Inc. NPC National Planning Commission QWIDS Query Wizard for International Development NPI Nonprofit Institutional Unit Statistics NPM New Public Management RMIS Revenue Management Information System NPSAS Nepali Public Sector Accounting Standards ROSC Report on the Observance of Standards and Codes OAGF Office of the Accountant General of the SAI Supreme Audit Institution Federation SAM Social Accountability Monitoring OBI Open Budget Initiative SAO State Audit Office OECD Organisation for Economic Co-operation SAR South Asia Region and Development SBAC Special Bids and Awards Committee ODA Official Development Assistance SIDS Small Island Developing State OGP Open Government Partnership SIGMA Support for Improvement in Governance PAA Public Audit Act and Management PAC Public Accounts Committee SOEs State-Owned Enterprises PDP People’s Democratic Party SSA Sub-Saharan Africa PDIA Problem-Driven Iterative Adaptation TAN Transparency and Accountability Network PEFA Public Expenditure and Financial Accountability TISS Tanzania Interbank Settlement System PER Public Expenditure Review TRA Tanzania Revenue Authority PETS Public Expenditure Tracking Survey TSA Treasury Single Account PFM Public Financial Management UMIC Upper-Middle-Income Country PFMRP PFM Reform Program UNCITRAL United Nations Commission on International Phil-GEPS Philippines Government Electronic Procurement Trade Law System UNDP United Nations Development Programme PforR Program-for-Results USAID U.S. Agency for International Development PO-PC President’s Office—Planning Commission VAT Value-Added Tax PPA Public Procurement Act WAEMU West African Economic and Monetary Union PPAA Public Procurement Appeals Authority WDI World Development Indicator PPSAS Philippines Public Sector Accounting Standards WEO World Economic Outlook PPMO Public Procurement Monitoring Office WGI Worldwide Governance Indicators PPRA Public Procurement Regulatory Agency WDR World Development Report PSAO Provincial Satellite Audit Office x  Political Economy of Public Financial Management Reforms Executive Summary U sing fiscal resources to achieve results is critical for equi- stakeholders and external supporters can better calibrate their table development. Low- and middle-income countries approaches to PFM reforms. Regarding the latter, the intention is to face many critical expenditure pressures, from expanding start a discussion and provide an initial set of ideas and empiri- access to and improving the quality of services to improving infra- cally grounded insights, rather than presuming that a single effort structure to developing an ability to respond effectively to shocks can provide a definitive set of answers. such as droughts, natural disasters, or volatility in the prices for In line with the World Development Report (WDR) 2017, the key imports or exports. report also asks whether the improvements in PFM systems pursued Accordingly, many countries have sought to strengthen their had tangible impacts. Evidence on this is still difficult to assemble, public financial management (PFM) systems. PFM reform efforts are and one part of the recommendations emerging from this analysis widespread. As a rough indicator, nearly 150 countries had under- is to ensure that ‘functional progress’ made needs much greater taken at least one Public Expenditure and Financial Accountability efforts at monitoring. (PEFA) report as of early 2017, which serves to assess the status The report focuses on the expenditure side of public finances, of PFM, as well as typically leads to the preparation of (further) considering reforms across the budget cycle. The quality of PFM reform efforts. Development partners (DPs) support efforts at and changes over time constitute the main result to be explained PFM strengthening with about US$1.3 billion annually and nearly (‘dependent variable’), but as noted, comprising also questions US$20 billion in total since 2002, based on Organisation for Eco- about whether reformed systems are functioning as intended. In nomic Co-operation and Development (OECD) data. terms of explanatory factors, the analysis deploys a wide lens. It However, following nearly two decades of intensive effort, what starts off with considering potential factors for which data is avail- has been achieved, where, and why is not very clear. From an impact able across countries, including broad economic, fiscal, and politi- perspective, it is critical to understand whether efforts at reform- cal characteristics. ing PFM have been successful, how progress may have varied, Regarding the how and why of PFM reforms, drawing on five and what accounts for different rates of progress among different case studies, the analysis hones in on several key factors:first, lev- countries. els and durability of political commitment, high-level policy goals, and fiscal trends; second, institutional and legal starting points and conditions, including the relationship between the executive Approach and Intention of this Report and the legislature with regard to their roles in PFM as well as This report pursues two main lines: first, it maps out what PFM prog- whether citizens are demanding such reforms; third, it looks at ress looks like across countries, regions, and income groups; then, the specific reform intentions as set out in the PFM reform plans, based on this mapping, it drills down into specific experiences and and one by one at how these were pursued over a period of 10 to issues of how efforts at strengthening PFM have progressed or strug- 15 years. gled. The first aspect provides important insights as to where and Inevitably, there are limitations to the approach. The first limi- how much progress has been made among all low- and middle- tation is that including a wider range of countries as drill-down case income countries for which relevant data is available. The second studies would be desirable, but was not feasible in practice. A sec- aspect builds on this mapping and seeks to understand in detail ond limitation is that while information on overall official devel- what PFM strengthening efforts were made in a small N sample opment assistance (ODA) and ODA dedicated to PFM reforms is of countries and explores the underlying nontechnical drivers and considered, the process tracing conducted does not look at the constraints.1 interventions and support provided by various DPs in detail for The intention is for the resulting insights to provide a basis each country. Tracking such support comprehensively, project-by-­ for reflection, as well as some concrete guidance on how reform project for each country, would have absorbed a level of effort beyond what has been feasible.2 A third limitation is that it is not 1. Nontechnical drivers and constraints include institutional incentives and constraints, as well as political commitment, capabilities, and 2. The statistical analysis and broad country-by-country data on levels of demand to pursue reforms, interacting with institutional aspects; see also PFM support suggest that the level of effort by DPs to support or leverage Diamond (2011). PFM strengthening is not related to the degree of actual progress made. The World BAnk  xi always possible to trace in full detail who promoted or obstructed past 15 years.3 Legacies of instability have been greatest in Nepal progress of PFM reforms and based on what specific motivations. which experienced 10 years of conflict from 1996 to 2006,4 while The report documents some specific instances and hence pro- Georgia and Nigeria also experienced conflict, and the Philippines vides, at least selectively, an understanding of such specific stake- experienced it in certain regions. Natural resource dependence holder dynamics. has been high in Nigeria. In Tanzania, a resource boom was antici- pated, but large-scale offshore gas production remains delayed.5 The PFM performance of the five cases and trends over time Key Findings on Patterns of Progress offer interesting contrasts. Tanzania had a notably high PFM per- and Drivers of PFM Performance formance relative to income group peers in the early 2000s, but saw a decline over the following decade. Georgia took the opposite Across Countries path, from a low performance in the early 2000s to an overall good Aggregating data on PFM performance by income groups suggests that performance by the mid-2010s. In Nigeria, PFM reforms had been global progress on PFM reforms has been somewhat disappointing initiated, but PFM performance remained very low when the first in recent years, especially among middle-income countries (MICs), PEFA assessment was undertaken in 2012, and the Country Policy while low-income countries (LICs) have made some gains starting and Institutional Assessment (CPIA) data similarly does not yet from lower levels. Across regions, Europe and Central Asia (ECA) indicate an improvement over time (although some reforms were stands out as a group of mostly MICs that has seen significant implemented since 2015 as discussed in the report). Nepal and progress. In ECA, this includes strong progress on ‘accounting the Philippines both achieved some progress. and reporting’, the PFM dimension which tends to be the weakest Thus, the country characteristics that are statistically relevant among other countries. The progress registered by LICs is region- explain some of the specific cases’ individual status and trajecto- ally concentrated in Sub-Saharan Africa (SSA), where most cur- ries, while also leaving a number of puzzles. Change in Georgia was rent LICs are located. exceptionally rapid and has gone farther than regional and income During regression analysis of which countries performed better group comparators on average. Tanzania’s regressing trend would or worse, the following picture emerges:  the quality of PFM systems not have been predicted, given its continued growth and political is most significantly and robustly associated with two variables, a stability. Nepal’s progression is stronger than would be expected country’s income per capita (positively) and a high share of rev- given its postconflict situation and still low per capita incomes. enues being obtained from natural resources (negatively). In addi- For the Philippines, the overall observed strengthening of PFM tion, we find statistically weaker associations with being a small systems has been broadly in line with the country’s characteristics. island developing state (SIDS) (negative) and per capita growth, political stability, and having a more democratic regime (positive, but small coefficients). Programmatic parties appear to possibly Tracing the How and Why also have a positive impact. The significance and impact of these of Progress in Detail additional factors varies, depending on how PFM performance is Chapters three to five drill down into how these specific dynamics—in measured (PEFA or CPIA 13) and what observations are included how and how far PFM systems have strengthened—have come about. (all PEFA assessments or only most recent ones). The analysis proceeds from a ‘big picture’ perspective on political Looking at dynamics over time, that is, potential patterns commitments and fiscal trends to the institutional and legal situ- between characteristics and PFM improvements, the strongest associ- ation, and to the specific processes of pursuing individual reform ation is with the relative initial performance— that is, countries that ‘packages’ such as improving budget preparation, adopting new had an initially weaker PFM performance show relatively greater accounting rules, and establishing financial management informa- progress than those with already stronger PFM. Recent growth also tion systems (FMISs). shows a positive relationship, and higher population growth shows a negative one. In addition to the main results, some less robust associations appear with being a SIDS and changes in the level of tax revenue. 3. Georgia transitioned to lower-middle-income status in 2003 and to The five case study countries cover a diversity of situations in upper-middle-income status in 2015. Nigeria transitioned to lower- middle-income status in 2008. terms of income levels and other characteristics identified as associ- 4. Nepal was included on the World Bank’s list of fragile and conflict ated with PFM performance. Nepal and Tanzania have continuously situation (FCS) countries until FY2014. For IDA18, Nepal has been classified as “exceptional FCV risk mitigation regime” along with three been LICs, the Philippines has been an MIC, and Georgia and other countries (Niger, Guinea, and Tajikistan). Nigeria have moved from low to middle income status over the 5. Production is expected to begin in the 2020s. xii  Political Economy of Public Financial Management Reforms The analysis starts with three broad factors: the intensity and Institutional Factors durability of political commitment to pursuing better governance in general terms, the relative strength and nature of rent-seeking The potential role played by institutional factors is reviewed in Chap- interests, and the fiscal situation and trends for each of the five ter 4, indicating some important aspects of variation with regard to three countries over the period 2000–2015. While political commit- main institutional factors: the relationship between the executive ment is widely held to matter in PFM reforms, how it does so and the legislature regarding budget approvals as well as approvals exactly, or how it could be quantified has not been systematically of reform legislation, the setup of central finance agencies (CFAs), considered. Among the case studies, Georgia and the Philippines and the structure of intergovernmental relations. Having strong stand out as countries that had governments with strong commit- parliamentary powers over budgets can lead to a situation where ment to governance reforms during the period reviewed, with the budget approvals are significantly delayed, as has been the case in initial mandate being strongest in the former. Vested interests and particular in Nigeria, with negative impacts on budget execution, nonproductive rents have been particularly significant in Nigeria, while in Nepal, budget delays emerged due to extended political linked to natural resource production, but were also quite promi- deadlock. The setting up of CFAs has been particularly fragmented nent in Georgia in the early 2000s, in the Philippines, in the form in the Philippines and Nigeria. In the Philippines, which had a of a small economic and political elite, and in Tanzania, as a close government that was in principle committed to reforms, this has alliance between business and political interests formed in the contributed to challenges with getting reforms done. 2000s. Intergovernmental relations play a crucial role in terms of the Fiscal trends have been relatively benign during the period acceptance of PFM reforms as well as for getting from core PFM under investigation. All five countries saw relatively strong growth reforms to actual functional improvements. Remarkably, all five and declining debt-to-GDP ratios during the decade 2000–2010, countries either had federal systems in place (Nigeria) or had at in line with general trends across low- and middle-income coun- least some ongoing discussions over greater decentralization or tries. Nigeria and Tanzania additionally benefited from debt relief. introducing federal systems, potentially requiring substantial addi- Georgia and Nepal used this period to increase their revenue to tional investments in subnational-level PFM capacities (notably gross domestic product (GDP) ratios. Revenues flatlined in Tan- in Nepal). Effective oversight and accountability across levels of zania and the Philippines at around 15 percent and 18 percent government poses challenges, particularly so in highly decentral- respectively, with both falling behind the trends in their respective ized or federal systems. income groups toward increasing revenues (with global averages Regarding organic budget or PFM laws, three of the five coun- for LICs and LMICs reaching 21 percent and 29 percent of GDP tries had new comprehensive legislations under discussion, but expe- respectively). Nigeria experienced rising revenue in nominal terms, rienced difficulties in getting these passed in parliaments. These but had a steep and extended decline in its revenue-to-GDP ratio challenges are rooted in having constitutional arrangements more as GDP rapidly expanded due to rising oil prices, and a GDP re- prone to legislative deadlock and may also be rooted in the fact basing was undertaken in 2014, while few efforts were made to that some members of parliament (MPs) represent rent-seeking expand non-oil revenue collection. more than ‘good governance’ interests. Remarkably, over the Given the largely benign environments, the selected cases and past 15 years, all five countries passed new procurement legisla- period offer only limited insight into the question whether fiscal pres- tion, even though this is an area where rent-seeking has typically sures may stimulate PFM reform efforts. Some link can be seen for been concentrated. It appears that international pressures and the early years in Georgia, as it experienced serious cash shortfalls incentives for passing new procurement laws have been greater until the start of reforms which targeted both revenue increases and more focused than with regard to improved PFM legislation and better expenditure management. In Nigeria, fiscal pressures in general. While, as highlighted by WDR 2017, laws may not appear associated with a leap forward since 2015 in completing consistently be observed in practice, improved PFM rules are an the establishment of a Treasury Single Account (TSA) which had important foundation for seeking better de facto management of previously lingered for several years. In Tanzania, the particularly public resources. high PFM performance at the start of the period reviewed was the A further important aspect which Chapter 4 considers is the result of a previous reform episode in the late 1990s associated role of demand-side stakeholders in PFM reforms. The analysis with fiscal pressures and seeking to achieve heavily indebted poor across the five countries indicates that demand-side involvement country (HIPC) completion. was relatively stronger in the Philippines, but relatively weak in the four other cases. Budget monitoring activities by civil society The World BAnk  xiii organizations (CSOs) are present but have tended to remain frag- saw increases in revenue, have the best aggregate credibility and mented. PFM reforms are typically not well-known or of specific improved credibility in allocation across sectors. interests to citizens at large, and their specifics can be difficult to With regard to budget execution reforms, Chapter 5 consid- understand. The main link between citizens and PFM reform tends ers the introduction of TSAs, FMIS, and of new accounting and fiscal to be through the long route of accountability, that is, through reporting standards. All three of these reforms have been extremely electing governments that make electoral pledges to improve gov- widespread across countries, including being pursued across all ernance, including the use of public resources. five case study countries. TSAs are a fiscally attractive reform for central governments, but are generally not attractive for line ministries and subnational Tracing Individual PFM Reforms governments (if they are to be included); a further potential source of Chapter 5 turns to the specific aspects of what and how progress on resistance are commercial banks which lose valuable liquidity. While, specific PFM reforms were pursued. It maps out to what extent the in principle, a relatively simple reform which requires limited reform intentions across the cases were differentiated or similar, capacity strengthening or systems rollout, a lack of trust among and finds a strong tendency toward a similar ‘menu’ of reforms. stakeholders on the one hand, and various interests to maintain Furthermore, important cross-cutting features are that reforms tend separate accounts on the other hand, make TSAs a reform that to remain partially completed in many cases for extended periods has remained partially implemented in several of the countries of time, and that the advocated expected impacts of reforms tend examined. Georgia and, to some extent, Nepal pursued a deliber- to exceed the actual impacts of such reforms across countries. The ate gradual strategy with good results. In the Philippines, a TSA chapter, furthermore, explores the extent to which PFM reforms was pursued, but still only extends to revenue accounts. Nigeria are embedded in a wider set of public sector reforms. Pursuing a experienced a particularly dramatic process. While the TSA reform broader governance agenda can be a signal of stronger government effort was first started in 2010, by early 2015, MDAs still held the commitment in achieving deeper changes in outcomes such as equivalent of US$11 billion in accounts with commercial banks. improving service delivery and in containing corruption. Among a brewing fiscal crisis due to fast shrinking oil revenues, Reforms to budget planning have targeted the introduction of the newly elected President ordered all MDAs to comply. The large- Medium-Term Expenditure Frameworks (MTEFs) and of developing scale transfer of funds triggered a public row, as the transferring program budgets, but tend to neglect some of the aspects that are MDAs were required to pay a 1 percent fee to an electronic plat- most critical to preparing and approving credible budgets on time. form channeling the funds to Central Bank accounts (e-Remita), The introduction of MTEFs has been pursued across all five coun- that is, around US$100 million, with the fee to be split between tries, and program budgeting across three of the five. In terms of the provider, commercial banks, and the Central Bank. Thus, seek- reform achievements, the MTEFs are seen as reasonably estab- ing to compensate one opposed stakeholder, that is, commercial lished and effective in Georgia and the Philippines, but in the banks, the government triggered new allegations of improper man- other three countries the reform remains partially implemented agement. Still, TSA rollout was judged to be completed by the and has at best limited impact on actual medium-term budget International Monetary Fund (IMF) by early 2016. allocations; in the case of Tanzania, this is so even after nearly two FMIS reforms—which tend to be relatively expensive and con- decades of producing MTEF documents. Program or performance sidered challenging in terms of change management—proceeded in budgeting has been pursued since 2006 in Georgia and more all five countries, but (a) experienced a number of challenges and recently in the Philippines and Tanzania; so far, real impact has (b) had significantly less impact than anticipated. After considering a remained limited. In Georgia, demand either from the parliament commercial off-the-shelf (COTS) software system, Georgia decided or from the wider public has remained weak, and so the incentives in favor of gradually developing a system in-house, an unorthodox for ministries, departments, and agencies (MDAs) to demonstrate approach which worked reasonably well given its strong commit- good performance are limited. ment and an ability to mobilize substantial information technology Considering the ‘basics’ of timely budget approvals and the (IT) capacities. In Nepal, the establishment both of a TSA and a expected outcomes of budgets that are credible, progress is still quite rudimentary (and still fragmented) FMIS took a deliberately grad- limited. Timely budget approvals are common in Georgia, the Phil- ual route, in the absence of much support, but also limited oppo- ippines, and Tanzania, while delays have remained substantial in sition from national MDAs. Tanzania has the longest established Nigeria. In Nepal, eventual constitutional reforms enabled timely FMIS, which was introduced during its main earlier wave of PFM approvals in the most recent years. Moreover, credibility in sectoral reforms which started in the late 1990s. While on the one hand, budget allocations remains low in four of the five countries, and it demonstrates that an FMIS could be successfully introduced even aggregate credibility is low in the Philippines and, to a lesser in a low-income environment, it also shows most clearly that the extent, in Nigeria. Georgia and Nepal, the two countries which system has not fulfilled expectations in terms of ensuring effective xiv  Political Economy of Public Financial Management Reforms commitment controls and credible and accountable budget execu- offices saw greater development than either internal audit or par- tion. In the Philippines, the government pursued ambitious plans liamentary follow-up and oversight. On the latter dimension, none of introducing a nationwide FMIS. However, following a drawn- of the five countries was rated better than a D+ by the most recent out process of system specification, the procurement eventually PEFA assessments. These continued low ratings reflect the general spectacularly failed, due to the refusal of the President to approve disincentives for MPs to pay attention to how budget funds were the contract award shortly before the end of his constitutionally used ex post, as well as some specific factors such as the absence mandated single term. A much smaller core system is now being of a regular parliament for several years in Nepal. pursued, likely to provoke less opposition. A third element of budget execution reforms is the introduction of accounting standards following International Public Sector Account- Key Implications and ing Standards (IPSAS) guidelines—intended to be accrual-based in Recommendations Georgia, Nigeria, the Philippines, and Tanzania, and cash-based in The analysis leads to several implications and recommendations, Nepal. This transition requires a substantial re-tooling of account- intended both to be immediately useable as well as stimulating a ing capacities, as well as valuing a wide range of public sector wider discussion on the what and how of PFM reforms. assets and liabilities. Of the five countries, Tanzania has been the first to actually start the use of IPSAS, first adopting the cash- • A perspective on nontechnical drivers for PFM reforms suggests based, and most recently accrual-based standards. Introduction a need to reconsider and nuance arguments about the sequenc- plans are the most ambitious in Nigeria, which seeks to introduce ing of PFM reforms. Improving ‘the basics’ is critical in terms IPSAS across all three main levels of government—in contrast to of key expected impacts of PFM reforms. However, it is also all other PFM reforms which are separately introduced at the fed- the relatively harder aim to achieve. Including ‘nontechni- eral and state levels. Given the reform status, it is as yet too early cal drivers’ into sequencing considerations implies that it is to assess impact. A key concern is that accrual standards may important to continuously keep some focus on the basics, not be accurately applied—intentionally or unintentionally—and without assuming that these are likely to be completed that oversight will be insufficient. Without adequate explanations, before advanced reforms are initiated. accrual accounts may also be more challenging for demand-side Furthermore, since de facto, real sequencing between stakeholders to review and interpret. basic and advanced reforms has rarely happened consis- In addition to these core budget execution reforms, procure- tently, many countries present a mixed challenge of various ment stands out as an area in which the five countries have sought partially implemented bits of both types of PFM reforms. reforms, although this is a politically and particularly charged expen- Reformers and DPs should keep track of where ‘basic’ diture aspect. The use of public contracting to divert resources for functional improvements stand and how they can be con- political financing as well as personal enrichment is widespread tinuously supported, also with a view to avoiding backsliding globally. All five countries adopted new procurement legislation from improvements that have already been achieved. This over the past two decades. Different sequences were deployed includes paying attention to whether ‘advanced reforms’ in terms of establishing e-procurement platforms and e-bidding that have been (partially) implemented are contributing to as an option for suppliers. In none of the countries are procure- basic functionalities (such as whether FMIS is actually used ment rules and processes fully settled due to a range of remaining for controlling commitments). concerns. Also, remarkably, there is little existing evidence about whether value for money has improved. • While transformational—rapid and substantial—PFM strength- The final two substantive sections of Chapter 5 address internal ening is rare, slow and incremental improvements are com- and external audit. In principle, establishing effective audit systems mon, but risks of backsliding also exist. As the quantitative is not very attractive from a political economy perspective, in the analysis indicates, countries with initially lower PEFA scores sense that effective audit capacities reduce the discretion of deci- saw greater improvements over time than those with already sion makers. All five countries have struggled to strengthen these better systems, however, the gains are relatively small over dimensions, but there has also been some progress. In Tanzania, a 10-year period. At the same time, global indicators also while several other PFM reforms saw backsliding, the capacity suggest limited aggregate improvement due to cases of and efficacy of the external audit office improved, and this also backsliding; illustrated among the case studies by Tanza- triggered efforts at and some progress with strengthening internal nia’s experience. Efforts at strengthening PFM systems audit. Georgia experienced a remarkable turnaround from an initial need to be calibrated to these different ‘opportunity envi- control chamber which was itself bribing audited agencies to a ronments’. This should include having some in-built flex- fairly capable State Audit Office (SAO). Generally, external audit ibility to change between a more incremental and a more The World BAnk  xv fundamental reform agenda, that is, following changes in • Addressing and engaging with stakeholder interests more government. During non-transformational periods, a deliber- explicitly. An understanding of the motivation and aims of ate focus on (a) developing a reform agenda and (b) moni- stakeholders is important to understand what reforms are toring and seeking to counteract backsliding risks should be likely to be feasible and ‘stick’. In this regard, it is impor- pursued. tant to avoid overly simple categorization of stakeholders into ‘reform proponents’ and ‘reform opponents’, as most • A further crucial implication from the analysis is that windows stakeholders have somewhat complex motivations, for exam- of opportunity are very important—but it can be difficult for ple, they may endorse some reforms, but may be cautious DPs to effectively provide support to incoming, highly reform- about pushing too far in a difficult environment, such as oriented governments. The opportunities for PFM reforms can with regard to procurement reforms. vary considerably between one government and the next. A government that is strongly interested and motivated to pur- In addition, motivations and interests are likely to differ sue reforms can achieve more in two to three years than it across categories of stakeholders. Decision makers such is likely to be achieved with a government not interested as Prime Ministers or Presidents may be particularly inter- in improving governance and the management of public ested in PFM reforms if they have made electoral pledges to finance over a period of eight or ten years. improve service delivery and governance and seek to show progress. Senior civil servants and professionals will often • One important question is how best to use windows of be keen to show that they are aligned with the ‘vision at the opportunity—developing new support within a few months from ­ top’.6 At the same time, they play a critical role in shaping new leadership being in place to the closing of the peak ‘honey- the specifics of PFM reforms; and it can be important to moon period’ is difficult. One option can be to have an exist- discuss and mull over the specific reform approaches they ing project in place that straddles an election period with seek to pursue.7 Stakeholders in line ministries and at sub- sufficient built-in flexibility to be adjusted to a new govern- national levels are a further important group to consider and ment’s needs and intentions, either through a rapid restruc- engage with. One important aspect for reform stakeholders turing and reauthorization process, or through within-project is that initial progress can help to build the credibility of flexibility. More ‘advanced’ options such as a pooled mecha- reforms. Especially at early stages, reform stakeholders may nism with flexible funding can also be envisaged, but none want to be careful about reform approaches that entail very of the five case studies offers an example in this regard. long planning phases or that are very complex or likely to • For the majority of time outside of windows of opportunities, it face stronger resistance. remains worth pursuing improvements, but progress is likely to be incremental and frustrating to those seeking a faster • Being clear about institutional arrangements and roles is essential for assessing bottlenecks and likely difficulties, as pace and tangible results. A key potential benefit is to ‘have well as for identifying priorities for engagement on institutional something ready to move’ when windows of opportunity arise changes. Institutional arrangements, including the setup of and—also importantly—to sustain some of the gains that CFAs, the executive-legislative relationship and powers over may have been made during previous ‘windows’. fiscal and budget matters, and intergovernmental arrange- • Fiscal pressures can play a role in accelerating reforms and ments including those for oversight and accountability, the links between fiscal trends and PFM reforms should receive are critical and can enable or pose significant obstacles greater attention. When governments are motivated to pursue to reform progress. On the one hand, coordination chal- PFM reforms due to fiscal challenges, they are likely to seek lenges, challenges of parliamentary approval, or subnational reforms both on the revenue and expenditure side, with impli- resistance should truly be taken into account when consid- cations for the design of support. To overcome a fiscal crisis ering reform strategies. On the other hand, it can also be typically requires measures on both sides, and attention to revenue trends should be reflected routinely in PFM opera- 6. Conversely, these stakeholders typically would be reluctant or have tions (for example, the decline in revenue was neglected in at best limited ability to ‘swim against the tide’ of top-level leadership Nigeria for some time before the fiscal situation became preferences. 7. This report has not explicitly explored whether particular organizational severe). Governments are likely to seek support on multiple arrangements, such as creating dedicated reform units, have been aspects, including debt management, fiscal policy, revenue associated with better results. From a political economy perspective, administration, and expenditure management. This poses a particular issue of interest is whether more committed governments achieve more when adopting certain mechanisms such as reform units some challenges for how (at least broadly) coherent support and/or interministerial steering committees. These are issues for further can be provided. exploration in future analytic work. xvi  Political Economy of Public Financial Management Reforms important to consider whether seeking to change existing • As international norms and standards are promoted, it is also institutional arrangements is essential to be able to achieve critical that national stakeholders—accountants, auditors, tangible progress. Whether this is needed, and whether nongovernmental organizations (NGOs), or staff of parliamen- and when a window of opportunity might arise can only be tary budget offices—have training opportunities to understand judged on a case-by-case basis, but should be considered if these in depth. This is critical for enabling such stakeholders there is evidence that progress has gotten stuck due to such to have an informed discussion both about how principles institutional challenges. and standards should be adopted and adapted into national systems, as well as to judge whether applications once • Clearly identifying priority functional gaps is important, as is introduced are done appropriately. Conversely, if standards being realistic about what instruments are likely to deliver are widely promoted, but with little check on how they are which kind of improvements. The need to focus on functional applied in practice, this can incentivize mimicry, that is, an improvements is highlighted in WDR 2017. Functional gaps imitation of standards without a real intention or capacity to such as effective controls against leakages and excessive use these toward intended effects. commitments, as well as ensuring that funds are spent in a cost-effective way remain substantial in most countries. It is • Among the cases studied, especially ‘short route’8 demand-side critical to identify what key gaps are at the outset of devel- factors have been rather weak; nonetheless it may still be bene- oping reform efforts and operational support. While coun- ficial to explore further how these can be strengthened. The key terparts may be more used to developing lists of technical finding that citizens and civil society have not been strongly assistance proposals, many will appreciate a more function- involved in, demanding, or encouraging PFM reforms raises ally focused and ‘problem-driven’ discussion about what are the question whether stakeholders seeking to pursue and the most urgent issues in PFM and how these affect service support such reforms should encourage and stimulate citi- delivery. zen engagement. A key challenge is then to prioritize what can be done and how. • Overall citizen expectations are crucial for electing governments For example, to ensure that a medium-term perspective is promising to deliver governance reforms as has happened in really integrated into the annual budget preparation process Nigeria and Tanzania in 2015, and in Georgia since 2004, creat- or that excessive commitments are monitored and trigger a ing a ‘long route’ of demand for improving PFM. However, since leadership reaction if needed. It is critical to move beyond such a ‘long route’ is not very specific, complementing it by the facile assumption that simply introducing a certain stronger specific citizen interest and demands regarding good reform instrument will trigger the expected change in the use of public funds is potentially valuable. PFM results chain. In principle, citizen and media interest has the potential • Moreover, it is important to be more realistic about what func- to incentivize political commitment to sustain PFM reforms tional improvements a given instrument is likely to deliver given and make them impactful. However, to make this hap- continuous challenges. This requires some honest stock-­ pen, better information and communication is likely to be taking of why bottlenecks to certain functional improve- required: Among governance reforms and compared to other ments exist. As part of prioritizing which functional gaps aspects of economic governance such as taxation, PFM to pursue and which instruments to utilize, it would also be reforms are among the most obscure for citizens, both in desirable to develop a more systematic understanding of the terms of happening ‘inside government’ and being difficult costs of various PFM reform packages. to grasp. Reformers and DPs can do much more to open up PFM reforms and to explain what is being pursued and Regarding risks of ‘isomorphic reforms’ and ‘isomorphic mim- why in plain language and concepts. Balancing revenue icry’, DPs may need to become more reflective of the interna- and expenditures and keeping track of how much funds are tional norms and standards they promote. Setting international spent and on what are concepts that are in principle quite norms and encouraging their adoption is not good or bad per clear to any household. se, but depends on what these target and how they are used. A political economy perspective suggests that it is most At the same time, it is important that citizens are enabled to important to have international norms for those issues that see whether funds are used well and responsibly, rather than go ‘against the grain’ of key national stakeholders—such engagement primarily being channeled toward demanding as establishing an independent external audit institution or improving procurement—as these can help national reform 8. The ‘long route’ and ‘short route’ refer to the two stylized paths for stakeholders to leverage improvements. citizens to influence state action, as set out in WDR 2004. The World BAnk  xvii more funds for a particular group or locality, but not being and its direct follow-up with audited agencies significantly able to see the bigger picture. The emergence of more NGOs compensate for the lack of a direct link to the legislature specialized on PFM issues as well as caring about sector and of parliamentary follow-up. performance, combined with the spreading of provisions in favor of fiscal transparency, are opportunities for moving in • Tracking what functional improvements are being made and sustained is critical in terms of incentivizing real reforms, as this direction. As the Philippines experience indicates, it is well as for planning further steps. Tracking of PFM systems important that fiscal transparency involves not just making and functions over time provides a chance to understand information available, but also coherence and clarity in the what has been achieved, to make corrections where nec- accounts presented. essary, and to plan further reform efforts. Assessments • Should reform stakeholders deliberately target partial reforms focused on expected outputs and outcomes, are critical to and unorthodox approaches? As is covered in Chapter 5 of assess to what extent the introduction of particular instru- this report, many PFM reforms remain partially completed ments or ‘forms’ are actually associated with functional for many years. There appears to be some benefit to intro- improvements. Government leaders and specific institutions ducing reforms deliberately in an incremental way rather tasked with undertaking PFM reforms are more likely to care than seeking all out comprehensive reforms which are at a about real impacts if these are being tracked (and results higher risk of failure, as has happened in some of the cases are made public). Without such tracking, pursuing façade covered. Overall, it is likely that for countries and time peri- reforms remains more attractive. ods that are not windows of opportunity, deliberately target- Repeat PEFA assessments have become one important ing partial reforms is a smarter approach than seeking more instrument for tracking what progress is actually being made; comprehensive reforms that get stuck. but they leave some important gaps which would need to be A related question is whether to consider ‘unorthodox filled by complementary efforts. For example, PEFA assess- approaches’, that is, approaches to PFM reforms that con- ments cannot assess in detail whether accounting standards tradict conventional notions of ‘best practice’. Both Georgia are actually applied as intended, whether procurement and Nepal offer examples of rather successfully introducing reforms have actually resulted in greater value for money, ‘unorthodox’ approaches to treasury automatization. Some or whether funds reach frontline service delivery units more may consider ‘unorthodox’ approaches as inherently attrac- reliably. Assessing these functional improvements requires tive, in line with criticizing ‘best practices’ as typically lack- complementary efforts, but these have been made only to a ing a good enough fit and promoting isomorphic mimicry. rather limited extent in recent years in most countries. However, there are also some important caveats. Certain ‘unorthodox solutions’ such as nationally programmed rather • Reform support tends to be focused on moving from one set of issues to the next, without sufficiently considering full implemen- than ‘off-the-shelf’ IT systems can be costly and become tation and sustention over time. New processes take time to be failures if national IT capacity is weak, or if the firms or fully embedded and routinized, thus leaving them at consid- staff to develop these are not selected based on meritocratic erable risk of backsliding for a considerable period following criteria in a weak governance context. the initial introduction. This can happen, in particular, related Thus, while unorthodox solutions can be attractive in terms to changes in government. Therefore, targeting the sustaining of offering a good fit with specific reform needs and oppor- and continued implementation and ‘use as intended’ of new tunities, they require due consideration of potential risks systems should receive as much attention from DPs support- and downsides and how these stack up relative to pursu- ing PFM strengthening as the introduction of new systems as ing a more standard approach. A significant benefit of con- such—even if a flatlining of results indicators over a period sidering unorthodox solutions is the opening up of options of time looks less attractive relative to defining progressive relative to exclusively considering a binary choice between rollout indicators for a new area of reforms. leaving things as they are and pursuing all out best practice In addition, the report sets out the following specific impli- reforms for a given aspect of PFM. cations for operational designs: Thinking beyond the box of standard reform approaches 1. It is critical to explore with counterparts what they perceive as should be encouraged considering what is technically and their core problems and how they seek to address them. While politically feasible and likely to have at least a limited but there may be different views and not a complete or easy set tangible impact. For example, in the Philippines, the right of solutions that emerges, such an approach should help of the Central Accounting Office (CAO) to publish its reports avoid just introducing another reform tool. xviii  Political Economy of Public Financial Management Reforms 2. In the same vein, raising the question of potential blockages. more broadly is more appropriate when there is less clarity As we increase knowledge from various countries about or agreement among stakeholders, and therefore less clarity blockages—for example, subnational governments/power ­ about where progress is most likely. The overall principle holders, line ministries—we can probe whether such block- of focusing on functional improvements and sustainability ages are likely to play a role in a given context and discuss rather than on introducing various new tools still applies, what this implies for reform design. even when pursuing a broader agenda. 3. Understanding and discussing with the client what has gone 8. In addition, PFM operations must pay greater attention to the wrong or has not worked as expected in the past and why. Much intended results chain, and in many situations this involves better knowledge about ‘what has gone wrong where’ can some operational engagement on how funds move across levels also help teams test what risks might exist based on experi- of government to frontline units. Where core systems are a ence from other countries. full-blown bottleneck, this may be premature, but in most 4. Probing the views and buy-in from stakeholders beyond core situations, the flow of funds, as well as of performance infor- CFAs is critical for most PFM reforms. As the experiences mation across the expenditure chain is important to con- discussed indicate, resistance to reforms can come from sider to ensure that operations actually enable the improved parliaments, subnational levels, as well as banks or other functioning of governments. stakeholders that are often not fully considered. Subnational 9. With regard to investing in embedding already existing or governments and frontline spending units may have impor- recently established systems versus breaking new ground, tant perspectives on what the key bottlenecks are toward greater attention should be focused on the former. After one or ensuring that PFM contributes to better service delivery. two decades of PFM reforms being pursued in most coun- Broad enough buy-in is also important within CFAs, as in tries, embedding systems and ensuring that they work as many countries, ministers come and go, but at least core intended, or adapting them to move functionality forward staff often stay. is critical. Certain expansions in functionality may be (very) 5. Encouraging frank discussions between teams, practice manag- difficult given political economy constraints, but it can be ers, and Country Management Units (CMUs) about whether an important to probe these boundaries, as often at least some engagement is likely to be able to use a ‘window of opportunity’ further gradual improvements are feasible. or is more intervening during a period of relative stagnation. 10. Investing more in monitoring whether systems are actually used This should help to set realistic expectations. It also offers as intended is critical. Reforms are more likely to be pursued an early opportunity to consider potential points at which it with a view to generating functional improvements if these are may be opportune to restructure a project, for example, if actually monitored. Monitoring the effects of reforms can be a a window of opportunity might arise following an election. direct operational target, or (more weakly) can be built into the 6. Contributing to setting realistic expectations rather than ‘will monitoring and evaluation (M&E) framework of an operation. work as advertised’. Especially in a Program-for-Results 11. Considering the interface with other ongoing public sector (PforR) design, there is the scope as well as a critical impor- reforms and with other DPs. Especially for reforms such as tance to be realistic about what particular PFM reform tools program budgeting, there are often critical links to other will achieve. ongoing reform efforts to consider, such as the establish- 7. Among operational design choices, one key choice is whether ment of central M&E functions, or the introduction of perfor- to include a range of components or seeking to be (truly) selec- mance management tools for civil servants. A performance tive. In many situations, there are more challenges than can management system that actually makes a difference is be reasonably worked on by one project on the one hand, more likely to emerge if these different aspects establish while multiple donors are engaged on the other hand—often coherent incentives and practices. A second interface that is involving some degree of friction and inconsistencies. Espe- often very important is an interface with other DPs working cially in more challenging situations, an ideal approach on PFM reforms. could be a ‘pooled fund’ for PFM reforms that can respond 12. Raising ‘blind spots’ with counterparts and with other DPs work- with some flexibility to a range of needs—for example, ing on PFM. When institutional fragmentation or outdated developing better rules and systems, developing meaning- legal provisions are a true bottleneck to moving PFM reforms ful training systems for PFM related staff, and so on—in a forward, World Bank Group teams can potentially play an reasonably coordinated way. important role in convening discussions on reforms, also If stakeholders know and agree on what they want and bringing expertise and options from other countries to bear. how they will pursue this, a corresponding and relatively nar- The report concludes with outlining selected areas for fur- row design is likely to work best; while starting somewhat ther analysis. The World BAnk  xix Rationale for the Report and Methodology 1 P ublic financial management (PFM) reforms are widely seen as 2010 when it reached US$1.9 billion and has somewhat declined essential for achieving several aims:ensuring that a country’s since then (Figure 1.1). At the same time, support is evolving, own public resources as well as aid can be accounted for, expanding from reforms that primarily target public expenditure optimizing effectiveness and efficiency of spending across pub- systems to adjacent areas, including reforms of state-owned enter- lic sector tasks, and maintaining fiscal stability and appropriate prises (SOEs), procurement reforms, and reforms of revenue policy management of public assets. PFM reforms are one pillar of over- and administration systems. all fiscal reforms that countries have to periodically implement Reform packages targeting PFM reforms remain rather homoge- as they seek to undertake a growing range of tasks within often neous. They include a list of around 10 items: Medium-Term Expen- expanding, but still inevitably limited fiscal envelopes. Conversely, diture Frameworks (MTEFs) and program budgeting to achieve a shortcomings in PFM systems can lead to lack of fiscal discipline better policy orientation of budgets, introduction of new budget and macroeconomic instability, weaken the alignment between the classifications and accounting standards (including in many coun- allocation of public resources and national policy priorities, and tries, the ambition to shift to International Public Sector Account- contribute to greater waste and corruption in the delivery of public ing Standards [IPSAS]), establishing and upgrading information services. technology (IT) systems for managing public expenditures, better Because of the centrality of PFM systems, donors have been cash management, including the introduction of Treasury Single actively engaged in supporting the reform and strengthening of these Accounts (TSAs), strengthening internal audit and external audit, across many developing countries. The average annual official and ex post accountability. Reform packages typically focus on a development assistance (ODA) spending on PFM reforms since the selection of this set and, in some cases, most or all of these areas. early 2000s has been US$1.3 billion, with spending particularly The results of PFM reforms remain, however, very diverse. high from multilateral institutions. Support peaked in 2009 and Some countries have made rapid progress in recent years, while Figure 1.1.  ODA for PFM, 2002–2015 2500.0 2000.0 Million USD 1500.0 1000.0 500.0 0.0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: Organisation for Economic Co-operation and Development (OECD) Query Wizard for International Development Statistics (QWIDS) 2017.   1 (similar) reform agendas are stagnating in others. Data from repeat less strong than expected. Jointly, all the factors tested explain Public Expenditure and Financial Accountability (PEFA) reports about 40 percent of the observed variation (based on PEFA and indicate overall progress. More countries progress than regress, Country Policy and Institutional Assessment [CPIA] data to mea- and the range of indicators along which progress happens is wider sure a country’s quality of PFM systems). Figure 1.2 reflects the and more frequent. However, within this overall diagnosis, there is considerable divergence from the trend line when mapping per great variation. Countries tend to progress on some indicators but capita incomes and the quality of PFM systems. There are several not on others, and in many cases, repeat PEFA assessments record countries at various income levels that are well above, and others both progress and slippages. that are well below the line; and the variation is observable for The combination of considerable international efforts, countries from various regions.10 relatively standardized reform approaches,9 and yet quite widely A further possible explanation is that progress depends on the diverging outcomes raises the question of why this is the case. level of external support. However, as de Renzio, Andrews, and A first potential hypothesis is that differences in how much PFM Mills (2011) have shown, the divergence in PFM systems perfor- strengthening achieves largely depends on a country’s overall level mance is only weakly related to the degree of ODA support for PFM of development. In principle, countries that are wealthier are likely reforms. Our analysis in Chapter 2 confirms this finding. to have better preconditions—typically, they have a more educated In addition, there is concern about the fact that reform public sector workforce, better IT capabilities, and more experi- approaches are not sufficiently targeting functional improvements and ence with complex management tasks. that they do not sufficiently involve iterative problem solving (Andrews The quantitative analysis undertaken for this study indi- 2013). Such concerns are also consistent with the emphasis in cates that while there is some relationship between income lev- the World Development Report (WDR) 2017 that formal or de jure els and the relative strength of PFM systems, the relationship is changes may not be equivalent to real functional improvements Figure 1.2  Levels of GDP and PFM Performance PFM Quality and Income 3.5 ZAF MDA GEO KSV BFA RWA ALB AZE PER BTN Most Recent PEFA Score ARM CPVMAR 3 ETH CRI PHL IDN ECU MNE HND UZB TUN MOZ MLI PAK SLV MKD KGZ UKR COL BWA VNM JOR SRB MNG AFG NPL PRY MUS TMP WSM 2.5 LBR UGA BLR TJK STP ZMB NIC TON DOM TZA VUT BGD LKA MWI BOL GTM NAM SLB CIV KEN ZAR NER SLE WBG SWZ DMA BDI ZWE GMB KHM JAM SEN TUV GRD PAN GHA VCT 2 FJI GAB LAO BLZ LCA TGO GIN BEN LSO MRT SSD MDV SUR FSM CAF MDG PNG LBN MMR COM COG 1.5 MHL GNBHTI NGA 6 7 8 9 10 Log GDP per capita (PPP, lagged 5-year average) EAP ECA LAC MENA SAR SSA Note: (a) ‘PEFA score’ is the average of 21 indicators (PI-5 to PI-12 and PI-16 to PI-28 as per the 2011 PEFA methodology), that is, we exclude PI-1 through PI-4, which measure PFM outcomes, indicators PI-13 to PI‑15, which cover revenue administration, and D1 to D3, which are donor-related indicators; (b) scores from the ‘most recent’ assessment for each country were utilized (for example, Afghanistan has a 2005, 2008, and 2013 assessment; only 2013 is included here); (c) 13 observations were dropped; and (d) high-income countries were excluded. 9. For the five case study countries, reform approaches are discussed in 10. The only region for which most countries are on one side of the trend detail in Chapters 4 and 5. line is Europe and Central Asia (ECA) (above the trend line). 2  Political Economy of Public Financial Management Reforms (World Bank 2017). The considerable homogeneity of reform and often, also some of the micro-issues, such as the political intentions—also reflected in the five case studies on which this and other relevant ties of the Minister of Finance and possibly synthesis builds—suggests that there may well be problems with of other key government staff. However, overall, there is typically adequately tailoring reforms. However, countries which made sig- little systematic analysis of the wider political economy drivers and nificant progress on PFM reforms appear to also have pursued dynamics affecting the prospects for PFM reforms. Discussions relatively standard reform approaches. of political economy and related nontechnical drivers are also A further hypothesis is that there are considerable nontechnical not systematic in project documents. They are most commonly factors that play a role in influencing whether a government man- addressed in the risk section of program assessment, not always ages to pursue and achieve reform progress.11 These factors may with a clear or substantial link to project design. also account for whether governments are interested in pushing The intention of this report is to distill insights from a set of to solve particular functional bottlenecks, or are more likely to empirical work on potential nontechnical drivers of PFM reforms and halfheartedly pursue solutions that are perceived as ‘international to propose a set of potential implications. The work is based on a standards’. Traditionally, such nontechnical factors have been des- quantitative analysis on the patterns of progress with PFM reforms ignated as ‘political will’ to pursue reforms, and the importance and correlations with various country characteristics, as well as of such ‘will’ is widely assumed among practitioners. As Allen, five in-depth case studies (see further detail on the methodology Hemming, and Potter (2013) have emphasized, “PFM is no longer in Section 1.1). viewed a purely technical finance and accounting topic (as it once Based on the insights gained, Chapter 6 sets out implications in was); rather, it has become a subject where institutions and politi- the form of proposals for what could be done additionally or differently cal factors play an important role.” And further, “We anticipate to render approaches and support to PFM reforms more effective. The that, over the next ten years, the importance of political economy report hopes to stimulate both debate on what extent such aspects analysis as applied to PFM will continue to grow both as an area of can be integrated into PFM reform efforts and experimentation research and in its practical application.” (Allen, Hemming, and with actually doing so. Potter (2013), 6). Rather little effort has been made to-date to really explore such factors empirically, in particular for PFM reforms in low- and middle- 1.1 Approach and Methodology income countries. There has been some discussion about the pos- The focus of this report is on how public expenditures are managed, sible importance of ‘political economy’ and related nontechnical that is, the spending and utilization rather than the collection side factors for PFM reforms (Diamond 2012; Wehner and de Renzio of public resources. The two areas are linked; however, it would 2013). An earlier literature explored political economy drivers of exceed a reasonable scope of analysis to treat both in one report. budgeting and fiscal policy focusing on OECD countries (Wildavsky That said, this analysis includes some discussion of how efforts 1986; Von Hagen and Harden 1994). Furthermore, a small num- at strengthening revenue collection and expenditure management ber of studies describe political economy factors underlying dys- are linked. It addresses this in the quantitative analysis reflected functionalities in budgeting in individual low- and middle-income in Chapter 2, as well as where relevant in the discussion of the countries (Killick 2005 on Ghana and Rakner et al. 2004 on case studies. While this report can touch on whether or not gov- Malawi). ernments have tended to pursue both revenue and expenditure This report is an effort to address this gap and to begin devel- reforms simultaneously, going into detail on how revenue collec- oping a more systematic empirical exploration. The purpose is to tion reforms were pursued goes beyond the scope of this effort.12 ‘test’ whether such nontechnical factors in fact play a significant The understanding of nontechnical drivers used in this report is role and to capture how they do so. Moreover, if this is the case, based on the emphasis of the WDR 2017 regarding the importance of are there options for the international community to integrate a functional improvements, while for the operationalization of political recognition of such factors into support approaches in a way that economy drivers it draws on the framework developed in Fritz, Kaiser facilitates greater progress? and Levy (2009). Therefore, it covers stakeholder incentives, inter- Most teams working on support for PFM reforms have some ests, relationships and powers, existing institutional provisions, notion of how nontechnical drivers, including the wider institutional discrepancies, and changes being pursued, as well as structural context and political economy aspects, affect opportunities for factors such as fiscal dynamics (see Box 1.1). With respect to a reforms. The overall political make-up of the country is known; number of specific aspects such as the legislature’s role in budget 11. Nontechnical drivers and constraints include institutional incentives and constraints, well as political commitment, capabilities, and demand to pursue reforms, interacting with institutional aspects; see also 12. The World Bank Group launched a global study on revenue-related Diamond (2011). reforms in late 2016. The World BAnk  3 Overview of Nontechnical Drivers and Their Interaction BOX 1.1 with Technical Aspects The Problem-Driven Framework in a Nutshell The problem/issue for which a solution is being sought Structural drivers (e.g., global commodity prices) Economic and technical analysis Political economy analysis— Institutions (formal and of feasible solutions focused on informal) Stakeholder interests, constellations & dynamics Implications: what can best be done to ‘make reforms happen’/find a solution that delivers progress? Implementation of the identified approach Note: This framework can be used for country and sector/issue specific analysis; certain issues (e.g., PE dynamics between countries) involve additional layers. Stakeholders are identifiable individuals or groups that by stakeholders; while conversely, structural conditions or have specific interests. Stakeholders interact with each shifts in these conditions can influence the interests and other as well as with existing institutions (formal and infor- opportunities that stakeholders have. Prime examples are mal, including gaps in formal institutions). Stakeholders natural resource endowments and shifts in the valoriza- may seek to influence institutional changes. Structural tion of such endowments, as well as in global prices and drivers are those that are not subject to direct influence opportunities. Technical Analysis and GPE Perspectives Are Complementary Vulnerability/problem Country-level GAC-PE Thinking constraints to growth/poverty reduction; features & situation challenges to achieve results in sectors; unsatisfactory operations Strategizing Problem-driven governance and political economy diagnostics Technical diagnostic What can be done? dialogue 1. Identification of governance arrangements (setting out options) and underlying political economy drivers 2. Assessment of feasibility of policy options Deciding Decisions, approach to reforms, Doing engagement WB policy advice to government/ engagement with local stakeholders Learning did our engagement work better as the result of integrating PEA (and other DDD aspects)? 4  Political Economy of Public Financial Management Reforms approval, it draws on the existing literature on political economy could be another motivating factor, but these have been shown aspects of PFM/budgeting as outlined earlier. to often not produce effective and sustained reforms unless they This perspective on nontechnical drivers also links to a wider coincide with other motivations. Chapter 3 of this report considers debate about innovative approaches to challenging reforms and insti- to what extent motivations to improve PFM were present in the tutional strengthening. Contributions to this debate have included case studies and across successive governments. an emphasis on the need for adopting more iterative and adap- The report deals with several topics that have emerged in recent tive solutions rather than more standardized designs (Andrews, discussions on PFM and wider public sector reforms, while its main Pritchett, and Woolcook 2017), as well as on taking a ‘power and objective is to consider operationally useable and useful implications. systems approach’ when considering change (Green 2016). One Recent discussions have focused on concerns that the introduction overarching umbrella for these ideas are the proposals for ‘doing of systems has not been matched by actual functional improve- development differently’ (Bain, Booth, and Wild 2016). ments, that reform approaches entail superficial copying and not The specific focus and contribution of this report is on how always a real intent to achieve improvements (‘isomorphic mim- to consider political economy drivers, how these affect opportunities icry’), and calls for more adaptive as well as for more ‘politically for PFM reforms, and how such insights can be used to shape future smart’ designs (World Bank 2012 and 2017). A further strand of PFM reform efforts. The report is focused on understanding political the discussion has focused on questions of how best to sequence commitments and interests in PFM reforms, as well as obstacles reform efforts and lessons learned in this regard (Diamond 2012). and challenges, how these shaped actual reform progress, and the This synthesis report is based on a combination of quantitative likelihood of such reforms to realize expected impacts. and qualitative assessments of experiences with PFM reforms over the There is frequently an implicit ‘functional’ assumption that PFM past 10 to 20 years. The first component is a quantitative analysis reforms are in public interest and hence should also be in the interest of patterns of PFM performance and associated country character- of decision makers and participating stakeholders. Such a perspec- istics. The analysis explores which characteristics are associated tive can easily overlook that stakeholders expected to propose, with better or worse PFM systems, and also traces the degree to authorize, and implement PFM reforms may well have divergent which countries have experienced improvements between suc- interests, for example, to accommodate excessive demands for cessive rounds of PEFA assessments. The second source are five budget allocations even when it is known that actual revenue will country case studies, which trace the specific processes through be insufficient; allow powerful allies in the cabinet to use more which PFM strengthening was pursued, the stakeholders involved, funds than allocated; generate resources for election campaigns and the actual improvements achieved (see also Figure 1.3 regard- through allocating contracts to allied firms, which are then allowed ing the expected PFM change and results chain). to overcharge and split the difference; or obscure how exactly The case studies include countries of different sizes, income funds are allocated and used through limited quality of account- levels, and regions. The five cases are Georgia, Nepal, Nigeria, ing, reporting, and auditing. Philippines, and Tanzania. They span four global regions (Europe Conversely, from a political economy perspective, an over- and Central Asia [ECA], South Asia Region [SAR], Sub-Saharan arching question is why politicians would ever be interested in PFM Africa [SSA], and East Asia and the Pacific [EAP]), and include reforms that reduce discretion and opportunities to maintain power. three middle-income, and two low-income cases, one resource rich Taking political economy interests into account, it can appear as case (Nigeria), and one postconflict context (Nepal, with other if poor PFM practices may be a rather stable negative equilib- cases also experiencing shorter-term or regional conflicts). In all rium, as these provide discretion and opportunities to elites to of these, there has been substantial pursuit of PFM reforms by maintain power and influence in ‘limited access order’ contexts governments and support by the international community. Nige- as described by North et al. (2013). However, there are several ria and Tanzania, in particular, stand out as countries with large- potential motivations, that we can surmise, to seek moving out of a scale support for PFM reforms. While smaller in absolute terms, negative equilibrium of poor PFM practices. One is the fact of fac- relative to the population and the overall ODA, support targeting ing increasing citizen expectations and seeking to respond to these PFM reforms has also been substantial in Georgia, while Nepal within limited fiscal space, which in turn requires to constrain the and the Philippines represent countries with more limited levels discretion and opportunities for diverting funds of stakeholders of support. along the public expenditure chain. External or other significant The time frame for each case study is approximately a decade, security threats—which have long been analyzed to motivate the that is, typically since the mid-2000s up to mid-2016, with some creation of more effective states—may also compel stakeholders to variation in the most relevant starting points, as well as selected seek establishing systems and processes which allows to track that key reforms that may have happened earlier (for example, PFM funds are (by and large) spent as has been decided, and to con- reforms in Tanzania and procurement reform in the Philippines in strain opportunities for diverting funds. External conditionalities the late 1990s). The specific PFM reforms traced were mostly set The World BAnk  5 Change Management, Political Economy, and PDIA BOX 1.2 Perspectives on PFM Reforms Considerations about change management and the politi- a certain reform and for overall goals, such as improving cal economy of PFM reforms are related, but cover distinct governance and using funds effectively for service deliv- aspects of reform authorization, implementation, and effec- ery. This includes being clear about likely limits and volatil- tive use. ity of political commitment. For example, volatility due to uncertain support for a government in the legislature, due Change management has been developed as a busi- to upcoming elections, or due to growing security concerns ness management concept and is focused on how which can draw attention away from economic governance leaders can bring organizations and people along, reforms. typically following a top-level decision that change is needed, for example, to increase productivity and sus- Depending on the level and solidity of political com- tainability of a company. It refers to the ‘process of help- mitment to PFM reforms, it is then sensible to design ing people understand the need for change and to motivate a reform strategy that corresponds to the relative ‘win- them to take actions, which results in sustained changes dow of opportunity’ and a change management effort. in behavior’ (World Bank 2015). The process is important, As discussed in this report, for contexts with more volatile as the introduction of new ways of working will only deliver or limited political commitment, pursuing a gradual and results if they are widely accepted and actively utilized, partial strategy may be most reasonable while situations of rather than resisted or circumvented. While change man- strong commitment provide a basis for pursuing bolder and agement ideas have originated in the private sector, they more comprehensive reforms. Change management then have also been applied to public sector organizations, with entails deliberate efforts to communicate effectively within most observers noting some specificity (Van der Voet 2014 affected organizations on why a certain change is being and Kuipers et al. 2014). made, what to expect in terms of sequencing of reform steps, setting out what training will be needed and offered, A political economy perspective in contrast is primar- and so on. This is particularly relevant for reforms which ily concerned with whether leaders have an interest to affect a large number of staff and of how things are done, pursue change. As is discussed in Chapter 3 of this report, for example, the introduction of a new accounting system political commitment has several roots, including whether or of large-scale IT applications. This type of change man- there is a perceived need and/or demand from citizens, agement may also involve nongovernment stakeholders, and whether top-level decision makers have a sufficiently for example, suppliers, when rolling out changes to pro- strong mandate. One indication of political commitment curement systems. is whether decision-makers are willing to commit politi- cal capital to a reform, such as explicitly backing changes As this report discusses, in particular, in Chapter 5, many against resistance from some stakeholders. PFM reforms remain only partially completed or utilized, often for many years. In such contexts, seeking to identify In addition, recent thinking on problem-driven iterative areas of potential gradual improvements through PDIA- adaptation (PDIA) has emphasized that reform efforts type processes may be particularly useful. Consulting with should entail a more iterative and adaptive process stakeholders involved along different parts of the public (Andrews 2013). Regular discussions with stakeholders finance and results chain (for example, MoFs, planning and would bring out ideas, demands, and insights as to what budget directorates of sector ministries, regional bureaus, is working and what is not. Greater flexibility in donor sup- and frontline organizations), can be very helpful in identify- port (‘from logframes to searchframes’) could then enable ing what is working and what is not and to generate ideas adapting projects and programs in place. about improvements. At the same time, one has to man- age the risk of facing an overwhelming set of problems or of demands for additional funding needed to generate These perspectives on PFM reforms can be combined improvements. Moreover, greater adaptiveness of donor in actual operational work. One is to be clear and hon- programming would be a great asset for responding in a est about the level of political commitment that exists for timely and pertinent way to windows of opportunity. Source: Authors, based on referenced material. 6  Political Economy of Public Financial Management Reforms Table 1.1.  Support to PFM Reforms in US$, millions 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Georgia 7.7 9.1 12.6 28.2 42.3 17.6 24.4 27.1 9.3 37.2 21.6 Nepal 0.4 0.4 0.2 0.3 0.7 1.0 9.8 5.7 15.6 6.9 5.7 Nigeria 11.5 15.8 48.7 24.3 39.5 389.6a 200.3 81.5 81.9 167.4 19.1 Philippines 0.6 3.6 2.8 1.8 3.5 5.1 3.0 12.2 11.8 17.3 8.1 Tanzania 59.3 77.9 85.0 83.1 64.2 62.0 11.7 45.5 72.7 72.6 66.6 Source: QWIDS 2016. Note: a. The 2010 figure for Nigeria includes a US$350 million IDA Development Policy Operation with PFM-related triggers. Policy operations are also included for other countries. out by governments in some form of PFM reform plans, while the initiation and support for PFM strengthening, and what factors and level of detail of these varies. For each case study, the main fac- stakeholders have limited or constrained such reform efforts. Based tors covered include fiscal dynamics, institutional arrangements on this, the case studies explore the extent to which forward- of PFM functions, wider political constellations and goals—such looking implications and guidance can be derived. This is par- as the stability in government and overarching government aims— ticularly relevant with regard to whether support to PFM reforms stakeholder interests and incentives related to the relevant PFM can be calibrated in ‘smarter’ ways, cognizant of opportunities and reform aspects, and, as the ‘dependent variable’, the actual prog- constraints. ress made with regard to key reforms. Case studies were developed by a dedicated analytical team Both the quantitative analysis and the five case studies draw for each country, with some staff participating in more than one case. on the data and information generated by PEFA reports to assess the Teams included members particularly knowledgeable about a par- status of PFM performance and the degree of PFM progress achieved, ticular country and relevant political economy dynamics, as well triangulated with other sources of information. Since being launched as PFM experts. Each case study involved one mission, as well as in 2005, PEFA assessments have become an important source follow-up discussions with country office-based staff on draft find- of information on the status of PFM across a growing number of ings. Two initial case studies, first undertaken in 2013, Tanzania countries.13 Furthermore, correlations were tested for robustness and Nigeria, were updated in 2015–2016. also with an alternative indicator, the CPIA assessments of PFM.14 Each case study followed a case-study protocol developed at the As is the case for any indicators, it is important to be mind- outset to provide a common framework (see Annex 3). The case-study ful of how PEFA data are generated. PEFA reports reflect primary protocol is based on the framework as reflected in Box 1.1, and data collection based on a framework first developed in the early corresponds to a number of the dimensions highlighted in WDR 2000s and subsequently updated several times, most recently and 2017. The protocol provides guidance on the specific information significantly in 2016. An additional point to consider is that while to collect and issues to focus on. In addition, individual teams there is a detailed framework of ratings on which PEFA assess- also sought to respond to specific issues of interest or concern to ments are based, there is still some discretion in how final ratings country teams. For each case study, issues covered include (a) key are established, and there can be some pressure to consider higher fiscal and political country features relevant for PFM reforms, rather than lower ratings when the information available suggests a (b) PFM system status and overall reform agenda, (c) specific PFM borderline case. A number of repeat assessments have highlighted reform aspects and underlying PE drivers, and (d) guidance to help earlier ratings as too positive. The study team still considers it bet- inform an overall story line and recommendations. ter to utilize this standardized source of information than to leave The case studies take a broad perspective on possible nontech- it aside, but these caveats should be noted to avoid an overly rigid nical drivers, while seeking to understand how these come to specifi- interpretation of results. cally affect PFM reform efforts (positively or negatively). Potential The main benefit of the case studies is the ability to under- drivers explored include main domestic and foreign policy goals, take ‘process tracing’ (Collier 2011) of who and what has led to the fiscal dynamics, including revenue and deficit trends, as well as the dynamic between the executive and the legislature, and 13. For extensive background information, see www.pefa.org. other institutional features which may have a significant bearing 14. Quality of Budgetary and Financial Management, Indicator 13 of on PFM. To the extent possible it also seeks to trace the specific the CPIA. Note that teams developing CPIA ratings are guided to draw on available information from recent PEFA reports wherever available, so policy goals and commitments of successive governments, as well the two indicators should not be considered as fully independent of each as the potential influence of nonstate stakeholders. The presumed other, but CPIA ratings are produced more frequently (on an annual basis) and draw also on other information. logical chain of PFM changes is reflected in Figure 1.3. The World BAnk  7 The synthesis draws on the combined findings from the quanti- The ultimate—and most challenging—aim of the synthesis is tative analysis and the case studies. This also allows exploring the to be operationally relevant and forward looking. Given what we can comparative findings across the cases reviewed (Byrne and Ragin learn about where PFM reforms have made progress, why and how, 2009; Collier 1993; Lijphart 1971). For example, if we ascribe what can we recommend for efforts at moving such reforms forward the achievement of reform progress in one case to a particular fac- and supporting them externally? Chapter 6 focuses on the forward- tor, do we find that the same factor was also present in other cases looking implications, both in terms of generalizable insights that achieving progress on the same PFM reform aspect? Conversely, appear to be relevant for various potential situations, as well as on if we argue that something posed an obstacle to reform progress selected specific implications for the countries that were reviewed do we find that a similar pattern prevails in other countries (of the as a way of illustrating what this can look like. same causal factor having the same association with more or less progress)? Figure 1.3.  PFM Change and Results Chain Country choices and experiences of PFM reforms: reform outputs, intermediate and final outcomes achieved (with a focus on selected PFM results); and mapping of underlying stakeholders, incentives, and dynamics, with a view to develop (a) a systematic as well as practical way of mapping PE factors relevant for specific areas of reform and relevant ‘monitor- ing indicators’; and (b) guidance on whether and how typical obstacles, pitfalls, as well as opportunities can be addressed proactively. Inputs to PFM Intermediate Final Reforms Outputs Outcomes Outcomes Impacts Government Changes in laws, Transparency and Fiscal Improvements in inputs, including rules, and comprehensiveness discipline state capacity institutional setup procedures for PFM reform Links to policy, planning, Strategic Improvements in coordination Improved and delivery allocation of service delivery information resources DP support to PFM systems and Control, oversight, and [Improved reform efforts business accountability Operational management of processes efficiency in budget support] Harmonization and Cost-savings (reduction in public alignment of DP Changes in idle balances, short-term spending Credibility and support people, skills, and borrowing) trustworthiness of organizations the state Complementary DP inputs Changes in incentives and controls Underlying stakeholders, incentives, and dynamics 8  Political Economy of Public Financial Management Reforms 2 PFM Progress and Association with Country Characteristics—Key Findings from Analyzing Relevant Indicators T he aim of this chapter is to set out PFM performance and recent differ between these two sources,16 while one would still expect trends globally, regionally, and by income groups, to summarize them to generally show the same trends (positive, negative, or key findings from the correlation analysis, and to situate the five stagnating) with regard to PFM performance. case studies against this background. The detailed reform efforts and For all low- and middle-income countries collectively, both the underlying drivers for the five countries investigated in depth are PEFA indicators and the CPIA-13 indicator show only a small aggre- traced in subsequent chapters, especially Chapters 4 and 5. gate improvement for the respective time periods covered. CPIA-13 The chapter proceeds in three sections. Section 2.1 reviews shows near stagnation, with the aggregate average for all available the variation in PFM performance, including the trends as trace- countries moving from 3.35 to 3.37—on a scale from 1 (worst) able over the past 10 to 15 years based on CPIA and PEFA data, to 6 (best) between 2001 and 2014 (see also Vani 2012). The globally and disaggregated by region and income group. Sec- average of PEFA ratings—across all dimensions—improved slightly tion  2.2 summarizes key findings from a quantitative analysis from 2.4 to 2.5 (changing the letter ratings from D [worst] to A focused on what country characteristics are associated with better [best] to numbers 1–4) between the earliest and most recent PEFA or worse PFM performance.15 Section 2.3 looks at how the five assessments, corresponding to a ‘C+’ rating.17 cases reviewed in depth in this report perform on publicly avail- PEFA results by region show that ECA outperforms the other able indicators, and how this corresponds to the findings from the regions in all six PEFA dimensions(see Figure 2.1). CPIA data also analysis of country characteristics. shows ECA as the highest scoring region with a 3.79 average rating for CPIA-13 in 2014. The PEFA dimension on which ECA particu- larly outperforms other regions is accounting and reporting. 2.1 Variation in PFM Performance The other regions vary significantly across the dimensions. across Regions and Income Groups SAR shows the second strongest aggregate average, albeit with noticeably stronger performance on some dimensions than on oth- Across the world, the status of PFM performance and success with ers (see Figure 2.1). Aggregate performance among EAP, Latin reforms vary significantly as noted in Chapter 1. This section reviews America and the Caribbean (LAC), and Middle East and North what relevant indicators show regarding the range of performance Africa (MENA) is relatively close to each other, while SSA is the and performance changes over time. It looks at performance based weakest performing region based on PEFA data. However, PEFA on PEFA indicators and compares this with data for CPIA-13, that and CPIA data are somewhat inconsistent, with the latter indicat- is, the two standard data sets available for measuring the quality ing a stronger performance for LAC than for SAR (see Table 2.1). of PFM systems. CPIA-13 data has been collected for longer than While SSA is overall the weakest region, it outperforms EAP PEFA data and is generated annually for most low- and middle- in predictability and control of budget execution and LAC in external income countries, while ratings are only made publicly available scrutiny and audit. It underperforms all other regions (SAR, ECA, for countries receiving IDA-lending. A disadvantage of the CPIA and MENA) in budget credibility, comprehensiveness and trans- indicator is that it provides only a single indicator rather than a parency, and accounting and reporting. more disaggregated and detailed perspective on PFM performance as PEFA assessments provide. Thus, the time periods covered by the two sets of indicators, the level of detail and the frequency 16. The two indicator sets are not independent of each other since recent PEFA assessments are likely to influence the assessments of CPIA-13. 15. An earlier and more detailed version of the quantitative analysis can 17. Clearly, there are methodological concerns about such an be found in Fritz, Sweet, and Verhoeven 2014. The main aspects of this aggregation, but at the same time, there is also an interest to check analysis were revisited in 2016 to check for any changes in results due to whether any aggregate improvement can be observed, taking the additional observations becoming available. limitations of doing so into account.   9 Figure 2.1.  PEFA (most recent) by Region18 Figure 2.2.  PEFA (most recent) by Income Group Budget Budget credibility credibility 4 4 Comprehensiveness Comprehensiveness Revenue 3 and transparency Revenue 3 and transparency 2 2 1 1 External Policy-based External Policy-based scrutiny and budgeting scrutiny and budgeting audit audit Accounting Predictability/ Accounting Predictability/ and reporting control of budget and reporting control of budget execution execution EAP ECA LAC MENA SAR SSA LIC LMIC UMIC Note: (a) 4 = A, 3 = B, 2 = C, D = 1, NR = 1, NA/NU = blank; (b) scores from the ‘most recent’ assessment for each country were utilized (for example, Afghanistan has a 2005, 2008, and 2013 assessment; only 2013 was included in SAR and LIC); (c) dimensions are calculated based on PEFA methodology, except that the ‘Revenue’ indicators (PI-13, PI-14, and PI-15) were separated out of dimension 4 (predictability and control of budget execution) as a 7th dimension labeled ‘Revenue’; (d) 13 observations were dropped;’19 (e) high-income countries were excluded. Definitions of the income groups (LIC, LMIC and UMIC) can be found at https://datahelpdesk.worldbank.org/knowledgebase/ articles/906519-world-bank-country-and-lending-groups The variation across dimensions is also evident when examin- in Chapter 5 of this report. Considering ‘revenue’ as a separate ing PFM performance by income group (see Figures 2.1 and 2.2). dimension indicates that performance is somewhat stronger than While low-income countries (LICs) on average perform lower for the remaining indicators under ‘predictability and control of than MICs, there is significant variation by dimension. Low- and budget execution’. ­ middle-income countries perform at similar levels on policy-based Table 2.1 reflects the change in PFM performance by region budgeting, and LMICs and LICs also overlap or nearly overlap in and income groups over time based on CPIA data. As noted earlier, their performance with regard to revenue collection and to the because CPIA data are generated annually, it is more suitable for predictability and control of budget execution. 18 19 comparisons over time, in particular when trying to capture how Looking across PEFA dimensions, the lowest rated one is ratings evolved across regions or income groups.20 Looking at ‘external scrutiny and audit’ across all income levels and regions. aggregate performance evolution by income groups, progress has Accounting and reporting is the second weakest cluster, but as mostly been made by LICs, but not by other income groups. Both noted above, showing a significantly better performance of ECA LMICs and UMICs saw limited improvements. Regionally, the CPIA countries compared to all other regions. The issue of why these indicators indicate improvements in ECA and SSA. dimensions tend to perform relatively poorly is explored further Performance regressed in several regions, albeit only to a lim- ited extent, and in part accounted for by the addition of new countries or declines in small states which may have been overrated during 18. Due to the fact that the cluster ‘predictability and control in budget certain years. For SAR, which has a small number of countries execution’ in the PEFA framework covers both public expenditure (8 total), the negative trend in CPIA data is driven by two coun- management and revenue collection, the indicators covering revenue (PI-13–15 [PI-19–20]) are shown as a separate ‘Revenue’ cluster in tries, which saw significant decline—in one case, Bhutan from Figures 2.1 and 2.2. As throughout this report, the reflection of PEFA an improbably high (6.0 ‘perfect’ score) CPIA rating in 2001. For data follows the 2011 version of the framework as most assessments included here were done before the changes introduced in 2016. The EAP, the decline in aggregate CPIA performance is partially driven corresponding indicators of the 2016 framework are provided in square brackets. 19. Dropped assessments include Anguilla (United Kingdom), Bosnia and Herzegovina (no national assessment), Cook Islands (New Zealand), Mayotte (France), Montserrat (United Kingdom), Niue (New Zealand), 20. Using PEFA data for such comparisons poses difficulties since the St. Pierre and Miquelon (France), St. Helena (United Kingdom), Wallis country’s coverage differs for any time period, affecting average ratings and Futuna (France) as they are not considered countries. Ukraine was in particular for smaller regions. Broadly, using PEFA data shows similar dropped because it had multiple scores for some indicators. results, but with some contradictory findings for specific regions. 10  Political Economy of Public Financial Management Reforms Table 2.1.  Change in CPIA-13 by Region and Income Groups % Increase/ Decrease Group No. of Countries 2001 (average) 2007 (average) 2014 (average) 2001–2014 LAC 28 3.59 3.68 3.59 0 SAR  8 4.07 3.44 3.38 −17 EAP 22 3.50 3.43 3.34 −5 SSA 47 2.97 3.12 3.17 7 MENA  9 3.38 3.38 3.42 1 ECA 23 3.45 3.71 3.79 10 LIC 29 3.05 3.19 3.10 2 LMIC 49 3.49 3.41 3.33 −5 UMIC 47 3.83 4.02 3.63 −5 SIDS 29 3.24 3.20 3.16 −3 Total 137 3.35 3.42 3.37 1 Note: (a) Ratings are on a scale from 1 to 6 (best); (b) Number of countries is based on year 2014; (c) Income group classification is according to year of CPIA rating; (d) SIDS = Small Island Developing States (according to United Nations Development Programme [UNDP]). by the addition of new countries to the ratings (Myanmar and for countries (‘cross-section’), which factors are associated with Timor-Leste) and half-step declines in several SIDS. better or worse PFM; second, it examines the effects of these The SSA region has the lowest average rating in both indicator country characteristics on PFM performance by looking at changes sets at the outset but the gap has narrowed from other regions due to within the country over time (‘first differences’). modest improvements. The LAC and MENA regions remained rela- This analysis uses PEFA data to explore how ‘predictable’ the tively stagnant, with some countries increasing and some declin- relative strength and progress on PFM systems are based on key coun- ing in PFM performance. By income group, the CPIA data shows try characteristics. If the relative strength of PFM systems is rather a small improvement for LICs, while showing limited declines for closely associated with key country characteristics, this would sug- LMICs and UMICs. gest that specific institutional and stakeholder dynamics or reform strategies do not matter a great deal. It would also imply that in countries with adverse characteristics, the odds for achieving 2.2 Do Country Characteristics Drive progress against existing patterns are very high. This also helps PFM Performance? Key Findings us to situate the five cases considered in detail, as presented in Section 2.3. from the Quantitative Analysis The country characteristics analyzed include macro-social, The variation of levels and trends in PFM performance raises the ques- economic, fiscal as well as political variables. Specifically, the fac- tion of what drives such variation and the extent to which variation tors explored include the following: population size, levels of gross may be explained by structural factors such as income levels. Thus, domestic product (GDP) per capita, recent growth performance, before launching into an analysis of potential nontechnical driv- natural resource dependency, and experience of fiscal shocks as ers, it is critical to explore to what extent structural factors ‘set the main macro-fiscal variables, aid dependency and tax revenue the scene’ for PFM performance. As the initial exploration in Sec- relative to GDP with regard to sources of revenue, and political sta- tion 2.1 indicates, not all countries appear to have the same likeli- bility, political regime, and the presence of programmatic political hood of having well-performing PFM systems. parties as the main broad political characteristics for which indica- This section focuses on analyzing two types of relationships tors are available for a large number of countries.22 A description between structural factors and PFM performance:21 First, it explores of the variables used is provided in Annex 1. The results from the analysis are presented in Table 2.2, with additional robustness checks presented in Annex 2. 21. The original analysis, including data up to 2013, was published in The analysis indicates that the quality of PFM systems is most 2014 (Fritz, Sweet, and Verhoeven 2014). The analysis in this paper significantly and robustly associated with two variables: a country’s extends that same analysis through 2016 (cross section) and 2015 (first differences). In this update, the cross-section analysis uses all available PEFA assessments and clusters errors at the country level using pooled OLS. 22. For the specification of the variables, see Annex 1. The World BAnk  11 income per capita (positively) and having a high share of revenues that aid and tax remain insignificant. As additional variables, we test a are obtained from natural resources (negatively). In addition, we find variety of growth and fiscal shocks—as experiencing such shocks statistically significant associations with per capita growth (posi- could motivate pursuing PFM reforms. Neither growth nor fiscal tive) and being a SIDS (negative). Programmatic parties appear shocks are found to have a relationship with PFM performance.25 to possibly have a positive and strong impact.23 Smaller and/or Lastly, we employ the same models using only most recent less robust associations are observed with several other variables: PEFAs instead of all available ones, that is, multiple observations population size, political stability, and regime type. Levels of rev- for some countries. We find the main variables, that is, GDP per enue and ODA relative to GDP are not significant. capita and natural resource dependence, remain significant, and We perform a number of robustness checks to test the validity of the magnitudes of the coefficients remain broadly similar across the results and how sensitive they are to variations. First, substitut- models. However, per capita growth becomes insignificant as ing CPIA-13 for PEFA average as the Y variable, results in similar does political stability, and the significance of the regime type magnitudes and significance levels for GDP levels and resource decreases. Tax revenue to GDP and ODA both become significant. dependency.24 The three political variables (programmatic parties, This indicates that depending on the composition of countries and regime type, and political stability) also remain significant, and years included, these factors may or may not be relevant. Table 2.2.  Cross-Section Analysis: Average PEFA Scores and Country Characteristics   (1) (2) (3) (4) (5) (6) GDP per capita (log) 0.2378*** 0.1958*** 0.2568*** 0.2171*** 0.2052*** 0.1887*** (0.0403) (0.0495) (0.0580) (0.0391) (0.0387) (0.0489) Growth (per capita) 0.0276** 0.0335*** 0.0242** 0.0223** 0.0285*** 0.0317*** (0.0115) (0.0096) (0.0116) (0.0106) (0.0099) (0.0112) Population (log) 0.0411* 0.0579** 0.0481* 0.1085*** 0.0502** 0.0259 (0.0232) (0.0241) (0.0255) (0.0288) (0.0228) (0.0277) SIDS −0.2223* −0.1728 −0.2216* −0.2967*** −0.3336*** −0.2480* (0.1225) (0.1295) (0.1242) (0.1087) (0.1142) (0.1253) Resource −0.3386*** −0.3498*** −0.3342*** −0.3542*** −0.2777*** −0.2918*** (0.0911) (0.0864) (0.0931) (0.0841) (0.0881) (0.0911) Tax 0.0106 (0.0074) Aid 0.0017 (0.0037) Political Stability 0.0867*** (0.0212) Regime 0.0487*** (0.0159) Programmatic parties 0.4791*** (0.1639) Observations 262 244 258 252 256 231 R-squared 0.340 0.384 0.344 0.436 0.386 0.351 Note: (a) the Y variable, ‘PEFA score’ is the average of 21 indicators as per the 2011 PEFA framework (PI-5 to PI-12 and PI-16 to PI-28); (b) we exclude PI-1 through PI-4, which measures PFM outcomes, indicators PI-13 to PI-15, which cover revenue administration, and D1 to D3, which are donor- related indicators; (c) 13 observations were dropped, and (d) high-income countries were excluded. Standard errors in parentheses, clustered at the country level. Observations weighted inversely to number of PEFA assessments. No influential observations are dropped. *** p < 0.01, ** p < 0.05, * p < 0.1 23. When using CPIA-13 as the proxy for PFM performance, programmatic parties no longer appear as a significant factor. The measurement of which countries have significant programmatic parties and which ones do not is also open to some interpretation (for the source used, see Annex 1). 24. The main model utilizing CPIA-13 for PEFA is shown in Annex 2. The 25. The results for growth and fiscal shocks are shown in Annex 2. See other specifications are not shown but were tested. also Fritz, Sweet, and Verhoeven 2014. 12  Political Economy of Public Financial Management Reforms Jointly, the variables tested are found to account for 35 percent repeat PEFA assessments, which tend to show greater improve- to 40 percent of the variation in PFM performance as measured by ments over time compared to CPIA-13 data. PEFA assessments (Table 2.2) and around 27 percent to 35 percent We find the explanatory power of the variables in the first-­ using CPIA data, that is, well below half of the overall variation. differences analysis to be slightly lower (approximately, 30 per- Hence, there are some patterns of influence of such country char- cent) than the cross-section analysis. The results are presented in acteristics on the likely quality of PFM systems. However, the rela- Table  2.3. Per capita GDP growth is highly significant and at a tionship is relatively loose, that is, there are many countries whose large magnitude when using the first and the most recent PEFA PFM performance would not be well predicted looking at such assessments (but becomes less significant or insignificant using characteristics. Also, as noted above, only a few variables show a other specifications—see the following sections). There is also a consistently robust relationship, as the results are highly sensitive strong association with initial PFM performance—that is, countries to the countries and time periods included. with initially weaker PFM systems show relatively greater progress These findings are partially consistent with earlier findings, with over time. Counterintuitively, increases in tax revenue collection some differences. De Renzio et al. (2011) similarly found that the over time are associated with a small decline in PFM performance. political regime did not have a significant impact on the strength Testing these patterns for robustness using CPIA-13 data allows of PFM systems, but that political stability has an impact, using using a larger number of 130 countries and shows the following results a different set of variables to measure stability. While de Renzio (see Annex 2): The relationship with initial performance is also et al. (2011) similarly found GDP levels to matter, Andrews (2009) highly significant, and in addition, recent growth has a positive found that income levels were not significantly indicated using a effect, but with varying/lower confidence levels than when using sample from a single region (SSA). The finding that aid depen- the first and most recent PEFAs, and population growth a nega- dency is not associated with PFM performance is also consistent. tive effect—the latter more clearly so than when using PEFA data. Population factors had not been explored previously. There is some Using CPIA data, furthermore, being a SIDS or being resource rich contradiction with regard to the findings regarding fiscal shocks. has a negative effect on the rate of PFM improvements. In addi- Earlier analysis by Krause (2009) found a relationship between tion, moving toward a less democratic regime may have a small experiencing fiscal austerity and strengthening budgetary controls positive effect. In contrast, ODA increases are no longer signifi- in OECD countries—that is, for a different set of countries. cant. The overall explanatory power increases to 35–40 percent. In addition to the cross-section analysis, we explore within- These findings reinforce the overall implication that country country changes over time. A subset of repeat PEFAs can be used characteristics explain some, but not all variation in PFM performance to conduct a ‘first difference’ analysis and relate changes in PFM across countries and within countries over time. Only some variables quality to changes in country characteristics, thus diminish- show robust and consistent associations across different specifica- ing omitted variable bias (that is, inadvertently not considering tions, while other relationships should be revisited as longer time- an important factor that distinguishes between countries). As of lines become available. At the same time, the analysis underlines the end of 2016, repeat PEFA assessments were available for that not all contexts are equal—achieving PFM reforms is likely to 91 countries.26 Some countries have undertaken two PEFA assess- be harder in poorer and more resource-rich environments than in ments, while the maximum number of repeat assessments is five. countries with the opposite characteristics. Still, decision makers For countries with more than two assessments, the first and the and other stakeholders in a range of contexts, including those with most recent available assessments are used for the analysis.27 One more challenging characteristics, can pursue improvements. As important caveat is that there may be two types of bias: one, gov- the first difference analysis indicates, initially weaker performers ernments may be more likely to welcome repeat PEFA assessments can achieve progress. if they have made efforts at PFM reforms (this bias likely shrinks The results are also consistent with the findings from the com- with a broadening number of countries for which repeat assess- parisons across regions and income groups presented in Section 2.1. ments are available). Second, there may be some positive bias in The latter showed that LICs—which on average have been per- forming worse—have seen a gradual increase in performance over time, while performance in MICs has tended to stagnate over the 26. Our analysis includes around 80 of these countries due to data past ten to fifteen years. By implication, if the trend of gradual availability of the X variables (though declines to approximately 60 countries depending on the model and available data). improvements among initially worse performers was to persist over 27. The length of time between PEFA assessments varies by country, an extended period, the importance of country characteristics such so we compute the per year change in PEFA scores (that is, the overall difference divided by the number of years between two PEFAs) as well as levels of per capita income and possibly also natural resource as the per year change in country characteristics over the same period. dependence would gradually diminish. For some characteristics that are not subject to year-to-year changes (for example, SIDS, resource dependency), we use a dummy variable. The World BAnk  13 Table 2.3.  First Differences Analysis   (1) (2) (3) (4) (5) (6) (7) GDP per capita (% change) 0.8797*** 0.8685*** 1.2990*** 0.9420** 0.8054** 0.6902** 1.1012*** (0.3065) (0.3166) (0.4004) (0.4407) (0.3203) (0.3089) (0.3215) Population (% change) −0.4732 −1.8203*** −0.6073 −0.8019 −0.2921 −0.1237 −0.7428 (0.6365) (0.6294) (0.9291) (0.8756) (0.6815) (0.7958) (0.6018) Resource (dummy) −0.0076 −0.0078 −0.0062 −0.0041 −0.0067 −0.0059 −0.0022 (0.0197) (0.0197) (0.0194) (0.0222) (0.0202) (0.0206) (0.0214) SIDS (dummy) −0.0122 0.0029 0.0005 −0.0035 −0.0157 −0.0332* −0.0071 (0.0157) (0.0178) (0.0197) (0.0183) (0.0155) (0.0168) (0.0164) Initial PFM quality (PEFA) (level) −0.0723*** −0.0598*** −0.0547*** −0.0591*** −0.0741*** −0.0635*** −0.0735*** (0.0143) (0.0163) (0.0192) (0.0171) (0.0143) (0.0141) (0.0141) Initial GDP per capita (level in log) −0.0160* (0.0087) Initial regime type −0.0042 (0.0036) Tax (% point change) −0.0264*** (0.0089) Aid (ODA) (% point change) −0.0078 (0.0085) Regime (Freedom House, % change) −0.0066 (0.0333) Programmatic Parties (% change) −0.0913 (0.1507) Political Stability (% change) −0.0035 (0.0034) Observations 82 76 59 64 78 70 76 R-squared 0.276 0.406 0.293 0.238 0.302 0.296 0.332 Note: This analysis only uses countries with at least two PEFA scores where both have all the covariates in the basic regression. Out of the available scores, it keeps the first and the last. As a result, countries with a 2016 PEFA assessment are dropped since there is not enough X-variable data. Robust standard errors in parentheses: *** p < 0.01, ** p < 0.05, * p < 0.1 2.3 Country Characteristics and PFM Tanzania for the past decade, and for Georgia, until 2013. CPIA data for the Philippines is not public. Performance Trends in the Five Case According to the most recent PEFA assessments, Georgia has Study Countries the highest overall PFM performance. This is followed by Nepal and the Philippines, then Tanzania, and last, Nigeria (based on the Against the background of global and regional trends in PFM perfor- 2012 unofficial assessment). Georgia’s strong performance is also mance and the association of PFM performance with various country reflected in CPIA data, on which it improved from a 3.5 rating in characteristics, this section considers the specific performance and 2005 to a 4.5 rating by 2013, the most recent year for which its trends observable in the five case study countries. The countries cov- CPIA scores are publicly available. ered in depth are from four different regions: SSA (Nigeria and The PFM performance in the five countries is broadly consistent Tanzania), SAR (Nepal), EAP (Philippines), and ECA (Georgia). with the quantitative analysis, although countries also defy predicted PEFA data are publicly available for four of the five countries. In performance in a number of respects. Georgia’s strong performance addition, a draft PEFA assessment carried out for Nigeria in 2012 is consistent with overall rather strong PFM in the ECA region and is used.28 CPIA data are publicly available for Nepal, Nigeria, and the country’s middle-income status, while being ‘even higher’ relative to regional and income group averages. Tanzania’s and 28. The assessment was not endorsed by the Federal Government of Nepal’s PFM systems are overall quite closely in line with their Nigeria (FGN) for publication. relative levels of per capita incomes, while being somewhat ahead 14  Political Economy of Public Financial Management Reforms on specific dimensions. Nepal is slightly ahead of the average rat- Relative to its income per capita, it now performs well above the ings for South Asia as well as of other low-income countries in trend line (see Figures 2.3 and 2.4 (a) and (b)). Nepal increased all dimensions except external scrutiny and audit. Similarly, Tan- its performance somewhat between its first and second PEFA zania is performing somewhat above average compared to SSA assessments (2008 and 2015 respectively). However, CPIA-13 and low-income country averages in all dimensions except budget shows no change between those two years. While Tanzania still credibility. The Philippines moved from a performance below the performs above the trend line relative to its income per capita, its per capita income trend line to one that is above for four out of PFM performance deteriorated between the earliest and the most six dimensions, except for two dimensions—budget credibility and recent PEFA assessment, bringing its average performance level accounting and reporting. Nigeria’s low performance is clearly con- slightly below that of Nepal (Figure 2.4 (a) and (b)). CPIA‑13 data sistent with the finding that PFM performance is lower in resource- also shows a clear and even stronger deterioration from a rating of rich countries and is well below expected levels relative to its level 4.5 in 2006 to 3 as of 2015. Nigeria’s performance on the PEFA of income.29 assessment is very low relative to its income level (the CPIA-13 Based on repeat PEFA assessments and CPIA data available, the indicator for Nigeria is rated continuously at 3.0 for the entire following trends over time can be observed. Georgia saw a gradual, period). For the Philippines, PEFA ratings improved significantly continuous upward trend in its PFM performance since 2005. between the two assessments carried out in 2010 and in 2016, Figure 2.3.  PFM Quality and Income (most recent) PFM Quality and Income 3.5 ZAF MDA GEO KSV BFA RWA ALB AZE PER BTN Most Recent PEFA Score ARM CPVMAR 3 ETH CRI IDN ECU PHL MNE HND UZB TUN MOZ MLI PAK SLV MKD KGZ UKR COL BWA VNM JOR SRB MNG AFG NPL PRY MUS TMP WSM 2.5 LBR UGA BLR TJK STP ZMB NIC TON DOM TZA VUT BGD LKA MWI BOL GTM NAM SLB CIV KEN ZAR NER SLE WBG SWZ DMA BDI ZWE GMB KHM JAM SEN TUV GRD PAN GHA VCT 2 FJI GAB LAO BLZ LCA TGO GIN BEN LSO MRT SSD MDV SUR FSM CAF MDG PNG LBN MMR COM COG 1.5 MHL GNBHTI NGA 6 7 8 9 10 Log GDP per capita (PPP, lagged 5-year average) EAP ECA LAC MENA SAR SSA Note: (a) ‘PEFA score’ is the average of 21 indicators (PI-5 to PI-12 and PI-16 to PI-28 as per the 2011 PEFA methodology), that is, we exclude PI-1 through PI-4, which measure PFM outcomes, indicators PI-13 to PI-15, which cover revenue administration, and D1 to D3, which are donor-related indicators; (b) scores from the ‘most recent’ assessment for each country were utilized (for example, Afghanistan has a 2005, 2008, and 2013 assessment; only 2013 was included in SAR and LIC); (c) 13 observations were dropped;29 (d) 5 case studies are highlighted with circles; and (e) high-income countries were excluded. 29. Dropped assessments include Anguilla (United Kingdom), Bosnia and Herzegovina (no national assessment), Cook Islands (New Zealand), Mayotte (France), Montserrat (United Kingdom), Niue (New Zealand), St. Pierre and Miquelon (France), St. Helena (United Kingdom), and Wallis and Futuna (France). The World BAnk  15 Figure 2.4.  (a) and (b). PEFA Dimensions for Five Case Studies—Initial and Most Recent PEFA Assessments Budget credibility 4.0 3.5 3.0 External scrutinity Comprehensiveness and audit 2.5 and transparency 2.0 1.5 1.0 Georgia (2008) Accounting Policy-based Nepal (2008) and reporting budgeting Philippines (2010) Tanzania (2006) Predictability/control of budget execution Budget credibility 4.0 3.5 External scrutinity 3.0 Comprehensiveness and audit 2.5 and transparency 2.0 1.5 1.0 Accounting Policy-based and reporting budgeting Georgia (2013) Nepal (2015) Predictability/control Philippines (2016) of budget execution Tanzania (2013) Nigeria (2012) Note: The first Philippines PEFA assessment was performed in 2007, but not published until 2010. In the case where two or more indicators were missing for one dimension, the score is a 0, based on the assumption that nonavailability of information points to a problem. 16  Political Economy of Public Financial Management Reforms while as noted, CPIA data are not publicly available. As reflected transparency. The ratings cover the years 2006 to 2015 (see Fig- in Figure 2.4, furthermore, the PEFA performance for the Philip- ure 2.5). Among the group, Georgia and the Philippines perform pines is highly uneven, with strong improvements in some areas, the highest and are nearly on par on this indicator. For Georgia, notably policy-based budgeting and transparency, and little or this has been achieved through a significant improvement since none in others. 2006, while transparency was already high in the Philippines in Taking a more disaggregated look, between the first and most the mid-2000s. In Tanzania, budget transparency has stagnated recent PEFAs, Georgia expanded its performance across the board, at an intermediate level; Nigeria has seen a slight improvement in overtaking Tanzania in all dimensions. For Tanzania, performance the most recent years compared to 2006. Nepal stands out as hav- was quite strong in the mid-2000s but declined over time, in par- ing declined in terms of budget transparency. However, as the OBI ticular for budget credibility, predictability and control in budget country report notes, this is due to a one-off non-publication of a execution, and accounting and reporting, while external scrutiny key document, and ratings are expected to re-improve to earlier, and audit saw slight improvements. Nepal improved, in particular, intermediate levels,30 similar to those of Tanzania. with regard to budget credibility and policy-based budgeting, while In terms of country characteristics, the five countries covered its performance for external scrutiny and audit has stagnated at a in the case study approach stretch across a spectrum of socioeco- level below all others. nomic and political situations. As reflected in Table 2.4, the set The Philippines originally had a very low performance on a num- includes two LICs (Nepal and Tanzania) and three LMICs. Nigeria ber of dimensions, while for 2016, especially policy-based budget- moved from low to lower-middle-income status in 2010. Georgia ing and comprehensiveness and transparency are rated well. Budget was a LIC in 2003–2004, moving to lower-middle-income status credibility and accounting and reporting only improve somewhat in 2005 and to upper-middle-income status in 2016. For Nepal, and remain at low levels; so it has a much more ‘lopsided’ perfor- Tanzania, and the Philippines, the income status did not change mance than Georgia. As noted earlier, the qualitative assessment over the past two decades. Population sizes range from small undertaken for this report differs in some respects from the prog- (Georgia) to medium (Nepal, Tanzania) to large (the Philippines ress as reported in the 2016 PEFA report. and Nigeria). Average annual population growth has been highest For Nigeria—for which only a single unofficial PEFA is available in Tanzania (3.0 percent), followed by Nigeria (2.6); moderate in in addition to CPIA data—we cannot judge changes over time across Nepal and Philippines (1.2 and 1.7 respectively); and negative in PEFA dimensions. Compared to the PEFA assessments carried out Georgia (−1.3 percent). for the other countries in the 2010s, Nigeria underperforms in all Economic growth rates have been relatively robust across the six dimensions, except slightly outperforming Nepal on external five cases in the 2000s and 2010s, with a deceleration in three cases scrutiny and audit. Its CPIA rating remained unchanged at a ‘3’ for since 2015. Per capita growth rates were the highest in Georgia all years from 2005 to 2015. As is discussed in detail in the fol- (6.8 percent—combining robust growth and a shrinking popula- lowing chapters, PFM reforms have been pursued in a number of tion) and Nigeria (4.3 percent) and lowest in Nepal (2.5 percent). areas since the mid-2000s, notably for accounting and reporting, but actual progress has remained limited. 30. See http://www.internationalbudget.org/wp-content/uploads/ A further PFM-related indicator that is publicly available for all OBS2015-CS-Nepal-English.pdf: “However, the regression in five countries are the Open Budget Initiative (OBI) ratings for budget transparency observed in Nepal appears to be temporary in nature.” Figure 2.5.  Progress of Budget Transparency Measured by OBI for Five Case Studies 70 60 50 40 30 20 10 0 Georgia Nepal Nigeria Philippines Tanzania 2006 2008 2010 2012 2015 The World BAnk  17 Table 2.4.  Key Country Data and Dynamics Georgia Nepal Nigeria Philippines Tanzania Population (millions, 2016) 3.7 29.0 186.0 103.3 55.6 Per capita income (current US$, 2005) 1,391 293 483 1,355 385 Per capita income (current US$, 2016) 3,854 730 2,178 2,951 879 Average growth 2001–2014a 6.1 4.1 8.0 5.1 6.7 Average growth 2015–2016 2.8 1.6 0.6 6.5 7.0 Projections 2017–2019 4.0 4.6 1.5 6.9 6.8 Projections 2020–2022 5.3 3.8 1.8 7.0 6.5 Average per capita growth 2001–2016 6.8 2.5 4.3 3.4 3.5 Annual population growth (average –1.3 1.2 2.6 1.7 3.0 2002–2015) Natural resource dependenceb 1.2 5.5 33.2 2.7 7.2 (expected to rise) Political stability/conflict Brief war with the Civil war, Decreasing conflict Peaceful in most Largely peaceful, Russian Federation 1996–2006 in the South/Delta; areas, increasing some tensions in in 2008 increasing conflict tensions and Zanzibar in the North conflict in Mindanao Political regimec Partly free Partly free Partly free Partly free Partly free Federal or unitary Unitary Unitary in transition Federal Unitary (federalism Unitary (union with (2007–2015); under discussion) Zanzibar), some Federal (since discussions about 2015) potential changes Programmatic parties 0.25 to 0 0.75 0.25 0.5 0.25 (0 to 1—higher = more programmatic)d Sources: a. World Bank national accounts data and OECD National Accounts data files—World Development Indicator (WDI) central + IMF World Economic Outlook, April 2017. b. Natural resource rents as a percentage of GDP from WDI Central, averages for the years 2000–2013. c. Based on Freedom House ratings. d. Ratings for programmatic party from Database of Political Institutions (2016).31 Higher values indicate a more ‘programmatic’ , or ‘policy-oriented and organized’ party landscape. Overall GDP expanded most strongly in Nigeria during the period 2016. Georgia, Nepal, and the Philippines have a low depen- 2001 to 2014 (8 percent), followed by a sharp deceleration due to dency on natural resources in terms of revenue collection and falling oil prices and other factors since 2015. It was second high- exports. est in Tanzania, where robust growth continued in 2015–2016. Three of the five countries have experienced larger-scale con- Growth was lowest in Nepal, including a deceleration in 2015– flicts, and two have had limited regional conflicts or tensions. Nepal 2016. Except for Nigeria and, to a lesser extent, Nepal, growth is a postconflict country after going through a civil war from 1996 was expected to remain robust for upcoming years based on IMF to 2006. Nigeria experienced conflict in its Delta region until forecasts.31 2009, which consumed significant fiscal resources. Lower-level Natural resource dependency is most significant in Nigeria. conflict and tensions still persist and negatively affect petroleum Revenues from natural resources made up around 70 percent of production. In addition, fighting started in the northern states and total government revenue on average between 2005 and 2015.32 has continued since 2009 due to the Boko Haram insurgency. A rise in natural resource production has been expected in Tan- Two ‘frozen conflicts’ over breakaway regions have plagued Georgia zania, with initial larger-scale gas production starting in spring since the 1990s. In August 2008, the country went through a brief but damaging armed conflict with Russia. In the Philippines, the Mindanao region has seen conflict and tensions since the 1960s, 31. Updated version of Thorsten Beck, George Clarke, Alberto Groff, partially due to religious tensions between the regional Muslim Philip Keefer, and Patrick Walsh. 2001. “New Tools in Comparative majority and the national government, combined with discontent Political Economy: The Database of Political Institutions.” 15:1, 165–176 (September), World Bank Economic Review. in an economically less developed region. In Tanzania, there are 32. Based on data from IMF Art IV reports. demands for greater autonomy by groups on the island of Zanzibar 18  Political Economy of Public Financial Management Reforms which involve tensions, some attacks, and suppression, especially Nepal, political instability appears as the main negative con- during election periods. text characteristic. The Philippines and Tanzania both appear as In terms of the political regime, the five countries are closely intermediate. clustered in that they are somewhat democratic, but democratic pro- Considering relative PFM performance relative to country char- cesses and civil liberties are not fully consolidated. All are rated as acteristics, several puzzles stand out. One is the significant gap in ‘partly free’ by Freedom House, with Georgia, the Philippines, and performance between Nigeria and Nepal, with the latter perform- Tanzania being closer to being considered fully democratic. Civil ing well above the former, and slightly above the trend line, despite liberties are potentially an important precondition for enabling being a postconflict environment. A second puzzle is Georgia’s civil society engagement on PFM and for facilitating fiscal trans- rapid transformation, from relatively low performance based on parency. Moreover, holding regular elections without predeter- early CPIA data to performance well above the trend line in most mined outcomes creates a set of incentives for political leaders recent years. A further puzzle is why Tanzania saw backsliding that differs from countries where elections are strongly controlled rather than at least stagnating trends, given its political stability, and where citizens cannot truly exercise their voice through this solid growth, and at least until most recently, absence of natural channel. resource dependence. A further interesting parallel is that discussions over the reconfiguration of national-subnational relations are ongoing in sev- eral countries. Tensions around fiscal federalism have been most 2.4 Chapter Summary prominent in Nepal. There, discussions about federalism were As indicated at the outset, the intention of this section has been to ‘set responsible for long delays in agreeing a new constitution,33 as the scene’ in terms of taking a broad look at what PFM performance is further discussed in Chapter 5. In the Philippines, moving to a looks like globally and across key country groupings, what we can federal system is under discussion, while in Tanzania, the consti- discern about trends over time,and factors that appear to be associ- tutional review process includes potential changes to the unitary ated with better or worse performance. This broad perspective also system.34 Nigeria has an established federal system, with signifi- allows situating the five cases that are explored in detail in the cant powers allocated to the country’s 36 states and respective following sections. governments. Intergovernmental relations and changes to these Several findings stand out. One is that performance has not can have significant impacts on PFM and on how to approach seen a great deal of further improvements in MICs on average over PFM strengthening. the past 15 years, while LICs—on average—have seen gradual Nepal and the Philippines score relatively highly on the mea- gains. Regionally, progress has been greater among ECA countries sure for programmatic parties, while in the other three cases, only one and, conversely, less so among other middle income regions. or no politically significant programmatic party is present. Nepal, in Regression analysis indicates which country characteristics particular, has a collection of left-wing political parties, including are associated with better or worse PFM performance. Key factors Communists and Maoists, and various factions of these, as well as include income levels per capita—which are generally associated a more conservative party (Nepali Congress). However, as is dis- with the quality of institutions—and a negative influence of high cussed further in Chapter 3, the presence of programmatic parties natural resource dependency, also highly consistent with the avail- in this case does not align with a stronger ability to make credible able literature on institutional quality. Other variables such as commitments (as postulated by Cruz and Keefer 2010), due to a growth, political stability, and being a small island state, vary in high degree of political fragmentation and resulting instability of their significance and impact depending on how PFM performance governments. is measured (PEFA or CPIA 13) and what observations are included In sum, country characteristics are overall most facilitating (all PEFA assessments or only most recent ones). An analysis of in Georgia and least in Nigeria and Nepal. Georgia is a MIC and initial trends over time (‘first differences’) indicates that countries has experienced solid growth, while on the downside, it experi- with initially weaker performance on average saw gradually greater enced a period of instability/conflict in 2008. Nigeria is affected improvements, suggesting some potential convergence, albeit at by two negative drivers, high dependency on natural resources a slow pace. for revenue and exports, as well as significant instability. In Situating the five case study countries, Section 2.3 sets out their PFM performance relative to the distribution of performance 33. Constitutional debates lasted from 2008 to eventual adoption in relative to incomes and maps out what trends in PFM performance 2015. 34. In Tanzania, the main subnational entity demanding federalism is the can be observed. Two of the five cases appear as relative ‘out- island of Zanzibar. liers’: Georgia, which has developed a very strong performance, The World BAnk  19 and Nigeria, which has a weak performance relative to its current The following chapters are dedicated to tracing the how and level of income. The Philippines also performs rather well relative why of these contrasting trends in PFM performance across the five to per capita incomes (while as noted in the section, ratings are countries. Chapter 3 discusses wider country contexts, including highly uneven across different dimensions). levels of political commitment to governance reforms in general In terms of changes over time, available indicators suggest terms, high-level policy goals, and fiscal dynamics. Chapter 4 con- improving trends for three cases, declining performance for one, and siders the specific institutional and legal starting point and key stagnation for another case. Repeat PEFA assessments indicate changes made or stalled, including the setup of central finance improving performance for Georgia, Nepal, and the Philippines; agencies (CFAs) and key legal provisions and efforts at changing and in contrast, declining performance for Tanzania. CPIA-13 is PFM-related legislation. Chapter 5 takes a detailed look at what consistent with trends as reflected by repeat PEFAs for Georgia PFM reforms were proposed and what actual progress was subse- and Tanzania, shows a decline and then re-bound for Nepal, and quently made. stagnation in Nigeria over a 10-year period at an intermediate- low level. 20  Political Economy of Public Financial Management Reforms Political Commitments and Fiscal Trends 3 T 3.1 Political Commitment to Reforms— he purpose of this chapter is to look at the ‘big picture’ of incentives and constraints that can be expected to influence a government’s interest in PFM reforms. Section 3.1 considers What Would We Look For? political commitment and interests that may be relevant for the High-level or leadership commitment is widely reported in the litera- pursuit of reforms. First, it looks at high-level political commit- ture (for example, Robinson 2009) and among practitioners as being ment to reforms, drilling down into specific aspects influencing crucial for PFM as well as for a range of other reforms.35 However, few the intensity and stability of such commitment. Second, con- efforts have thus far been made to empirically capture or compare sistent with the emphasis of WDR 2017 on inequities and their levels of commitment, and high-level commitment largely remains effects on various aspects of governance, the section explores the in the realm of ‘anecdotal’ treatment. In many cases, political nature of economic rents and their distribution. A prevalence of commitment is assumed to be present, for example, because it is nonproductive rents and associated rent-seeking may create more expressed in national development strategies or in PFM strategies; interests opposed to reforms. As identified in Chapter 2, a high however, these are done with some degree of routine and often in level of natural resource dependency is associated with lower PFM response to donor requests linked to financial support, and there- performance, and so it is relevant to consider some of the mecha- fore may not reflect real commitment. None of the governance indi- nisms at play, as well as other forms of rent that are less readily ces available to date capture ‘government reform commitment’.36 captured by cross-country data on key characteristics. Therefore, there is no easy way of comparing the levels of commit- Section 3.2 provides a summary of the main policy goals of ment between different countries, or testing whether its assessed the governments holding office during the period under review in strength is actually associated with better or worse progress being this report. This includes not only the overarching policy goals on made. Weaknesses in reform commitment are often diagnosed ex the one hand, but also the fiscal pressures faced by these gov- post as a reason for why reforms were not able to progress. ernments on the other hand. The latter comprises fiscal trends, A relevant issue is therefore if and how we can capture govern- including revenue and expenditures, debt dynamics, and the level ment commitment; and related to this, whether we can better under- of aid relative to GDP. Together, these fiscal dynamics can influ- stand the link between high-level commitment and the actual reforms ence high-level policy options and incentives. The sections also undertaken. This section seeks to systematically describe the level explore to what extent, governments were able to expand the avail- of commitment to the public sector and PFM reforms in the five able fiscal space through revenue increases. countries that are analyzed. This includes looking beyond PFM Overarching goals and fiscal trends matter to understand a gov- and at the overall goals that each government pursued. ernment’s motivation for PFM reforms. Government decision makers will tend to be interested in PFM reforms not as an end in itself, but as a means to wider goals, such as ensuring a functioning 35. For example, Allen, Hemming, and Potter (2013, 5) “Politics and state, being able to deliver more public goods and services within politicians play a large role in the process of change [i.e., PFM reforms].” existing fiscal envelopes, or signaling reforms to external stake- 36. The index developed and updated by Cruz, Keefer, and Scartascini (2016) on programmatic parties is one effort at approximating the holders, as well as shorter term political goals such as maintaining ability to make binding commitments; but there are cases of political loyalty of key supporters and/or ensuring broader popular support. commitments without programmatic parties being present.   21 One key aspect of political commitment to reforms is the nature A third dimension is the issue of whether an incoming gov- and strength of a government’s mandate, and the promises made to ernment has a clear idea of what to do and how to pursue reforms. citizens during elections.37 Promises of reforms made during elec- This can depend on whether key members of the government with tion campaigns are often exaggerated relative to what politicians responsibility for PFM have previous experience in government actually intend to do or are able to do if voted into office; but they (specifically in PFM-related positions), and also, whether there are create expectations that many politicians will subsequently try to any existing reform plans or options that are ‘ready to go’ once a take some action on. Conversely, an absence of reform promises political mandate is available (see also Kingdon 1995). Without made during an election, typically, is a negative indication regard- clear plans, it will take a new government more time to establish ing a government’s intention to reform the public sector (see also a clear and specific reform agenda, which is costly given that the Keefer and Vlaicu 2008). window of opportunity for key reforms is often limited. In addition to the electoral commitments or promises made, the It is inevitably challenging to identify unambiguous indicators strength of the electoral mandate received also matters. Taking deci- and thresholds for each of these aspects, such as a specific share of sions and implementing measures is easier for governments that votes won to consider a mandate as strong or weak. However, it is have a strong and clear electoral mandate, while it is more difficult possible to capture all three aspects empirically with some preci- for governments with weaker mandates. For example, holding the sion. The main elections that are considered here are those at presidency without a majority in parliament can lead to political the expected ‘starting point’ of PFM reforms. In all five countries, deadlock or at a minimum, can make decisions on reform mea- additional elections and rotations in government have taken place sures difficult. Electoral rules also matter in this regard because since then. These subsequent cycles are discussed briefly for each they determine how a given share of votes translates into seats or country. into who wins a particular office. Winning 40 percent of the popu- Across the five cases, the strength of the political mandate was lar vote can translate into quite different shares of seats in the distributed as follows (see Table 3.1): legislature across different countries, depending on the electoral rules. • In Georgia, all three dimensions (electoral commitment, man- date, and preparedness) were relatively or very strong in 2004. The incoming leadership had based its electoral appeal on completely changing the government (aimed at eliminating 37. The group of countries reviewed here have all held regular elections corruption and regaining state effectiveness); it won a very during the period covered. Somewhat different considerations apply to countries that hold no elections or only highly controlled elections. strong and unified mandate in the elections for both the Table 3.1. Indications of Overall Political Commitment to Public Sector and PFM Reforms at the Outset Georgia Nepal Nigeria Philippines Tanzania Election Year 2004 2008 2007 2010 2005 Electoral Yes, electoral Somewhat, large No, commitment to ‘national Yes, significant No, some reform commitment to platform to overhaul parties favor greater unity’ rather than reform commitment to orientation and governance and governance inclusion of and after the first more or less governance reforms promises, but overall public sector service delivery democratic leadership continuity rather than reforms? for disadvantaged transition change groups Strong mandate? Yes, very high No, highly Yes, President with 70% Somewhat, President Yes, President won support, both fragmented vote of the vote elected with 42% of the with just over 80% of President (96%) and and difficult to form Parliament: 79 % PDP: 68% vote.a President’s party the vote; aligned party parliamentary majority a government House of Representatives: controlled the House wins over 80 percent (67% of votes) (Maoists as largest 28/36 governorships;b losing of Representatives, but of seats in parliament party win 229 out of strength due to death of only part of a coalition (264 out of 324 seats) 601 seats) incumbent in the senate. Being prepared? Yes, relatively No, initial incoming Yes, relatively strong and Yes, in particular, Somewhat, the same well prepared, government is experienced key technical well trained and party has held the key ministers had composed of former appointments experienced Secretary government since previously served in rebels of Budget independence, so in the government principle prepared to operate, but limited commitment Note: PDP = People’s Democratic Party. a. The electoral system does not foresee a runoff; the candidate with the highest share of votes is elected. The constitution limits each President to a single term. b. http://www.eods.eu/library/FR%20NIGERIA%202007_en.pdf. 22  Political Economy of Public Financial Management Reforms executive and the legislature, and the leadership assuming another through an electoral process, as happened in 2007 office was reasonably well prepared, given that several of its for the first time, was widely regarded as a key achievement members had previously served as ministers or in other posi- in itself. The electoral commitment was primarily to preserve tions in the government. The initial window of opportunity of ‘national unity’ in a federal and complex country with signif- the 2004 government lasted for about three years. In 2008, icant internal divisions, rather than to improve governance. the next elections were held, yielding a somewhat reduced The incoming government was relatively experienced. but still solid mandate (60 percent of votes and 80  per- The illness and passing away of an incumbent in office led to a cent of seats in parliament for the incumbent party and a weakening in the political mandate. The overall ability of the 54 percent majority for the incumbent President). The 2012 government to act was weakened by the illness of President elections then brought a change in the government,38 with Yar’Adua. His condition turned critical after two years into the incumbent government again being successful—with a the government’s term. After he passed away, he was suc- higher number of seats won—in 2016. ceeded by Vice President Goodluck Jonathan, who also won • In Nepal, there was a significant commitment to pursue political the subsequent presidential elections in spring 2011 (win- reform, in particular toward inclusion of traditionally disadvan- ning 59 percent of the vote). The 2011 government contin- taged groups; however, the electoral mandate resulting from the ued along similar lines as the previous PDP-led one; while 2008 elections was highly fragmented resulting in a rapid suc- struggling even more to address violence and conflict in the cession of different governments. Due to the electoral system north/north-east. adopted and the context of social fragmentation, 25 parties A major change then occurred in the 2015 elections, won by the won representation in the Constituent Assembly elected in candidate of the opposition coalition, running on an electoral 2008. Consequently, the mandate and the decision-making platform that emphasized the need for improving governance ability of the government being formed based on the dis- and combating corruption. The presidential contest was won tribution of seats was weak. Moreover, as no political force by Muhammadu Buhari running as the head of the All Pro- had a majority or even near majority, governments changed gressives Congress (APC), and winning with 54 percent of frequently, with five different governments being formed the vote. In the parliamentary elections held simultaneously, between 2008 and 201339 when elections were next held. the APC also won majorities in the House of Representatives The level of experience and preparation of key government (225 out of 360) and in the senate (60 out of 109). figures varied across these rapid successions. Some of the successive Prime Ministers and Ministers of Finance had • In the Philippines, President Aquino won the 2010 elections on previously been in government, others had no experience; an electoral platform with a clear commitment to fight corrup- and all faced a very short and uncertain tenure. In 2013, tion and improve governance, and the staff in key PFM-related voter support shifted toward the more conservative Nepali positions were well prepared, but the electoral mandate was Congress (winning 196 out of 601 seats), and the Commu- mixed.41 Similar to previous presidents, Aquino won with a nists (175 seats), with the Maoists losing support (down to plurality of the vote (42 percent) rather than an outright 80 seats), representation remained fragmented. Changes in majority (no runoffs are being held). His party only held a government have remained frequent since then.40 majority in the Lower House with the Upper House being ruled by a coalition agreement. From the start, his term was • In Nigeria, the 2007 electoral mandate for the second govern- affected by a difficult relationship with the judiciary, which ment led by the PDP was still relatively strong, but was not diverted political attention and contributed to slowing down linked to a commitment to reforms. The PDP’s electoral sup- the reform process.42 port had already weakened relative to previous elections, and the ability to transfer power from one president to The President’s political capital in the legislature waned over the period of his administration, reinforced by the single-term rule that excludes presidents from running for a second consecutive 38. Presidential elections followed in 2013. The same political grouping won the parliamentary and the presidential elections. 39. Prachanda: August 2008–May 2009; Madhav Kumar Nepal: May 41. http://www.eiu.com.libproxy-wb.imf.org/FileHandler.ashx?issue_ 2009–February 2011; Jhala Nath Khanal: February 2011–August 2011; id=1187558903&mode=pdf Baburam Bhattari: August 2011–March 2013; and Khil Raj Regmi: 42. The outgoing President, Arroyo, had appointed a new Chief Justice March 2013–February 2014. The latter was a nonpartisan caretaker in May 2010, two days after the elections were held, but before she Prime Minister until the newly elected government took over. left office. The President elect opposed this appointment and sought 40. Sushil Koirala (Nepali Congress): February 2014–October 2015; the removal of the Chief Justice, which happened in 2012, following Khadga Oli (Communists): October 2015–August 2016; and Pushpa impeachment procedures by Congress (for failure to comply with asset Dahal (Communists) since August 2016. declaration provisions). The World BAnk  23 term. Internal government coherence also declined from to be seen as untainted by corruption. John Magufuli, who 2014 onward when the relationship between the President had held ministerial portfolios for labor, rural development, and the Vice President became strained.43 In 2016, the and agriculture, ran on a programmatic platform commit- Philippines held nationwide elections, which resulted in the ted to combating corruption and was regarded by the public win for a candidate, Rodrigo Duterte, from a different politi- as an outsider to the political establishment. He won with cal party, with 39 percent of votes. Duterte ran on a plat- 58 percent of the vote. As President, Magufuli developed a form centered on populism, anti-crime, and anticorruption. reputation for austerity following several measures to reduce The political style and policy orientation shifted significantly expenditures, improve the delivery of essential public ser- with the change in administration. vices, and fight corruption. • In Tanzania, the government that was elected in 2005 had a Overall, there is considerable variation across the five cases (relatively)44 strong mandate being backed by 80 percent of the regarding the high-level commitment to governance and public sec- vote, but was primarily committed to preserving the status quo; tor reforms that governments had at the outset of the specific periods while some reform momentum re-emerged in the 2010s. Tanza- reviewed here. Clearly, a political mandate for reforms is rarely nia has been governed by a single party, the Chama Cha Map- as strong as that held by the incoming government in Georgia in induzi (Party of the Revolution, CCM), since 1977, making it 2004. However, the turn to more reform-committed governments the longest running ruling party in the region. The system has in Tanzania and in Nigeria in 2015 shows that political swings in been relatively institutionalized and enabled a succession of favor of governance reform can happen in quite different coun- presidents every 10 years since 1985. Constitutionally, mul- tries. Furthermore, institutional rules matter for how strong politi- tiparty elections have been permitted since the early 1990s, cal mandates are likely to be. For example, Nepal had no threshold with the first competitive elections held in 1995. for parties for the allocation of parliamentary seats selected through proportional electoral rules for its 2008 elections, result- During President Kikwete’s first and second term, from 2005 ing in a very fragmented parliament. Electoral rules in the Philip- to 2015, the need to please different factions to maintain the pines, such as the President and the Vice President running on CCM’s unity affected the government’s interest in public sector separate tickets, for a single term, and without a runoff also favor governance. Increased political competition was accepted to distribution of power, which can involve some costs in terms of the continue the legitimacy of the ruling party.45 The govern- strength of political mandates and coherence of decision making. ment’s interests evolved in the specter of likely significant In the round of elections considered as ‘reform starting points’ future revenue windfalls from gas production under prepara- electoral commitment to reforms was thus strongest in Georgia, fol- tion during this period. The 2010 elections saw a further lowed by the Philippines. Between these two, the political mandate decline in electoral support to 60 percent (on a turnout of was considerably stronger in the former, while technical readi- 43 percent), still delivering a second term for the President ness was good in both. In Nepal, the largest political parties were and a continuation for the ruling party, but also signaling committed to a government more oriented toward the poor and growing discontent among citizens. In reaction, in 2012, excluded groups, but the political mandate remained diffuse. In the government adopted a large-scale reform strategy ‘Big Nigeria and Tanzania, the 2007 and 2005 governments held rea- Results Now’ (BRN), modelled on Malaysian reforms. sonably strong electoral mandates and had some good technical For the 2015 elections, the CCM selected a candidate who ran leaders in relevant positions, but neither had a commitment to on a strong anticorruption platform. To best confront increased reforms. competition, the ruling party sought to select a candidate The case studies also illustrate that political commitment to most likely to be popular, which also required the candidate reforms can change considerably over time and with subsequent elec- tions. In the Philippines, political commitment waned over time, due to a confluence of factors, and shifted to other policy areas 43. The Vice President is elected on a separate ticket and in 2010 following the 2016 elections. On the contrary, reform commitment was elected from a different party than Aquino’s by a narrow margin. Still relations were initially good due to a long family relationship; but became stronger in Nigeria and Tanzania from 2015 onwards, turned sour in 2014 with corruption allegations being raised against the after having stagnated for the previous decade. Vice President, possibly in the context of his popularity relative to the All five countries had reasonably well qualified technical teams preferred successor for the outgoing President. 44. Such a high level of support would be considered ‘very strong’ in a in place at the starting points considered here, reflecting the fact that fully democratic context; but in the wake of a one-party tradition was at PFM reforms have been pursued in many countries for some time. least partially a sign of discontent with the ruling party. 45. Formally, alternative parties have been allowed to register and The level of preparedness was relatively lower in Nepal, given compete elections since 1992. the background of the initial post-2008 Ministers of Finance; at 24  Political Economy of Public Financial Management Reforms the technical level, technical staff were committed and relatively economic and fiscal liberalism had delivered growth, but without experienced, but this level has also suffered from frequent rota- sufficiently benefiting many people especially in rural areas. The tions as discussed later in the report. The specific link between the incoming government continued to seek closer integration with political level and the technical level, including key appointments western countries, in particular the European Union (EU). Given in CFAs, is discussed further in Chapter 5. this significant external anchor, the commitment to governance improvements continued.47 In Nepal, key common goals of the left-wing parties succes- 3.2 Overarching Policy Goals sively heading the government were to foster socioeconomic trans- For most governments, PFM reforms are a means to an end, rather formation, to strengthen Nepal’s position in relation to India, and to than an end in itself; therefore it is important to understand the high- move from a unitary to a federal state; although agreement was lack- level policy goals, and how they relate to a potential interest in PFM ing on how to pursue these aims. Given the short duration of succes- strengthening. Higher-level goals may involve policy aims such as sive governments and the fact that none held a majority, pursuing ensuring better social cohesion and reducing poverty, seeking to major policy goals remained highly constrained. There were long transform an economy, or pursuing foreign policy goals such as periods of political deadlock and multiple tenure extensions for building a good relationship with particular countries or groups the Constituent Assembly. Eventually, the main political parties of countries, often involving foreign trade opportunities, and the came to an agreement on a new constitution in September 2015, extent of external aid dependency and relationships. Goals may supported by a large majority of voters. However, politics remained also be more short-term and ‘non-policy’ goals such as pursuing unstable even after this milestone was attained. survival of the current elite group or leader in power. In addition, For the Nigerian Government elected in 2007, key high-level high-level policy goals can be influenced by fiscal trends: fiscal policy goals included preserving and strengthening national unity. crises can focus attention on resolving these through PFM and Key issues in this regard were to manage the transition from a revenue-related reforms—even if a systematic impact on strength- southern to a northern president as well as addressing the inten- ening PFM appears to be lacking, as explored in the quantitative sifying conflicts first in the south and then in the north. External analysis.46 Conversely, expected fiscal windfalls may raise many dependency of the incoming government was low and had been new demands for spending. further reduced due to Nigeria reaching the completion stage of This section first considers general high-level policy goals as the Heavily Indebted Poor Country (HIPC) Initiative in 2006. This they can be identified retrospectively for the five countries analyzed. reduced Nigeria’s debt by US$18 billion, an overall debt reduction It then moves to looking at potential fiscal drivers, including rev- of about 60 percent.48 As a large country within its own region, enue and expenditure trends, debt dynamics, and the levels of aid external dependencies have not played a major role. relative to GDP. Preserving political stability and addressing security threats For the Georgian Government elected in 2004, two overriding remained key issues on the political agenda. Political stability came policy goals stood out. First, to resurrect the country as a function- under increased pressure when the sitting President passed away ing state and to reverse the extensive state capture and corruption. in May 2010. Political power then passed to the Vice President, Second, to more closely integrate with Western Europe and the Goodluck Jonathan, who again was a southerner. Considerations United States, and reunite the country’s territory. Related to this about preserving political stability dominated the 2011 elections, was a vision of establishing a libertarian model of a state, which in the shadow of the intensifying conflict with Boko Haram in the several key government leaders subscribed to. north/northeast. A major policy shift eventually took shape in the The opposition party elected in 2012 in turn ran on electoral 2015 elections, in which Jonathan lost to a challenger (Buhari), a promises to curtail the excesses of the Saakashvili Government, to northerner who had united several parties around a policy platform restore more normal relations with Russia, and to promote social of improving governance and combating corruption.49 equity. While the outgoing government had managed to trans- The Philippine Government formed in 2010 came into power form governance from a previous equilibrium of pervasive corrup- with three main policy goals. First, it had campaigned on a strong tion, it had also deployed harsh measures, including accusations anti-corruption platform, promising to reverse the course of the that it condoned the use of torture by law enforcement agencies. previous administration. Second, reviving a struggling economy Confrontation with Russia had been economically costly, and 47. Georgia is a priority country under the European Neighborhood Policy and the Eastern Partnership. In June 2014, the EU and Georgia signed 46. For example, if fiscal challenges motivate many governments to seek an Association Agreement, which includes a Deep and Comprehensive PFM reforms, but not all succeed to the same extent, the quantitative Free Trade Area (AA/DCFTA). analysis would not detect a pattern when focusing on the actual extent of 48. https://www.oecd.org/countries/nigeria/48481192.pdf. PFM improvements. 49. This also represented a shift back to a northerner as President. The World BAnk  25 with a strong focus on employment, while also promising a gradual In Georgia, before the start of the 2004 reforms, the state had elimination of the budget deficit. A third goal was to bring political been captured to a significant extent by ‘thieves-in-law’ (mafia type stability to areas with significant Muslim populations, by pushing networks that had formed during the Soviet and early post-Soviet forward negotiations with the Moro Islamic Liberation Front (MILF) period), which operated without restraint and often in collabora- in Mindanao, among other rebel groups. tion with the police, the justice system, and other parts of the state In Tanzania, President Kikwete, in office from 2005 to 2015, apparatus (Shelley, Scott, and Latta 2007). Government offices focused his efforts on four policy goals. First, continuing economic were for sale, in particular in customs, because much of the illicit reforms necessary after a period of structural reform and debt relief activities focused on imports and exports; while revenue collection that the country had undergone.50 Second, preserving national from tax and customs had become very low for an MIC.51 Given unity, by forging a political accord between two opposing parties in the absence of large natural resource rents and of large industrial Zanzibar. Third, tackling poverty reduction, and improving delivery centers and companies in the country, Georgia did not have the of public health, and of social services. Fourth, improving public large-scale oligarchs as they had emerged in Russia and some governance and fighting corruption. Following declining support other post-Soviet countries during the 1990s. for the CCM in the 2010 elections, the government subsequently A potential source of rent available in Nepal is having unique focused more on public perceptions of service delivery in the run- tourism assets, while similar to Georgia, there are few (other) natural up to the 2015 elections. It adopted the BRN program to address resource rents. As a relatively small country, business has been the lack of progress made toward service delivery improvements. dominated by a few large business groups, including one inter- The new administration elected in October 2015 campaigned nationally operating conglomerate. Most of these are involved in on promising to bring change with regard to reducing corruption and trading and light manufacturing of food and beverages, construc- elite enrichment and of delivering better services for ordinary citizens. tion and development, hotels, banking and insurance, and tele- As part of this, the government declared its objectives to improve communications. As elsewhere in Southeast Asia, many of these public revenues by curbing tax evasion and strengthening tax groups grew from original family traders; and were often given administration through efforts to increase compliance. some form of protected markets during the period of the royal government to promote economic and sectoral development.52 Business groups have sought favorable treatment in procurement 3.3 Rents and Rent Distribution contracts. Furthermore, in a socially and ethnically highly frag- In addition to electoral mandates and overall policy goals of govern- mented society, business houses are also seen as associated with ments, the nature of available rents and how they are distributed and a subset of groups. In Nepal, social networks and having the ‘right allocated typically plays a significant role in shaping reform options connections’ (afno manche) have a strong tradition, reinforced by and implementation (Levy 2014). Rents—defined as returns the royal form of government that formally only ended in 2008 exceeding the opportunity cost of resources that might otherwise (Subedi 2014). be deployed in a competitive market—are common in all econo- As a resource-rich country, Nigeria has been a classic example mies; particularly large-scale rents are available in resource-rich of a country with large-scale rents and associated significant elite economies (see also Barma, Kaiser, and Le 2012). The large nega- capture of the state and of available rents. Since the 1960s, large- tive impact of natural resource rents that appear in the quantita- scale corruption has affected many aspects of public sector opera- tive analysis also indicates that the nature and distribution of rent tions at all levels of government. This includes diverting revenue matters, and can affect PFM performance. flows from oil, mismanagement in public procurement, inflated The nature of the rents, and how they are being allocated is payrolls, and other forms of leakage (Okonjo-Iweala 2012). Non- crucial (Khan and Jomo 2000). While rent-seeking matters most transparent allocation of oil and mining licenses also negatively directly for economic policies, such as restrictions to competition, affect the revenue that accrues to the government (Global Witness rents and rent allocations can also affect the quality of PFM. Uncon- 2015). One mechanism for extracting rents has been oil bunker- trolled rent-seeking can undermine revenue collection and conse- ing, which is estimated to have accounted for losses to the treasury quently affect cash management. Another aspect of PFM that can of US$6.7 billion in 2013.53 Beneficiaries of an economy built suffer significantly from rent-seeking is procurement. Where rent- seeking is significant and benefits narrow groups, it can become 51. Entrenched Corruption Begins at Georgia’s Border (2002). http:// more difficult to effectively pursue governance improvements even www.eurasianet.org/departments/insight/articles/eav062702.shtml. 52. http://www.gefont.org/uploads/publications/9733_The%20Big%20 for a government that has made electoral commitments to reform. Business%20Houses%20in%20Nepal.pdf; http://bossnepal.com/top- group-companies-nepal/ 53. William Wallis, “Nigeria: The big oil fix,” Financial Times, May 26, 50. Tanzania completed the HIPC program and received debt relief in 2015. Other sources cite a somewhat lower amount of US$1.25 billion 2001. https://www.imf.org/en/News/Articles/2015/09/14/01/49/pr0148. per quarter or US$5 billion per year. 26  Political Economy of Public Financial Management Reforms around rents from the oil sector have been elites composed by politics with multiple family members over time—and the share classes of politicians, connected business interests, high-ranking of seats controlled by individuals linked to such dynasties signifi- civil servants, and military officers (Lewis and Watts 2015). cantly exceeds similar trends in most other countries.54 The fam- In addition, Nigeria also has large business groups, which have ily groups acquire both power and wealth, thus entrenching their flourished due to several factors, including favorable regulatory treat- electoral advantages. ment, as well as business acumen. The latter includes an ability to In Tanzania, economic liberalization and privatization in the work in a diverse market that international competitors found diffi- late 1990s offered an opportunity to political insiders to purchase cult, not least due to widespread corruption (Akinyoade and Uche formerly state-owned companies, or channel such purchases to asso- 2016). Nigeria’s richest business family is engaged in cement ciates, who would in turn provide funding support to electoral cam- rather than oil, riding a wave of population growth and expansion paigns. The patronage relationship seems to have evolved as a of buildings and infrastructure. response to the limited progress made in pursuing genuine reform Although Nigeria has established a number of dedicated institu- strategies. There have been repeated allegations of collusion tions to constrain corruption and rent-seeking, these have often not between senior politicians and large businesses around procure- functioned as intended. Transnational and local companies have ment contracts. The Richmond case of 2006 provides an example frequently engaged in shadow practices to participate and ben- of competing powerful interest groups within the executive, with efit from available rents. Elites benefiting from large-scale rent-­ close ties to business, redistributing resources to secure elites’ seeking include various ethnic, regional, and religious groups. support for the ruling party. At the same time, as part of decentral- Nonetheless, the system is deeply exclusionary, limiting most ization, local governments have been given increasing control over gains to a narrow group of elites from various communities. local service delivery and its finances, whereas line ministries have Against this background of deep-seated rent-seeking and scale lost both legal control over resources and associated rent-seeking of nonproductive rents available, the government elected in 2015 has opportunities. pledged to improve governance and bring about a state that works for Available natural resource rents in Tanzania are expected to all citizens. This electoral shift occurred against a background of a increase significantly. Major natural gas deposits were first con- sharp decline in the value of oil since mid-2014. Sharply shrink- firmed in 2011, and additional deposits have been discovered ing fiscal resources (discussed further also in Section 3.3), began since then.55 Such an initial natural resource boom period creates to negatively affect the government’s ability to deliver services and new rent-seeking opportunities and in turn affects the stakes in pay civil servants, as well as to co-opt various groups. Main chal- elite competition. lenges included a worsening conflict and perceptions of economic Looking at the nature and scale of rents across the five coun- neglect in northern Nigeria, after previous high spending to resolve tries, Nigeria stands out in terms of rents being available from one par- the conflict in the south. Therefore, the new government pledged ticular source (oil production), and its direct and negative impacts on a to improve the collection and use of fiscal revenue. However, swim- government’s incentives related to manage public resources well. As is ming ‘against the tide’ of rent-seeking interests and achieving suc- discussed further below, rising oil prices in the late 2000s provided cess in a period of much reduced fiscal resources and economic an easy source of additional revenues and, consequently, disincen- deceleration has proven to be extremely challenging. tives to seek revenue collection through taxing the non-oil economy, In the Philippines, public procurement and industrial policy as well as careful husbanding of expenditures. At the same time, have been two major sources of rent. Natural resources have not consistent with the findings from WDR 2017, the discussion also played a significant role. Further specific rent-seeking opportuni- indicates that most countries have some form of entrenched inter- ties exist in the domestic manufacturing sector and concentration ests which benefit from the fiscal and regulatory status quo. of ownership in the financial industry and airlines (North et al., It is important to stress that rents and rent distribution are not 2013). In an effort to curb corruption, President Aquino, elected static, but evolve over time due to shifts in opportunities. Sources of in 2010 on a good governance platform, continued the procure- rent change in value over time, most notably with regard to natural ment reform started under President Estrada and initiated other resource rents following the end of the ‘commodities super cycle’ fiscal transparency-related initiatives. An important aspect of the Philippines is the presence of well- 54. According to Acemoglu and Robinson (2013), “60% of congress- established political dynasties (North et al., 2013; Tusalem and people elected in 2007 had a previous relative who were also in congress.” See: http://whynationsfail.com/blog/2013/1/9/political- ­ Pe-Aguirre 2013; You 2014). Highly unequal land distribution dynasties-in-the-philippines.html. See also Cruz, Labonne, and Querubin with a majority of land owned by a small group of landowners, (2014). emerged in the first half of the 20th century and served as a 55. For example, see Oilprice.com: “$8bn Natural Gas Find Re-affirms Tanzania’s Status as a Gas Giant,” James Burgess, March 28, 2016. basis for political dynasties. A key feature of dynastic politics is http://oilprice.com/Energy/Energy-General/8-Billion-Natural-Gas-Find-Re- that the same families are represented in regional and national Affirms-Tanzanias-Status-As-Gas-Giant.html. The World BAnk  27 which has affected oil-related rents in Nigeria, or the discovery Regarding the second presumed dynamic, in the cross-country of large-scale new resources in Tanzania. Technological change analysis we find a weak positive relationship between tax revenue and can disrupt rents, for example, as in the telecommunications PFM performance as measured by PEFA. The five case studies also and to some extent in the financial sector over the past decade. show a positive relationship between revenue performance and Some countries, such as Georgia and Nepal inherently have fewer PFM quality (as discussed further below). Countries with increas- large-scale sources of rent, which in principle, makes it easier to ing revenues and expenditures experienced strengthened PFM focus economic ambition on ‘value adding’ activities. However, performance, while countries with lower, stagnating, or declining privileged access to contracts or licenses, for example, for imports, budgets saw less PFM progress. However, as is discussed fur- can still bestow important advantages on some economic actors ther in Chapter 4, the presumed mechanism—that this relation- over others. Conversely, voters in the countries studied here have ship is due to a greater demand for financial accountability from at times favored governments committed to constraining rent- citizens—is not clearly supported by the case studies. ­ seeking and to improve governance. As the cases also indicate, The following sub-sections consider several fiscal dynamics to electoral support for better governance is not consistent over time capture the situation of the five countries analyzed: Revenue and and can erode if economic benefits are not forthcoming. expenditure trends over the period 2000–2015, their level of debt (gross debt relative to GDP), and levels of aid dependency. It also sets them in a context of performance by income groups globally. 3.4 Fiscal Trends—Limited Pressures Revenue and Expenditure Trends to Reform Figure 3.1 reflects global trends in revenue and expenditure over the Existing literature and research provides two main hypotheses on past 15 years by income group. The figure indicates that the gap potential links between fiscal dynamics and government commit- between income groups have somewhat declined over this period. ment to seeking PFM reforms. One assumption is that fiscal crises High-income countries’ spending levels have remained relatively or austerity periods can be a trigger for PFM reforms (de Gramont stable, at around 35 percent to 40 percent of GDP, and somewhat 2014; Krause 2009; Pretorius and Pretorius 2009). Reforms ini- higher during the global financial crises in 2008–2009. UMICs tiated in Tanzania in the 1990s, in Turkey in the early 2000s, and LICs have seen the highest growth rates in both expenditure and in several OECD countries are widely seen as having been and revenue levels, narrowing the gap between income levels. precipitated by fiscal challenges. The specific reforms initiated LMICs have also seen a rise in spending and revenue levels, but have differed—focusing on greater efficiency through New Public at a slower pace compared to the other groups. Most recently, rev- Management (NPM)-inspired reforms in several OECD countries, enue appears to level off especially in UMICs and LICs, possibly versus strengthening and centralizing controls over cash manage- due to the end of the commodity super cycle, while expenditures ment in Tanzania. still continue on an upward trend—which may also signal some Another strand of literature has proposed that a higher share renewed greater fiscal challenges ahead. of tax-based revenue leads to better PFM through stronger citizen Overall, the graph indicates a global trend toward increased demand for better services and greater accountability (Moore 2004; revenue and increased expenditures over this period. One probable Prichard and Leonard 2010). This would imply that countries with reason for this is a growing expectation of what states should larger (non-resource related) fiscal revenues experience a rising deliver, or coproduce with the private sector: including free and demand among citizens for spending funds effectively, which in universal basic services, infrastructure, effective regulation, con- turn should result in a process of strengthening PFM systems. sumer and social protection, and internal and external security.56 The analysis of correlations could not find a link between the At the same time, rising amounts of funds collected and spent by experience of fiscal crises and progress on strengthening PFM, while governments is one of the key reasons why further strengthening the case studies indicate that for each country, fiscal trends are PFM is critically important. important, but how this happens varies. For the period tested, the Against this global background, the countries covered in the statistical relevance is weak or disappears depending on time lags case studies show widely varying fiscal performance, with some used (four versus five years). As is discussed further below, the observable parallels between fiscal and PFM performance trends. All case studies indicate that there is a potential link between facing the five countries reviewed in principle sought to achieve some fiscal challenges and political interest in PFM reforms—but with variation in whether and when strengthening of PFM systems actu- ally proceeds, and a limited incidence of fiscal shocks during the 56. The tendency toward increased government expenditure accompanying economic development was identified as ‘Wagner’s law’ main period reviewed. This may explain the absence of a system- (named after Adolph Wagner, 1835–1917). High-income countries have atic relationship across a large number of countries. sought to constrain further growth of revenue and expenditures, but still show a slight upward trend on average. 28  Political Economy of Public Financial Management Reforms Figure 3.1. Revenue and Expenditure Trends (2000–2016) by Income Group 45.0 40.0 Revenue LIC 35.0 Revenue LMIC 30.0 Revenue UMIC % of GDP 25.0 Revenue High 20.0 Expenditure LIC 15.0 Expenditure LMIC 10.0 Expenditure UMIC 5.0 Expenditure High 0.0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Source: IMF World Economic Outlook database, October 2016. Figures reflect total revenue, including grants; and general government total expenditure. Note: Income groups are identified using the World Bank list of economies (December 2016). increase in revenue. However, only two—Georgia and Nepal—actu- situation. Especially after the end of the conflict in 2006, Nepal ally achieved notable revenue increases (see Figure 3.2). Notably, experienced an increase in revenue and expenditure growth rela- these countries also progressed relatively further on strengthening tive to earlier years. By 2010, both revenues and expenditures PFM systems. In the Philippines, revenues have flatlined at a com- reached about 18 percent of GDP up from 12 percent in 2001; paratively low level, while PFM saw some progress—with a strong and by 2013/14, total revenues including grants reached a high tilt to certain areas, but not others as noted in Chapter 2. Tanzania of 21.6  percent of GDP. The main proximate drivers of revenue saw a boost in revenue collection in the early 2000s, when it also expansion have been high import growth fueled by remittances on registered continuing improvements in its PFM systems, and then the one hand and administrative reforms in the Inland Revenue saw a flatlining in revenues since 2007, when PFM performance and Customs Departments on the other hand. Revenue perfor- also began to stagnate and slide backwards.57 mance as a percentage of GDP is now even somewhat above other Georgia achieved substantial progress in both revenue perfor- LICs (see Figure 3.1 above). In parallel, as discussed in Chapter 2, mance and PFM. At the start of the period covered here, Georgia had PFM performance improved between the two PEFA assessments in a low revenue to GDP ratio for a middle income country, of around 2008 and 2015. Tracing CPIA ratings for PFM further back to the 15 percent, and suffered from associated cash management chal- early 2000s indicates that this strengthening was in part a recov- lenges. Increasing revenue was an early priority for the govern- ery toward earlier levels of performance.58 ment following the Rose Revolution in 2003/04. The government Nigeria stands out as a negative outlier relative to global trends applied considerable effort to reduce tax evasion and broaden the as well as the other case studies during this period, because rev- tax base, which resulted in a large rise in revenue, peaking at enue mobilization declined substantially between the early 2000s and 31 percent of GDP in 2008 (see Figure 3.2). Higher revenues in 2015 relative to GDP. While Nigeria benefited from rising global oil turn allowed the government to raise spending—including invest- prices until the early 2010s, revenue as a percent of GDP had ments in better services as well as a better functioning public already declined sharply since 2000 (see Figure 3.2)—in the con- sector. The specific sequence of reform steps is discussed further text of a significantly expanding GDP combined with an increas- in Chapters 4 and 5. ingly narrow tax base. From a peak of over 30 percent of GDP in Nepal saw steadily increasing revenue collection and some the early 2000s, total revenue and expenditures dropped to only progress on PFM strengthening—despite a politically complex 8.5 percent and 10.5 percent of GDP respectively by 2015, and 57. For CPIA-13 (budgeting and financial management), Tanzania achieved the highest rating in 2005 and 2006, at 4.5 out of a maximum rating of 6, one of the highest ratings of any IDA country on this 58. The CPIA-13 rating for Nepal was 3.5 in 2000, then declined to a indicator. low of 2.5 by 2010, again increasing to a rating of 3 since 2013. The World BAnk  29 Figure 3.2. Revenue and Expenditure Trends (2000–2016) Georgia Nepal 40 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Revenue (incl. grants) Expenditure Revenue (incl. grants) Expenditure Nigeria Philippines 40 40 35 35 30 30 25 25 20 20 15 15 10 10 5 5 0 0 2011 2000 2001 2002 2003 2004 2006 2007 2008 2009 2010 2012 2013 2014 2016 2005 2015 2011 2000 2001 2002 2003 2004 2006 2007 2008 2009 2010 2012 2013 2014 2016 2005 2015 Revenue (incl. grants) Expenditure Revenue (incl. grants) Expenditure Tanzania 40 35 30 25 20 15 10 5 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Revenue (incl. grants) Expenditure Source: World Economic Outlook (WEO) 2017. 30  Political Economy of Public Financial Management Reforms even further in 2016.59 For several years, readily available revenue A stimulus package was adopted in 2008/09 to cushion the that continued growing in nominal terms made it possible for the negative impacts of the global financial crisis on the country, and government not to be overly concerned with this decline and not the 2010 elections further prompted spending increases. Following to seek strengthening revenue collection, although problems with the 2010 elections, and bolstered by the expectation of future cash management were clearly present. Debt relief achieved in increases as anticipated hydrocarbon revenues come onstream, 2006 further facilitated this policy stance. Tanzania increased its public spending as a share of GDP further The fiscal situation worsened dramatically further in 2014– above the level of revenue increases (Tilley 2013), with expendi- 2016 when oil prices declined sharply, triggering a decline in revenue tures reaching 20 percent of GDP as of 2015. even in nominal terms. As a consequence, the incoming govern- Debt Dynamics and Levels of Aid Relative to GDP ment in 2015 was confronted with an urgent need to strengthen revenue collection as well as to improve spending. Among the five Related potentially relevant drivers include debt dynamics and aid case studies, Nigeria has experienced the most acute fiscal crisis, dependency. High or rising debt levels and growing risks of debt albeit only at the very end of the analysis period (in addition, as unsustainability could be a trigger for fiscal austerity and associ- noted above, Georgia had gone through a period of fiscal crisis in ated efforts to make the most of available funds. Higher debt lev- the 1990s and early 2000s). els could also increase the relative leverage of aid and associated In the Philippines, revenues remained relatively steady for more demands for strengthening PFM systems. It is therefore impor- than 15 years relative to GDP, while some improvements in PFM were tant to consider the debt levels and trends across the five cases being achieved. Since 2000, revenues and expenditures have hov- analyzed. ered at an average of 18 percent and close to 20 percent of GDP In all five case studies, gross debt declined in the early 2000s respectively. While the size of the budget increased by 134 percent and remained at sustainable levels until 2015, albeit with debts again in nominal terms over the last decade, public spending however, growing in some cases. As Figure 3.3 indicates, significant declines did not change as a percentage of GDP. Compared to its income in debt levels were common in the early 2000s—due to a com- group peers, the Philippines fell behind, as LMICs on average saw bination of the government’s own efforts as well as debt relief in an increase of their revenue and expenditure relative to GDP dur- several cases. Since 2007, the debt trends among the five coun- ing this period (Figure 3.1), reaching closer to 30 percent of GDP. tries have become more mixed. Nonetheless, none of the countries Tanzania achieved gains in revenue collection mainly in the analyzed here has faced significant problems with unsustainable early 2000s, in parallel to improvements in PFM performance, fol- debt levels during this period, while some might again face chal- lowed by a leveling off in revenue and a gradual decline in PFM lenges in future. performance. Between 2000 and 2005, Tanzania achieved an Among the five countries, Nigeria in particular saw a decline increase in its level of revenue collection by about 5 percent of in debt levels due to debt relief. According to IMF figures, Nigeria’s GDP. While these efforts allowed the revenue to reach 16 per- national debt was reduced from 88 percent of GDP in 2001 to cent of GDP, it still falls short of LIC averages, which increased 11  percent of GDP in 2006 following the achievement of HIPC even further to above 20 percent on average. Similar to Nigeria, criteria and being granted relief (see Figure 3.4). As is discussed Tanzania also benefited from debt relief achieved in 2001. Both further in Chapter 5, the prospect of debt relief provided some revenue collection and PFM efforts were grounded in an earlier impulses for PFM reforms up until 2006, while subsequently, reform period from the mid to late 1990s. Those reforms included some of the efforts that had been initiated but not completed establishing an independent revenue authority, with strengthened remained lingering for several years. incentives to pursue revenue collection. On the expenditure side, Both Nepal and the Philippines saw a steady decline in their cash management problems were brought under control (Diamond gross debt without receiving debt relief. The Philippines achieved and Khemani 2005; Gray and Khan 2010), and overall, Tanza- a continuous decline, from 64 percent in 2000, to a gross debt nia reached a fairly high level of PFM performance for an LIC by of 35 percent of GDP by 2015. As aggregate spending was kept the first PEFA assessment undertaken in 2006. However, in both broadly in line with available revenue, external and overall gov- areas, there was backsliding over the decade from 2005 to 2015. ernment debt have declined over the past decade. Thus, while corruption has been a major concern in the Philippines and reduc- ing corruption was a major electoral commitment of the incoming government in 2010, fiscal or debt pressures were less urgent. 59. 2015 IMF WEO figures. A further issue is highlighted by Nigeria: uncertainty about GDP size and growth can make it difficult to fully Nepal achieved a significant decline in its debt from almost identify revenue trends, given that growth and inflation typically lead to a 60 percent in 2000 to 24 percent of GDP in 2015. Much of this is continuous expansion of nominal revenues. Only after Nigeria’s GDP was re-based in 2014, did it become clear how dramatically revenues had the result of budget surpluses from sluggish execution of capital declined relative to the size of the economy. expenditure and strong revenue growth. Nepal was one of the few The World BAnk  31 Figure 3.3.  Debt Trends by Income Groups 140 120 100 LIC % of GDP 80 LMIC 60 UMIC 40 High 20 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Source: WEO database, October 2016 (general government gross debt). Note: Income groups are identified using the World Bank list of economies (December 2016). Most of 2016 data are estimates of the WEO. Figure 3.4. Gross Debt Trends, Five Case Study Countries (% of GDP) 80 70 60 50 40 30 20 10 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Georgia Nepal Nigeria Philippines Tanzania Source: WEO 2017 (general government gross debt). countries that in principle qualified for the HIPC Initiative started In contrast, Georgia and Tanzania initially saw a decline in 1996, but eventually decided not to request debt relief. Fol- in gross debt up until 2006–2007, but increases since then. Both lowing the end of the civil war, key government and economic countries had somewhat lower initial debts (about 50 percent of stakeholders felt that engaging in the HIPC Initiative might hinder GDP) compared to the other case studies in the early 2000s and rather than boost opportunities for economic revival.60 both saw a similar initial decline—Tanzania due to the debt relief programs and Georgia due to increased revenue and significant 60. See https://thehimalayantimes.com/business/nepal-shies-away-from- privatization receipts. However, since 2008, domestic and foreign hipc/. Nepal’s eligibility is discussed in Pant and Subedi (2006). borrowings in the two countries have increased again, raising gross 32  Political Economy of Public Financial Management Reforms debt to 42 percent and 35 percent of GDP for Georgia and Tanza- The analysis found that there is a correlation between having bet- nia, respectively. ter PFM systems and greater budget credibility—as expected, but Importantly, during previous decades, debt levels had been not between stronger systems and lower deficits (Fritz, Sweet, and much higher in several of the five countries: In Nepal, debt had Verhoeven 2014). Thus, even after countries have undertaken sig- peaked at around 60 percent in 1999/2000 and the Philippines nificant PFM reforms and can keep spending in line with approved saw external debt reaching close to 100 percent of GDP in the budgets, governments may still decide to overspend. late 1980s.61 In Nigeria, external debt reached over 200 percent The importance of aid relative to the economy and the budget is of GDP by 1993, and in Tanzania 166 percent of GDP in 1993– a further potentially relevant fiscal aspect. In particular, the extent 94. Several of the countries carried high-debt burdens for several to which public expenditures depend on external contributions years, before bringing these under control. Thus, the period from has been considered as a potential driver for countries to pursue 2000–2015 is one of comparatively low external public debt in PFM reforms (Therkildsen 2000 and 2001), but also conversely these countries. as potentially undermining incentives for public sector and PFM Overall, for the 15 years reviewed here, there has been no clear reforms (Bräutigam and Knack 2004). In the quantitative analy- pattern or link between debt trends and success of PFM reforms, in sis, we found no statistically significant relationship between vari- the context of benign economic trends. Debt levels have mostly been ous measures of aid dependency and PFM performance, in line reduced, in some countries facilitated by significant debt relief, with earlier findings by De Renzio, Andrews, and Mills (2011). as well as due to significant economic growth during this period. The fact that there is no clear pattern between levels of aid Given these trends, for the past 15 years, debt levels did not play dependency and efforts toward PFM reforms also largely holds true a role in inducing efforts at improving PFM. To further explore across the five case study countries. Figure 3.5 maps the level of whether high debt levels trigger PFM reforms (and with what prob- ODA relative to GNI for the five countries, and Table 3.2 pres- ability) would require considering additional time periods and a ents a broader range of measures, including the annual U.S. dollar focus on cases which experienced unsustainable debt. amount per capita and the share of aid in total public expendi- Looking at the reverse relationship, at least one case (that is, tures. The spike in aid to Nigeria in 2005–06 is due to debt relief Georgia) contradicts the potential assumption that better PFM systems received during those years. Generally, as all the economies expe- will be associated with constraining debt levels. This observation rienced substantial growth during this period, the aid to GDP ratios is consistent with findings from the quantitative research that somewhat declined. looked at the effects of the quality of PFM on fiscal aggregates. The case studies confirm that aid dependency is not an unam- biguous driver of PFM reforms. Tanzania remained the relatively most aid-dependent country over the period, but as noted in Chap- 61. For Nepal, the Philippines, and Georgia, data are available both for central government debt and for external debt. For Nigeria and Tanzania, ter 2, saw a slippage on PFM performance from an initially inter- data are only available for ‘external debt’. WDI Central Database. mediate level. Georgia and Nepal both have some aid dependency, Figure 3.5.  ODA (% of GNI), Five Case Study Countries 16 14 12 10 8 6 4 2 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Georgia Nepal Nigeria Philippines Tanzania Source: WDI 2017. The World BAnk  33 Table 3.2.  Levels of Aid Dependency Georgia Nepal Nigeria Philippines Tanzania Averages for LICs Averages for MICs Aid per capita 121 25 16 5 50 48 10 (average 2000–2015) Aid in % of GNI 5.5 5.6 1.5 0.3 9.9 11.9 0.4 (average 2000–2015) Source: WDI Central 2017. and additional external incentives (EU association) exist for Geor- are whether leading politicians made commitments to reforms dur- gia as reflected in Section 3.2. Particularly strong PFM reforms ing the past elections, combined with the strength of mandates were observed in Georgia, and intermediate progress in Nepal. received. Furthermore, Section 3.2 explores high-level policy Nigeria and the Philippines both receive little aid relative to GDP. goals, and whether pursuing PFM reforms provided a good fit with In Nigeria, aid as a share of central government spending is still such goals. Georgia stands out as a case with strong public sector substantial—but the central level only accounts for a limited share reform commitment following the 2004 change in government, of overall spending due to a federal system. The two countries which helped to break out of a previously poor governance equilib- saw diverging trends on reforms—with some intermediate PFM rium. Subsequent governments then continued the path of improv- strengthening in the Philippines, but very little progress in Nigeria ing public sector performance. between 2006 and 2015. In Nepal and the Philippines, political pledges to improve At the same time, the case studies also indicate that aid is not governance and to make the state work better for the poor played totally delinked from PFM reforms. As noted above, there is some important roles in the 2008 and 2010 elections respectively, but indication that the promise of debt relief (provided in 2006) in both cases, there were also important constraints at work. In induced initial PFM and fiscal reform measures in Nigeria and Nepal, electoral fragmentation and a political focus on a new led to the initiation of a wider PFM reform agenda—albeit one constitution, and how to constitute intergovernmental relations, that was subsequently left lingering when the main incentive had made it challenging to pursue reforms consistently. In the Phil- been realized. In Tanzania, the pressures from high debt, high- ippines, vested interests against a strong reform push from the aid dependency, and the prospect of debt relief combined in the central government were particularly significant; and reform prog- late 1990s, and this period was associated with significant PFM ress remained particularly uneven. In Tanzania and Nigeria, over reforms. In the 2005 CPIA, Tanzania reached a 4.5 rating (on a the past decade, electoral promises focused more on preserving scale of 1–6) for budget management, indicating strong PFM per- peace, national unity, and the status quo, rather than governance formance, before beginning to slide backwards. Thus, fiscal pres- improvements, at least until the 2015 electoral cycle. The latter sures and the prospect of debt relief provided important incentives brought in much greater government commitment to reforms, and to pursue PFM reforms for some time. However, reforms pursued there are at least some initial indications of progress. in reaction to such pressures were only partially sustained over As the case study experiences indicate, the relationship time, and underlying issues subsequently resurfaced. between fiscal trends and efforts toward PFM reforms is more complex than might be expected. Across the globe, many governments in low- and middle-income countries have sought to increase revenue 3.5 Chapter Summary and expenditures in recent years (as reflected in Figure 3.1). The This chapter covers the broad picture of incentives and constraints five cases suggest that governments in countries with low revenue which governments face, with the aim to identify whether and how to GDP ratios typically seek revenue increases. As is discussed these factors shape the interest in and commitment to PFM reforms. further in Chapter 4, this was a powerful motivation in Georgia, As emphasized by WDR 2017, commitment is a key feature with and seems to also have played a role in Nepal as well as Tanza- regard to the likelihood that policies are actually and effectively nia. At least for Georgia and Nepal, this went hand in hand with implemented. improving expenditure management. However, as is discussed fur- As discussed in Section 3.1, it is possible to provide at least ther in the following chapters, there is less of a clear link between an approximate account of the political commitment to governance increased revenue and increased demands for accountability—and or public sector reforms and strengthening, and the cases suggest an for PFM improvements specifically—than is assumed in parts of association with the pursuit of PFM reforms. Important ingredients the literature. 34  Political Economy of Public Financial Management Reforms Debt seems to have been a greater driver of PFM reforms in the remained non-urgent—and not clearly visible—as revenue was ris- 1990s and up to HIPC, but has not been a strong driver for the coun- ing in nominal terms and GDP was underestimated, thus masking tries and years covered here.62 The decline in debt ratios indicates the relationship. The picture in Nigeria only turned into a strongly that governments have sought to maintain or improve fiscal stabil- negative revenue trend once GDP was re-based in 2014. Revenue ity and to avoid unsustainable debt levels. However, levels of debt even exceeded expenditures for several years before the global evolve continuously. There is the risk that even in countries where financial crisis which depressed oil prices in 2008–2009, facili- PFM improvements have been implemented, and where debt and tating a stimulus during those years, and followed by a renewed deficits were brought under control through an increase or recov- rally in prices from 2010–2014. A substantial fiscal crunch then ery of revenue levels, new periods of excessive expenditures can emerged since mid-2014 due to the drop in oil prices combined emerge. Among the case studies, this is especially an issue in with other factors and there are some early indications that this is Georgia.63 contributing to a reviving of efforts at fully implementing earlier Fiscal shocks were rare during the period reviewed, except in conceived PFM reform measures. Nigeria toward the end, and Georgia at the very outset. While the level The following two chapters zoom into how PFM reforms pro- of revenue relative to GDP had declined in Nigeria for a while, this ceeded in greater detail. Chapter 4 discusses a set of four more direct drivers which are assumed to affect PFM reforms: institu- tional arrangements and legal provisions, the political-technical interface, focusing in particular on the selection and tenure of 62. For a review of PFM improvements related to HIPC efforts, see de Renzio and Dorotinsky 2007. Ministers of Finance, the role of the demand side for PFM reforms, 63. Another case is Mongolia where external debt was brought down and that of ideas and ideology. Chapter 5 then hones in on the from 100 percent to 40 percent of GDP between 2003 and 2008, and a range of PFM reforms were implemented in the 1990s and 2000s. specific aspects of PFM and how each of these have developed Nonetheless, debts rose to over 180 percent of GDP by 2014. across the cases. The World BAnk  35 Institutional Conditions and Dynamics for Advancing PFM Reforms 4 O verall political dynamics and policy goals and fiscal trends, accountability, but also the adoption of legal changes, such as a as discussed in Chapter 3, are important motivators for new organic budget, procurement, or external audit law. A fourth pursuing PFM reforms, specifically for providing high-level key aspect is intergovernmental arrangements—that is, the role political backing to such efforts. This chapter hones in on some played by subnational entities in managing public funds, and their of the specific challenges and opportunities that are embedded rights and responsibilities in this regard. in the institutional environment in which PFM reforms are being sought (see also Dressel and Brumby 2009). 4.1.1 Public Finance Laws In particular, this chapter considers the institutional and While laws often do not fully reflect what happens de facto—as legal setup (and potential efforts at changing this) including the underlined also in WDR 2017—legal provisions, their status, and the status of key legal provisions, the organization of central finance degree to which they need to be updated or changed matters(see also functions, the relationship between the executive and the legisla- Lienert 2013a). With regard to organic budget laws or broader tive, and the nature of intergovernmental relations and the related laws on public finance, countries range from having very old and distribution of responsibilities for managing public funds. In addi- outdated laws (Nigeria) to those that have been updated multiple tion, the chapter looks at the link between the political and techni- times in the 1990s and 2000s (for example, Tanzania). Apart from cal level and how this influences reform commitment and traction. organic budget laws, key public finance laws include those govern- ing public procurement, debt, and/or on fiscal sustainability, as 4.1 Institutional and Legal well as supreme audit offices. On the revenue side, this is comple- mented by tax laws and laws governing other forms of revenue. Arrangements Table 4.1 summarizes the status of the main PFM-related laws One important element for making reforms happen are existing insti- across the five countries covered. Three of the five countries are tutional and legal arrangements. These concern both the directly currently in the process of updating or, for the first time, introduc- PFM-specific institutions—which matter for how far the system ing comprehensive organic budget legislation (Nepal, Nigeria, and is from functioning well—as well as the institutions of the wider the Philippines). Georgia and Tanzania have recent organic budget authorizing environment. This wider institutional setting can facili- legislation adopted in the 2000s. The latter two and Nepal also tate making changes and improvements, or in a negative scenario, have current legislation pertaining to external audit. In Nigeria, the can make this particularly difficult. In line with the importance of role of the Auditor General is mainly regulated by the Federal Con- these issues, institutional and legal aspects of PFM systems have stitution.64 Efforts to adopt a full-fledged law failed in the 2000s, started to receive greater coverage in PEFA and other PFM-related and again failed before the 2015 elections. Following the change reports in recent years. in administration, a new effort was started in 2016. In the Philip- Several types of institutions and institutional relationships mat- pines, the main legal basis for external audit is similarly a set of ter. A core aspect is the legal provisions that establish the rules constitutional provisions as well as a presidential directive, rather of how public funds are to be allocated, used, and accounted for. than a dedicated law.65 This legal situation reflects challenges A second aspect is the structure of the Ministry of Finance (MoF) in the relationship between the executive and the legislature itself, including how it is internally organized, what functions it comprises, and what other institutions play a role in managing 64. Articles 85 to 87 for the Federal Level, and Articles 126ff concerning key PFM aspects (see also Allen et al. 2015). A further critical the requirement of states to establish state level Auditor Generals. 65. See: http://www.coa.gov.ph/index.php/2013-06-19-13-06-03/ institutional aspect is not only the relationship between the execu- constitutional-mandate; and http://www.coa.gov.ph/phocadownload/ tive and the legislative regarding budget approvals, oversight, and userupload/Issuances/rules-and-regulations/PD1445.pdf.   37 Table 4.1.  Key Public Finance Laws in the Five Case Study Countries Country Budget Preparation and Execution External Audit Procurement Debt Georgia Budget Code (2009) Audit Office (2008) Law on State Procurement Economic Liberty Act (2010) (2011) (limiting government spending and borrowing) Nepal Financial Procedures Act (1999) and Audit Act (1991) Public Procurement Act Public Debt Act (2002) Regulations (2007) Under development: New Fiscal Responsibility and Budget Management Act Nigeria Public Finance (Control and Management) Office of the Auditor General Public Procurement Law Fiscal Responsibility Act Act (1958) regulated by the 1999 (2007) (2007) Last attempt to pass a new law: 2009 Constitution, adoption of separate law failed in the 2000sa Philippines No unified PFM act External audit is regulated by Government Procurement Foreign Borrowings Act Budget Reform Bill drafted and initial the 1987 Constitution and Reform Act (2003) (1966); Central Bank Act discussions in parliament held in 2016; a Presidential Directive (1993) revised bill was submitted in May 2017 No. 1445, issued 1978 as amended No separate act Tanzania Public Finance Act (2001; revised 2004) Public Audit Act 2008 Public Procurement Act Loans, Grants and (2016; amending the Act Guarantees Act (1984) of 2011)b Source: Authors, based on publicly accessible laws and case study reports. Note: a. An original Audit Act dates back to 1957 (that is, predating independence). This is partially overridden by the 1999 Constitution, but not absolutely. A separate standalone Audit Bill was resubmitted to the legislature for approval in early 2016. A new Audit Bill was passed by the legislature before the 2015 elections, but was not assented to by the President in time to become law. b. See Chapter 6. The PPA 2011 replaced an earlier act adopted just 7 years earlier (2004). The 2011 Act only became effective in December 2014, following extensive delays in publishing the regulations. Since the 2011 Act was viewed as not providing value for money in procurement, amendments were passed in June 2016. discussed below: where this relationship is difficult and prone to When and how a particular new law was passed, failed, or deadlock, it can be particularly challenging to achieve improved stalled typically has deep roots in wider nontechnical factors. The PFM legislation. passing of a new law has typically proved to be far more difficult The area where recent international ‘best practice’ models in complex countries with relatively powerful legislatures (Nepal, seem to have had the greatest impact is on legal provisions for public Philippines, and Nigeria), than in Georgia and Tanzania where the procurement. In this area, an international ‘model law’ was devel- incentives or powers of legislatures to block legislation have been oped first in the 1990s, and then updated in 2011, by the United relatively smaller. In the Philippines, the overall constitutional Nations Commission on International Trade Law66 (‘UNCITRAL setup as well as political constellations generally make the passing model law’), as part of the global move toward greater trade liber- of any legislation difficult (Matsuda 2013). Such situations can alization and market integration (see also Sanchez 2013). Many leave PFM systems without sufficiently clear legal underpinnings countries initiated public procurement reforms since the 1990s for extended periods of time. also as part of their accession to the World Trade Organization and regional trade agreements. All five countries covered here adopted 4.1.2 Organization of ‘Central Finance new procurement legislation in the 2000s or 2010s. As is dis- Functions’67 cussed in Chapter 5, these reforms mostly involved external sup- port and advice based on international best practices, and a key What Ministries of Finance are responsible for and how this links to thrust was to make government procurement more open to national the roles and responsibilities of other ministries, departments, and and international competition. Tanzania provides an example of agencies (MDAs), varies substantially from country to country. As first moving toward international ‘best practices’, then going back- Allen et al. (2015) note, this issue has thus far received rather ward, and then pursuing another round of reforms through suc- cessive changes and amendments of its procurement legislation. 67. This section reflects the current organizational arrangements in each country and related reform challenges; while discussing most important changes or new institutions established over the past decade. It would be 66. See: http://www.uncitral.org/uncitral/en/uncitral_texts/procurement_ difficult to fully reconstruct the institutional setup of the mid-2000s for infrastructure/2011Model.html. each country. 38  Political Economy of Public Financial Management Reforms limited attention in the literature.68 However, these arrangements this newly created ministry. The minister now no longer has a for- are likely to have an impact on what Ministers of Finance care mal oversight role over budget preparation, while he/she also no about and may also entail gaps, blockages, or friction if the roles longer has a broader economic management or reform mandate. of various MDAs are not sufficiently well defined or are partially The Office of the Accountant General and the Debt Management overlapping or competing. Office continue to be separate offices from both the ministries. A further area where countries notably diverge is with regard In the Philippines, the central finance institutional landscape is to their institutional setup to coherently pursue planning, budget highly fragmented. There is no Ministry of Finance, but rather four preparation, and budget execution. The lack of good enough links main CFAs. The Department of Budget and Management (DBM) is between planning and annual budget allocations is a widely noted the primary agency tasked with preparing annual budgets, as well issue (Allen 2009; Dabla-Norris et al. 2010; Wilhelm and Krause as administrating and controlling budget execution. It also man- 2008), and ‘getting things right’ can be challenging. In terms ages the government’s compensation and position classification. of reform efforts, the standard tools that are being pursued for The Department of Finance (DoF) is primarily tasked to handle improvements are the introduction of MTEFs and of program bud- revenue policies as well as the administration of the revenue gen- gets. However, in contexts where the institutional setup is dysfunc- eration, debt management, accountability of state-owned corpora- tional or not functioning well enough, injecting these tools without tions and assets, and (indirectly) treasury functions. The Bureau making other changes may not to be effective. of Treasury (BTr) (which falls under the purview of the DoF but Among the five cases, Nigeria and the Philippines have the most remains autonomous) manages cash transactions and maintains complex setup of CFAs, Nepal is an intermediate case, and Tanzania the book of accounts of the national government, as well as for- and Georgia have currently relatively concentrated central finance eign and domestic debt. In addition, the National Economic and functions, coming closer to what one might describe as a ‘standard’ Development Authority, which is a cabinet-level agency respon- model of a MoF. In Nigeria, the setup of CFAs was changed in late sible for economic planning, policy coordination, and coordination 2015, following the election of the new government. Both before functions regarding public investments. Lastly, the Commission on and after the change, the Federal MoF had as its core technical Audit (COA) is not only tasked with external oversight of all govern- departments, one for Economic Research, one for Reform Coordi- ment accounts but it also oversees the accounting function, which nation, and a Department for International Economic Relations. presents a conflict of interest. There are several other agencies Neither budget, nor treasury, nor debt management functions were involved in PFM activities including the Public Procurement Policy directly under the ministry. Board and the Central Bank. The main budget preparation functions were located in the In Nepal, the Ministry of Finance69 has direct responsibility Budget Office of the Federation, budget execution functions in for budget preparation and for asset management while the remain- the Office of the Accountant General, and debt management in ing CFA functions are shared between the MoF and other agencies the Debt Management Office. In addition, there was a National or departments. Treasury and cash management are the functions Planning Commission (NPC) with partial responsibility for the of the Financial Comptroller General Office (FCGO),70 headed by preparation of investment plans. At the same time, at least for a ‘special class’ officer. The FCGO is also responsible for inter- the period 2011–2015, the Minister of Finance was concurrently nal audit. Revenue management falls under the Revenue Admin- the Coordinating Minister for the Economy (CME), with a remit istration Division within the MoF as well as the Inland Revenue focused as much on reviving agriculture and seeking economic Department,71 another one of the six departments under the MoF diversification, as on upgrading public financial management. The (that is of the same status as the FCGO). A Public Procurement main revenue collection responsibility sits with the Federal Inland Monitoring Office (PPMO) was established in 2008 as an agency Revenue Service (FIRS), and similarly, procurement oversight is under the Prime Minister’s Office. established as a separate function. An element of complexity concerns the role of the National In late 2015, the institutional setup was changed by the new Planning Commission (NPC), due to its significantly overlapping government, but only partially defragmented. The main change mandates with regard to the annual budget. The NPC is tasked was the creation of a Federal Ministry of Budget and National Plan- with formulating and monitoring development plans and policies ning (FMBNP), combining the functions of the Budget Office of at a macro level. However, its role as against the MoF and other the Federation and of the NPC, and becoming responsible for the CFAs has grown over time. It not only approves projects at the preparation of the entire federal budget, including recurrent and capital expenditures. A Federal MoF continues to exist alongside 69. See: http://mof.gov.np/uploads/cmsfiles/file/Structure%20of%20 MOF_Eng_20130730125648.pdf. 70. http://www.fcgo.gov.np/about-fcgo/ This office is one of six 68. Existing discussion and analysis include Allen and Grigoli (2012); semiautonomous departments under the MoF. Allen and Krause (2013); and World Bank (2013). 71. https://www.ird.gov.np/home/index. The World BAnk  39 central and district level during budget preparation, but also does which is the ministry’s IT department in charge of managing and so again during budget execution, after formal approval of the further developing the ministry’s IT systems. annual budget by the parliament. This contributes to delays in As the overview of these different arrangements indicates, the executing capital expenditures, and low overall budget execution complexity of the institutional setup can vary significantly, and affects rates. The NPC has also become highly politicized. Debt and aid the potential for pursuing PFM reforms. Institutional fragmentation management functions are distributed between the MoF, FCGO can go far beyond the issue of integrating or not integrating plan- (ensuring timely payments), and the Central Bank. ning functions into the MoFs and can instead also extend to areas In Tanzania, a number of key central finance functions are typically thought of as core functions, such as budget preparation. included under the MoF: budgeting, treasury, accounting, internal The experience across the case studies indicates that complexity audit, and public procurement appeals, while planning has moved tends to become an obstacle to pursuing PFM reforms, as has in and out of the ministry. The revenue function is organized been in particular the case in Nigeria and the Philippines. separately in the Tanzania Revenue Authority (TRA), established It is possible, but challenging to change the arrangement of in the mid-1990s as an autonomous revenue agency. The Pub- CFA functions. Georgia is a case where, as part of the reform efforts lic Procurement Regulatory Agency (PPRA), established through made since 2004, the institutional design and allocation of CFA the 2011 Public Procurement Act (PPA), has general policy over- functions was brought more explicitly under the control of a single sight of procurement. The planning functions for investments have ministry. A similar trend seems to materialize in Tanzania, fol- moved in and out of the MoF. In 2008, a Planning Commission lowing the 2015 elections. The potential cost and benefits of an was again established in the President’s Office—the President’s institutional change should be carefully considered. In particu- office-Planning Commission (PO-PC),72 with a broader role to for- lar, in highly fragmented environments, the potential short and mulate and monitor implementation of national plans, as well as longer-term benefits of greater integration are likely to be substan- to plan and oversee major public investments. The latter respon- tial. One option can be to seek consolidation when governments sibility entailed some tensions between the MoF and the PO-PC. change, because structures and office holders are less entrenched Following the October 2015 elections and renewed reform drive, at those points in time. This has been done in the cases of Nigeria the planning function was moved back to the MoF. The external and Tanzania in 2015–16, while in the case of Nigeria, further audit function in Tanzania is exercised by the Controller and Audi- steps appear as needed. tor General (CAG). Georgia has developed the most concentrated allocation of cen- 4.1.3 Relationship between the Executive tral finance functions among the group.73 The MoF holds responsi- and the Legislative bilities for budgeting (capital and recurrent), treasury, tax policy and administration functions, and public debt. This concentration The relationship between the (central) executive and legislative is of functions has been pursued as part of the reforms made over the also a crucial factor affecting PFM systems. In the recent donor past 10–15 years, including the full integration of the treasury and discourse around PFM, legislatures are often primarily consid- revenue functions into the ministry. Public procurement oversight ered as a body exercising oversight and accountability. Therefore, and execution functions are exercised through a State Procure- the emphasis of development support in this regard is typically ment Agency which has attained its current status through several focused on strengthening the role of parliaments to review budget rounds of reforms. External audit is the responsibility of the State execution ex post (see for example, a summary of such efforts Audit Office (SAO). provided by Hudson and Wren 2007). However, parliaments also One particular feature of the institutional setup in Georgia is have a crucial role to play in other PFM aspects: (a) approvals of the existence of ‘Legal Entities of Public Law’ (LEPLs). These are annual budgets (and MTEFs where these exist) and (b) passing of divisions within ministries or separate agencies which fall outside PFM reform legislation. the purview of the regular civil service. Generally, LEPLs have a In three out of the five case study countries, contestation right to generate and manage their own revenue and they can pay between the executive and the legislative has repeatedly led to sig- salaries outside of and above civil service pay scales. For example, nificant delays in budget approvals. As Wehner (2006), Lienert alongside its regular divisions, the MoF of Georgia contains four (2013b), and others have emphasized, the budgeting powers that LEPLs: (a) the Revenue Service, (b) the Service Agency, (c) an are allocated to legislatures, typically through a constitution, vary Academy of Finance, and (d) the Financial Analytical Service, significantly across countries. Further variations—within a country over time—result from the fact that the level of support in a legis- lature for a government can vary and change due to several factors, such as whether and how broad the executive has a supportive/ 72. http://www.mipango.go.tz/. aligned majority in the legislature. 73. http://www.mof.ge/en/4479. 40  Political Economy of Public Financial Management Reforms Delayed budget approvals are an important source of distor- changing de facto allocations during budget execution, as well as tions and malfunctioning of PFM systems. Delayed approvals create a bunching of expenditures toward the end of the fiscal year. There uncertainty for spending agencies, especially around nonsalary can also be a knock-on effect for ensuring proper accounting and expenditures; and they increase discretion within the executive timely completion of the annual budget cycle, including finalizing with regard to releases during the initial months. Nigeria is a par- financial statements and external audit reports. Furthermore, long ticularly extreme case, where disagreements between the execu- delays and multiple revisions of the budget negatively affect budget tive and the legislative have repeatedly led to delays between the transparency, because they lead to delays in the publication of offi- budget being submitted to parliament and final signature into cial budgets, and uncertainty about what constitutes the ‘original law of more than 150 days in recent years.74 Budget approval budget’ against which actual expenditures should be compared. delays have also been substantial in Nepal, due to the political As reflected in Figure 4.1, the timeliness of budget approvals uncertainties and challenges related to the postconflict political can improve over time. This has notably happened in the Philip- transition. Budgets have been submitted in mid-July to the Consti- pines and in Tanzania. In the Philippines, the Arroyo administra- tutional Assembly, but have been passed after one- to nine-months tion had weak support in the Congress, while this strengthened delays (the latter in 2012/13), leaving the government to operate during the Aquino administration. The Philippines as well as Nige- with short-term authorizations to continue spending.75 Timeliness ria have constitutions that seek to emphasize checks and balances improved following the passing of the 2015 Constitution. In the between the executive and the legislature with regard to control Philippines, delays were substantial during the Arroyo administra- over the budget, and such arrangements can be prone to deadlock. tion which had insufficient support in the Congress, but improved In addition to delayed budget approvals, legislatures can also significantly during the 2010–2016 Aquino administration, as affect the realism of budgets. For example, in some countries, MPs also reflected in the substantial improvement in PEFA ratings for request that estimated revenues are revised upward to accom- ‘timeliness of budget approvals’ (Table 4.2). However, this has modate additional spending requests. This can result in revised been due to political constellations rather than changes in relevant estimates of expected economic growth, and/or of expected rev- rules and may slide backwards whenever there is a period of diver- enue (for example, through upward revision of oil price estimates gence between the President and the majorities in the two houses for the upcoming fiscal year as happened in Nigeria). The degree of Congress. to which unrealistic budgets are ‘promoted’ by legislatures rather Delays in budget approvals almost inevitably reinforce the ten- than executives is more challenging to document than delays in dency of under-execution of investment expenditures, changes in allo- budget approvals, because it requires access to the drafts being cations during budget execution, and bunched releases and payments submitted to legislatures, as well as to the budgets as passed. at the end of the fiscal year. When budgets are adopted with sub- Given the potential importance of this issue, making the executive stantial delays, MDAs have no certainty over their allocations for budget proposals publicly available should be part of the global extended periods of time and hence will typically hold back in par- pursuit of greater fiscal transparency. ticular with moving investment projects forward. This contributes to Improving executive-legislative relations around budgets can be very challenging. Typically, the power of the legislature with 74. These include delays in budget approval by the legislative with regard to budget approvals is set out in the constitution and hence delays in executive assent to the budgets as passed by the legislature. requires particularly large-scale majorities to change. Such majori- Budget approval delays affect the federal budget, but not state budgets. ties are particularly unlikely to be a feasible prospect in those In the states, legislatures mostly ‘rubber-stamp’ budgets proposed by executives. However, in some states, executives include expenditures countries where long delays with budget approvals occur. As one favoring the political goals of various members of state legislatures, commentator on the Nigerian experience pointed out, everyone in even though these cannot all be funded, in the expectation of adjusting budgets as needed during execution. the MoF and the Federal Budget Office was so exhausted from 75. Nepal’s fiscal year starts in mid-July. the annual struggle to get the budget passed by the Legislative Table 4.2.  PEFA Ratings for Timeliness of Budget Approvals (PI-11 (iii)/[PI-18.3])a (iii) Timely budget approval by the legislature or similarly mandated body (within the last three years) Georgia Nepal Nigeria Tanzania Philippines Earliest A D D B (2009)b D Most recent A NA — C A Note: a. The main PEFA indicator cited is that as used in the 2011 PEFA framework. The indicator cited in [] refers to the 2016 version of the PEFA framework. b. In the 2006 assessment for Tanzania this dimension was not rated. The World BAnk  41 Figure 4.1.  Timeliness of Budget Approvals Compared Country Comparison on Delay of Budget Approval 250 Number of days of delay of budget laws 200 150 after start of FY 100 50 0 –50 Tanzania Nigeria Nepal Philippines 2010 2011 2012 2013 2014 2015 2016 Source: Data from publicly available information and governments documents. Data not available for Georgia, and for 2014–2016 for Nepal. Assembly that there was little appetite to even consider seeking is important to consider that despite the new procurement laws a constitutional change that would have clarified the respective having been passed, the change process continues rather than roles of the executive and the legislature with regard to the budget having reached a ‘steady state’ and full implementation in several and to reduce the likelihood of prolonged deadlock. of the countries. For example, in Nigeria, implementation of the A further key aspect that tends to be underappreciated is the procurement legislation is considered weak both at federal and role of parliaments in approving PFM reform legislation. Across the state levels.77 five cases reviewed here, three (Nepal, Nigeria, and the Philip- Moreover, despite considerable attention by international pines) have PFM reform legislation that is pending in parliament; development partners (DPs), as noted above, the oversight and and some have several pending bills (for example, a pending accountability role of parliaments has remained stunted. This is illus- organic budget law as well as pending legislation on the external trated by the fact that the PEFA ratings for legislative scrutiny audit function in Nigeria, and several laws in Nepal and the Philip- of external audit reports did not improve in even one of the five pines). For various reasons, legislatures can hold up such reform countries analyzed (Table 4.3). The lack of improvement in parlia- legislation for extended periods of time. mentary scrutiny stands in contrast to the performance of external The passing of procurement legislation in all five countries audit, which while still weak, did see improvements across the indicates that in principle, legislation targeting PFM reforms can be case studies (further explored in Chapter 5). This is also in line passed even in difficult environments and even if a law potentially with the cross-country analysis, which found that external over- goes ‘against the grain’ of vested interests. The new procurement sight and accountability is the weakest dimension across the six laws typically took about five years from initial design to even- PEFA categories, and this holds true across income levels as well tual passage.76 However, as is discussed further in Chapter 5, it as different types of political regimes (that is, democratic as well as non-democratic). Legislatures do not take up, discuss, and follow up on audit 76. In Nigeria, the first Country Procurement Assessment Report (CPAR) was produced in 2000, and the Public Procurement legislation reports for various reasons. For example, in Nepal, the parliament was passed in 2007. In Tanzania, the first CPAR was carried out in was either not in session or did not table audit reports for dis- 1996 and the Public Procurement Act (PPA) was passed in 2002, and cussion submitted to it before 2014. This is partly due to the then subsequently changed multiple times. In Nepal, about 5 years elapsed between the first CPAR and the passing of the PPA 2007. In postconflict environment of having no parliament and, hence, no the Philippines, public procurement reform began in 1998 and was approved in 2003. In Georgia, the first Law on Public Procurement after independence was passed in 1999. However, reforming procurement 77. A majority of states has followed the federal level in passing new began in earnest in 2010. procurement laws. 42  Political Economy of Public Financial Management Reforms Table 4.3.  PEFA Ratings for Legislative Scrutiny of External Audit Reports PI-28 [PI-31]: Legislative scrutiny of external audit reports Georgia Nepal Nigeria Tanzania Philippines Earliest C+ D+ D C D Most recent D+ D - D+ D PI-26 [PI-30]: Scope, nature and follow-up of external audit Earliest D+ D+ D+ D+ B+ Most recent B+ C+ - C+ Ba Note: PI numbers according to the 2016 PEFA methodology are provided in brackets. a. In the 2016 PEFA (based on the new methodology) this indicator is rated as C+; but when comparing between 2010 and 2016, the indicator is assessed as B. parliamentary committee (or Public Accounts Committee [PAC]), oversight roles rest with state-level audit offices and parliaments. but also due to the absence of an Auditor General for seven years. Therefore, efforts to support improvements in PFM have sought A somewhat different situation prevails in Georgia, where from to target particular states, for example, those with governors who 2012, the SAO was perceived as partisan due to appointments have signaled greater interest in using public funds well and in made by the previous government just before the end of its man- achieving better service delivery. Even in this setting of limited date. More broadly, there are also incentive issues: paying atten- direct influence, the federal government still has an important sig- tion to budget allocations ex ante offers more potential political naling role, because its laws, standards, and practices are often rewards than trying to discern ex post where money was not well replicated at the state level. spent, and getting the executive to make changes. Nepal stands out for the most far-reaching transformation being under way to its system of intergovernmental relations. The coun- 4.1.4 The Importance of Intergovernmental try changed from having a unitary system to a federal system of government in 2015, following intense and drawn-out discussions Relations and of Effective PFM Systems on constitutional reforms. The change is expected to also funda- across Levels of Government mentally affect the way in which budget plans are developed and An additional aspect of institutional complexity is the arrangement allocations are executed. When such large-scale changes happen, of intergovernmental relations, and given widespread trends toward they, in principle, require significant investments in training and increasing decentralization, PFM reform efforts will increasingly have retraining staff, setting up new offices, and so on—and funding to contend with these. Intergovernmental relations are particularly from a country’s own resources or from DPs may often fall short or important insofar as the overall goal of strengthening PFM systems not be on time. is to strengthen service delivery. In decentralized countries, even The Philippines, Tanzania, and Georgia are all unitary states, more so in federal systems, this means that PFM systems have to but each still holds some complexity in intergovernmental relations be effective and accountable across levels of government rather and the configuration of PFM systems. Discussions about how to than just at the center. (re)configure intergovernmental relations are present in all three Among the five cases covered, intergovernmental arrangements cases. With a population of under 4 million, Georgia is a compara- differ, and in several cases, they have either been changed in recent tively small country. Municipalities were traditionally simply exe- years or discussions on this are ongoing. Nigeria is the only one cuting levels of government, but have gradually received greater of the five cases that is an established federal state. This places degrees of autonomy since the early 2000s. Still, the central gov- important limitations on the degree of improvements that can be ernment has a full view of all financial transactions, through the achieved through PFM strengthening at the level of the federal unified treasury system, for example, and there is a single external government. State and local governments have a guaranteed right audit body. to substantial fiscal transfers from common resources (the Fed- Tanzania has central, regional, and local levels of government eration Account and the separate pool of value-added tax (VAT) with very limited subnational autonomy with regard to allocating and receipts, as well as a ‘derivation allocation’ for states where oil managing public funds. The budgetary votes of Level 1 (central is produced). Once funds are transferred, the federal government government ministries) and Level 2 (regions) are approved by the does not have direct oversight of or influence on how funds are national parliament. Level 3 (local government) is under the under spent to deliver public goods and services. The external audit and the authority of the Regional Administration and Local Government The World BAnk  43 under the President’s Office,78 which thus is a direct powerful and ensure that sufficiently well-trained staff is locally available, lever over spending at local levels. Heads of regions are appointed in particular in low-income contexts. Capacities for planning and rather than elected. While local government authorities (LGAs) supervising investment projects can pose a particular problem. are responsible for delivering basic health and education services, Increases in fiscal autonomy can also lead to gaps in oversight central ministries retain direct involvement. These arrangements that take time to address through strengthening local oversight are very much in contrast to those prevailing in Nigeria. mechanisms and capabilities. While the Philippines has a particularly dispersed geographic Third, many centrally implemented PFM reforms require buy- structure (with 7,000 islands) and a large population (over 90 million in from subnational stakeholders, and rollout or replication across as of 2010), it still retains a relatively centralized system of govern- subnational entities. For example, as discussed in Chapter 5, ment. The Local Government Units (LGUs) comprise 81 provinces, implementing a TSA raises questions about whether and how sub- 144 cities, 1,490 municipalities, and over 42,000 barangays (vil- national entities are part of the system, or should be allowed to lages). Somewhat similar to the situation in Tanzania, local gov- maintain separate bank accounts. Similar issues prevail for the ernments are in charge of basic service delivery, but the national implementation of automatization through a financial manage- government continues to play a significant role in both financing ment system and various potential modules of such a system. and implementing these. National government agencies are repre- sented in the field by a mix of regional, provincial, and/or district offices, and the national government implements various vertical 4.2 The Political-Technical Interface investment programs, for example, for local roads and health clin- for Driving Reforms ics. Regarding ex post oversight, the COA has regional offices, As sketched out in Chapter 3, politicians typically have overarching and has started building some Provincial Satellite Audit Offices goals and incentives that make them more or less interested in pursu- (PSAOs). ing improvements in how public finances are managed. These goals As this discussion indicates, intergovernmental relations are and incentives can be related to fiscal trends—for example, caring an important aspect of the PFM-related institutional context, and in more about the management of funds when a lack of resources particular for the intended link between improving PFM systems and threatens the ability to act and less so when there is a steady better service delivery. With variations depending on the particular increase—as well as to other wider issues, including external and intergovernmental arrangements, service delivery improvements internal threats and aspirations. can only be achieved if the flow and the use of public funds is Taking these motivations as given, the issue of how they are improved across the levels of government involved (see also Smoke translated into concrete action at the technical level arises; and also, 2015a, 2015b). Particularly, basic service delivery is typically the if and how anything can be achieved from within (senior) techni- responsibility of third or fourth levels of government, and hence cal levels during periods when wider political commitment to pur- there are multiple institutional layers between a national MoF sue reforms is weak or absent. The former aspect is particularly and a local school, health clinic, or agricultural extension service. interesting for incoming politicians who may want to prepare or This can make the pursuit of reforms particularly challenging in a make changes. The latter aspect is relevant for stakeholders at context such as Nigeria, which includes a high degree of subna- the technical level, and also for DPs who often seek to support tional autonomy and requires efforts by multiple separate sets of PFM reforms in environments where political commitment is more stakeholders.79 lukewarm or volatile. A second key point is that changes to intergovernmental The case studies indicate several interesting facets around the arrangements and responsibilities for managing public funds are political-technical interface. First, politicians as well as other stake- being pursued in a number of countries and can lead to new capacity holders pay attention to the appointment of Ministers of Finance. and oversight bottlenecks. New PFM reform challenges can emerge The appointments are typically made deliberately, and not as a when changes such as an increase in the number of subnational residual appointment such as sometimes made to more minor units or in the share of fiscal responsibilities are being decided. A portfolios. In some instances, appointments are made with a sig- basic challenge can be to establish new offices, roll out IT systems, naling intention to the international development and/or business community (for example, the two appointments of Ngozi Iweala in Nigeria). Furthermore, among the five case studies, several Minis- 78. This function has moved multiple times between the Prime Minister’s and the President’s Office. ters of Finance have been relatively high-level politicians who sub- 79. Intergovernmental relations also often entail their own political sequently moved on to the office of Prime Minister (for example, economy and fiscal dynamics, such as higher spending on some areas than on others, or less active oversight by the national level of Zurab Nogaideli and Nika Gilauri in Georgia, Baburam Bhattarai of subnational units run by important allies. These issues are beyond the scope of this analysis. 44  Political Economy of Public Financial Management Reforms the Maoist party in Nepal), while in other instances they have been implement the program, which no doubt compliments the anticor- primarily technical experts. ruption crusade.”81 Political leaders may also care about limiting The selection and mandate for reforms can be quite corruption and mismanagement if this has been a central aspect of explicit. The new 2015 governments in Tanzania and in Nigeria their political platform—and may do so to a quite detailed extent both explicitly appointed reform-oriented Ministers of Finance as as has been the case during the 2010–2016 administration in the a way to signal a renewed focus on governance and PFM reforms Philippines and the administration in Tanzania following the 2015 (Kemi Adeosun in Nigeria and Phillip Mpango in Tanzania).80 In elections. 2004, the post-Rose Revolution government in Georgia appointed Therefore, even if day-to-day involvement of political leaders is as the initial Minister of Finance Zurab Nogaideli, who was one of usually minimal, it matters whether other ministries and stakeholders the ‘internal’ reformers who had served under and then left the believe that the political leadership is backing PFM reforms being previous Shevardnadze Government in protest against its inability pursued by technical leaders. A number of PFM reforms require to address the subversion of the state by corruption and criminal changes that have to be implemented broadly across government networks. When Nogaideli was moved to become Prime Minister (for example, developing budgets in new formats, ensuring timely in 2005 (after the unexpected death of the incumbent), a succes- and adequate reporting of expenditures, and so on). As discussed sion of young, reform-minded successors was appointed, including above, legal changes as well as timely budget adoption will also a second Minister of Finance who then moved on to the post of depend on a majority in parliament approving these without exces- Prime Minister (Nika Gilauri). sive delays. Political backing of the reform agenda—as well as Second, beyond appointments, the direct involvement of polit- institutional factors such as the allocation of powers over budget ical leaders in PFM is typically limited, but at key decision points approvals—is therefore important, especially for potentially more explicit political backing can be critical. While Presidents or Prime contentious reforms and contentious points in reform processes. Ministers select Ministers of Finance intentionally—to reform or This is explored further in Chapter 5. rather to preserve the status quo—in many situations they will not Across the five case study countries, most Ministers of Finance engage on the detail of PFM reforms. In the words of a former min- have been in their positions for relatively short periods, hence com- ister (who had been involved with considerable reform progress) mitment and continuity provided by senior civil servants is poten- “the President did not understand or care about PFM reforms, tially important. Rather short ministerial tenures prevailed both except for tax collection; and his direct involvement was minimal.” Georgia—which saw the greatest degree of progress—as well in ­ This is likely to be true for many situations; government leaders as in Nigeria, which saw the least degree. Each of two countries may be keen on regaining or adding fiscal space as this opens their had seven Ministers of Finance within the decade up to 2015. margins for maneuver and for pursuing their key political agenda. The number rises to nine Ministers of Finance in Nepal during It is very unlikely that they will care about the PFM reform details the period 2003 to 2013. In Tanzania, the ministerial portfolio such as accounting standards or improvements in IT systems. changed hands five times between 2000 and 2015, and two to However, at key turning points and when resistance from three office holders for key CFA positions in the Philippines.82 stakeholders, in particular within the executive, is significant, clear Thus, tenure length as such is not clearly related to the likelihood political signals are crucial. For example, a 2016 senate hearing of reform progress.83 One feature noticeable in Georgia and also in Nigeria credited the President directly with forcing MDAs to to some extent in Nepal is that senior civil servants rather than close their accounts in commercial banks and to complete the ministers were also significantly involved in PFM reform efforts, TSA for the federal government: “However, full implementation of and at least in Georgia, these tended to serve in the ministry for a TSA (recollection) kicked off in March 2015 [that is, still under longer period. outgoing President Jonathan] and gained traction when President The level of prior experience of the Minister of Finance and Muhammadu Buhari mandated the closure of all FGN accounts senior appointments in the ministry, including specific experience in held in commercial banks by September 2015. This led to a a country’s PFM system can also be relevant. The background of massive one time surge, especially for September and October 2015, as MDAs rushed to comply. [.  .  .], the senate resolution setting up the Joint Committee, the Senate President and all the 81. Available at: http://www.elombah.com/index.php/special- reports/5624-the-full-senate-report-on-the-abuse-of-the-treasury-single- testifiers at the hearings applauded the sagacity, foresight, and account; accessed September 29, 2016. dogged determination of President Muhammadu Buhari to fully 82. The Philippines had two Secretaries of Finance between 2005 and 2016, three Secretaries of Budget between 2006 and 2015, and three Commissioners of Audit between 2001 and 2015. 80. As noted in Section 4.1.2., the role of the MoF in Nigeria is limited. 83. Some Ministers of Finance who are credited with substantial reforms The appointee as Minister of Budget and Planning, that is, the second have served quite long terms, for example, Trevor Manuel of South Africa and main CFA, is Udoma Udo Udoma, a former two-term senator from a (1996 to 2009), but such long tenure is quite rare, in particular in more southern state, Akwa Ibom. democratic settings. The World BAnk  45 Ministers of Finance is typically one of the following: from within 4.3 Demand Side—More a Potential the ministry (for example, former Directors or Directors General), being transferred from a prior post of minister of another ministry, than an Actual Driver for PFM Reform from the private sector—often a bank or other financial institution, The role of the demand side in driving PFM reforms across the five from being an economic and fiscal expert within a political party countries remained rather limited. This has multiple reasons: the constituting government, or coming from an international organi- technical issues are diverse and can be challenging to fully be zation or academia. conversant with for nonspecialists. Furthermore, while there is In principle, having previously served in the MoF bestows some typically some public demand to address corruption where this advantages in terms of being familiar from the start with key bottle- constitutes a major challenge, the specifics of PFM reforms are necks for progress and potential solutions. Such choices were made typically not widely discussed in the public; and may even be little following the ‘pro-reform’ elections of 2015 in Tanzania and Nige- known outside a relatively small circle of those involved in a par- ria: the newly appointed Minister of Finance in Tanzania previously ticular aspect.84 That said, there have been some exceptions, such served as a Deputy Permanent Secretary; while in Nigeria, the new as considerable attention that emerged on the implementation of Minister of Finance had previous experience with reforming PFM a TSA in Nigeria in early 2016, or on the issue of procurement systems in one of the country’s states (Ogun), while the new Min- reforms in Georgia, and on budget transparency in the Philippines. ister of Budget and Planning had served as Chairman of the Com- As noted in Chapter 3, all five of the countries are ‘partly free’ mittee on National Planning, Revenue Mobilization, and Poverty or democratic according to international indices, and hence all offer Alleviation, and Chairman of the Appropriations Committee during some space for citizen engagement and public debates. At various his tenure as a senator. points in time, discussions about and discontent with corruption Having a competent team, which has an articulated reform has also been running high across these cases. Thus, each of these agenda (Kingdon 1995) is helpful, and competency is likely to be a countries in principle offered some opportunities and reasons for necessary condition for success, but as the example of the Philippines greater citizen engagement on PFM reforms. underlines, it is not sufficient. As discussed further in Chapter 5, Selectively, nongovernmental organizations (NGOs) and public the heads of offices of key CFAs were quite competent and expe- media have been active on issues such as budget monitoring, pro- rienced and had at least a partially clear reform agenda. However, curement, and overall fiscal transparency. In Nigeria, NGOs have these positive factors were constrained in their effectiveness by engaged in some monitoring of budget allocations and to the the high degree of institutional fragmentation, by staking out an extent possible also actual expenditures both in individual states overambitious agenda, and by a political mandate not quite strong and at the federal level. At the state level, governors do not neces- enough to clearly go against strong vested interests. sarily welcome more active scrutiny and have at times threatened Some limited, incremental, and partial progress may still be civil society representatives or made it difficult for them to obtain possible even in a difficult environment if there are technical teams data on planned or actual expenditures in a timely way.85 Overall, seeking to pursue improvements; but this cannot have a transforma- according to Okonjo-Iweala (2013): “[i]n the case of Nigeria, it is tional impact on its own. Most notably, some reform aspects such as only recently that civil society organizations (CSOs) are engaging the partial and gradual rollout of a TSA still progressed in Nepal, more systematically in the budget process.” The author (and for- despite political fragmentation and repeated deadlock and very mer Minister of Finance) puts this more nascent engagement on frequent changes of Ministers of Finance. PFM issues also in contrast to notably greater civil society engage- While political backing is important for enabling PFM reforms ment with regard to the Extractive Industries Transparency Initia- to be carried through, conversely political disinterest (or interests tive and the repatriation of funds looted by the Abacha regime. opposed to improving the efficient and accountable use of funds) can Annual budgets, as well as selected specific topics, such as the negatively affect technical capacity. For example, one of the govern- contested establishment of the TSA, receive some media cover- ments in the sample replaced technically and managerially capa- age; including coverage of problems with budgets that relevant ble staff by weaker appointments over a period of about a decade, CSOs have identified.86 thus weakening the role of the MoF. The resulting declining trend In Nepal, during the postconflict period, considerable emphasis is reflected both in stagnating and declining PEFA ratings and was laid and external support provided for more participatory budget- ratings by general indices such as the Worldwide Governance Indi- ing. However, as a 2011 report finds: “Overall, public demand for cators (WGI) (for Government Effectiveness and Control of Corrup- tion). However, even in this case, training opportunities for MoF 84. See also the recent work by de Renzio and Mastruzzi (2016) focused staff have continued to expand, thus maintaining opportunities for on how CSOs use budget information. a turnaround of technical capacities as the political commitment 85. Focus group discussions in Delta and Bayelsa states, May 2013. 86. See for example, http://www.newsweek.com/nigerias-buhari-finally- re-emerges. passes-2016-budget-456739 on the (late) adoption of the 2016 budget. 46  Political Economy of Public Financial Management Reforms information on public finances is weak, linked to a lack of ‘budget four case study countries. NGOs are credited with having been literacy’ on the part of citizens and limited expertise on finan- closely involved in improving procurement legislation adopted cial and budgetary analysis among CSOs.” (Tamang and Malena in the early 2000s, as well as with increasing budget transpar- 2011). There was also some concern that investing in ‘participa- ency and participation in budgeting decisions. There are, however, tory budgeting’ at local levels can lead to expectations that a dis- at least two important limitations: as is discussed in Chapter 2 trict would receive more funds—rather than (or more than) putting and in further detail in Chapter 5, some of the envisaged PFM a clear emphasis on spending available funds more effectively and reforms were not successful; notably the implementation of the with less leakage.87 Government Integrated Financial Management Information Sys- In Tanzania, CSOs have had longer engagement on accountabil- tem (GIFMIS). Second, the formats in which budget information ity and developed some technical capacity, but at least up to 2015, is presented are not very clear (receiving a ‘C’ rating in the 2016 government resisted greater CSO involvement in PFM. As the coun- PEFA; see also IMF 2015), due to a complex structure, earmark- try report notes: “[CSOs’] direct involvement in budget and PFM ing, special purpose funds, and automatically authorized expen- reform has been limited and had minimal impact due to capacity ditures, as well as considerable powers of the executive to make constraints, resistance from government, and few openings for par- reallocations during budget execution. Thus, while information is ticipation.” CSOs have been involved in Public Expenditure Track- made available, including in Excel formats, it can still be challeng- ing Survey (PETS) and Social Accountability Monitoring (SAM); ing to analyze and interpret. and their engagement has focused on expenditures in particular In Georgia, civil society has played an intermediate role with sectors, such as education (HakiElimu88). Thus, CSOs have been regard to PFM reforms. Somewhat similar to the Philippines, several able to engage more on specific aspects of sector expenditures individuals have moved between prominent CSOs and positions in than on core PFM reforms. the government, albeit not as specifically related to PFM. Several The Philippine civil society community is one of the world’s CSOs are able to employ specialist staff, which allows them to more vibrant and sophisticated. The interests and effectiveness of monitor government operations, including PFM, overall fiscal man- these CSOs naturally vary quite significantly, particularly when it agement, and procurement. The organizations most significantly comes to advocating for better services or PFM. Several CSOs have engaged on governance issues include Transparency International, been actively involved in various ways to improve the PFM system the Georgian Young Lawyers Association, the Open Society Geor- over the years. Several CSOs were newly formed under President gia Foundation, and the Institute for Development of Freedom of Estrada’s tenure (1998–2001), which created the Transparency Information. CSOs have been invited to serve as members of a and Accountability Network (TAN). During the Aquino presidency, range of government committees and commissions. They are rep- CSOs have been actively involved in a participatory budgeting resented in the Dispute Resolution Council of the Procurement process. Agency, and on the PFM Coordination Council that was established Over the course of several decades, NGOs have built up in 2009. At the same time, many former public officials moved stronger technical capacity, benefiting from external financial and back to CSOs (as well as to the for-profit sector) after the 2012 technical support (by U.S. Agency for International Development parliamentary elections. [USAID], the World Bank, and others), and also from the fact that However, there are also continued weaknesses in civil society some former civil servants involved in PFM joined key CSOs (Dres- capacity to engage on PFM and related areas. CSO representatives sel 2012). Following the 2010 elections, several of the key CFA interviewed in spring 201590 were of the view that their role is still office holders appointed by the Aquino administration had close limited—both in terms of their capacity to engage and of the gov- links with civil society.89 The Aquino administration then started ernment providing opportunities to do so. Similarly, several media a process called ‘Bottom-up-Budgeting’ which engages LGUs and representatives interviewed indicated that there is limited public CSOs in identifying programs, activities, and projects for inclusion interest, and hence public finances and how they are managed in the National Budget. This was part of the administration’s com- are only infrequently covered in the news. Three areas that get mitment to the ‘People Power’ paradigm for ensuring that citizens some coverage are the annual budget, major public procurement participate meaningfully in public policy and programs (Dressel contracts, and bonus payments made to public officials. A particu- 2012). lar challenge is the fact that civil society itself can be perceived Overall, civil society influence on budgeting and on PFM as politicized, and hence its watchdog role can be rejected by reforms has been greater in the Philippines than in any of the other governments as politically motivated rather than representing the 87. Interviews, Kathmandu, May 2013. 88. See www.hakielimu.org. 89. Finance Secretary Cesar Purisima, as well as Secretary of Budget and 90. Interviews with various representatives of relevant CSOs, Tbilisi, April Management, ‘Butch’ Abad. 2015. The World BAnk  47 public interest at large. There is generally very limited discussion relatively less ideological in Nigeria and Tanzania. In the latter, the on PFM-related issues in the media. dominant party originally had a strong left-wing orientation, but Across the cases, CSOs, media attention, and citizen demand this had shifted toward a market-oriented position over the course thus have played some role in motivating certain aspects of PFM of the 1990s. reforms, but not a very strong one; while a potentially large role is In Georgia, libertarian ideas were especially carried out played by the ‘long route’ of accountability. As captured in Chap- by Kakha Bendukidze, one of the core members of the government ter 3, an important channel between discontent of citizens about and of the President’s inner circle between 2004 and 2009 when weak management of public funds and actual PFM reforms to he left the government.92 While some of his liberal reform efforts address this tends to run through the election of a reform-oriented were focused on business regulations (as Economics Minister) he government. This happened in Georgia in 2004, in the Philippines was also involved in the drafting of the 2011 ‘Liberty Act’ which in 2010, and in Nigeria and Tanzania in 2015. This long route can imposes strict limits on the fiscal expansion of the state at 30 per- be important in bringing about changes; but voter sentiment can cent of GDP, as well as limits on gross debt and on tax increases. of course also swing back. The quality of CSO engagement and of Liberal or libertarian ideas also expedited a loosening up of civil media coverage go hand in hand, in the sense that CSOs are one of service rules, which in turn facilitated large-scale changes of staff the key stakeholders that can provide commentary or criticism on in the areas of tax and customs administrations. It also enabled budgets and management of public funds, which the media then the establishment of LEPLs, as special entities within the public report on. On its own, media coverage tends to be very basic, such administration that can set their own wage scales, generate their as simply reporting the breakdown of the annual budget follow- own revenues, and so on. While an explicit libertarian orientation ing approval. This is also an area that would benefit from further has since disappeared, some of the institutional features this gave analysis and exploration of how the role of the demand side can rise to still persist. develop. In Nepal, political parties with a decidedly left-wing orientation—ranging from Communist to Maoist—won a dominant ­ share of votes in the 2008 elections, albeit as noted in Chapter 3, 4.4 The Role of Ideas and Ideology with a great deal of fragmentation. The basis for the popular support Ideas and ideology can play a potential role in shaping fiscal policies of such left-wing parties was widespread popular discontent with and PFM. One interesting example is the idea of a ‘flat tax’ which a traditional system of social exclusion (including a rigid caste has been long advocated by more right-wing or libertarian groups system), as well as large regional disparities. One might expect a in the United States—and which eventually found fairly wide government led by Maoists, emerging from conflict, and elected adoption in several post-communist countries since the 2000s.91 in reaction to long-standing exclusion of many citizens and dis- While the experimentation was led by the Baltic countries that tricts, to engage in a spending spree. However, rather surprisingly, have also been frontrunners in other state modernization projects despite the strong left-wing orientation as well as despite political such as e-government, the subsequent spread of the policy was fragmentation, spending has remained remarkably aligned with significantly related to Russia’s success in reducing tax evasion revenue in Nepal, as traced in Section 3.3. through the adoption of relatively low and easy-to-administer In the Philippines, a key dichotomy has been between govern- rates for personal and business profit taxes. Compared with fiscal ing in the interest of a few and in the interest of many; the pro-people policies, for PFM, the standardization of reform approaches (as and anticorruption agenda influenced the PFM reform agenda after discussed further in Chapter 5), has limited the role of ideas or 2010. Over the decades since independence, the Philippines has ideological orientation of governments, while some impacts are seen two popular uprisings against governments seen to rule in still noticeable. the interest of a few and as abusing public resources.93 In reac- Several governments across the five countries have had a nota- tion to a repeated prior history of successive corruption scandals, ble ideology or idea-driven policy approach. The ideological orienta- the 2010 administration was committed to improving government tion has ranged from Maoist (as the dominant party in Nepal after integrity, as well as to be more inclusive and transparent to citizens the 2008 elections), to a center-left government in the Philippines and CSOs (as discussed above). This commitment also shaped the from 2010 to 2016, to libertarian (as an influential streak in the Georgian Government from 2004–2012). Governments have been 92. Bendukidze was a biologist by background, had become an oligarch in Russia and a member of the liberal opposition there before returning to Georgia. 91. The Baltic states, Russia, Ukraine, the Slovak Republic, Georgia, 93. The People Power Revolution in 1986 against the Government of Romania, Macedonia, Albania, and others successively introduced flat Marcos, and People Power II in 2001 against the Government of Estrada, taxes, with the Baltics starting the ‘experiment’ in the 1990s and others after reports that he had gambled with and otherwise misused public following subsequently. funds. 48  Political Economy of Public Financial Management Reforms specific approach taken to PFM reforms—albeit being constrained authorizing and implementing environment, that is, the relation- and filtered by the institutional setup and legacies as discussed ship between the executive and the legislature, and intergovern- above, which made it difficult to effect and entrench changes. mental arrangements. All of these issues need to be considered As a large country, Nigeria is intellectually vibrant and diverse, when seeking to pursue PFM reforms, and also when identifying, but politics has been more dominated by oil rents, regional issues, and on the one hand, what aspects (most) are in need of reform and, the need for balance between different regions to maintain national on the other hand, the feasibility of reforms. unity, than by any particular ideology. As discussed in Chapter 3, A pivotal dimension for enabling PFM reforms and good PFM the issue of geographic zones and of managing conflict—first in practices is the relationship between the executive and legislature. the south, then in the north (and resurgent in the south) have In cases where this relationship is difficult or even marked by dominated political and public debates. Combating corruption has confrontation and where this coincides with institutional rules not come into sharper focus in the run-up to the 2015 elections; while designed to avoid deadlock, there is a high risk that budgets get in 2016, there was also some backlash against this.94 CFA posi- delayed, and PFM reform legislation becomes difficult to adopt. tions have been variously held by technocrats and by loyalists. In turn, the nature of this relationship is itself underpinned by a Tanzania has seen a somewhat similar pattern, and similarly set of rules, typically at least in part constitutional provisions, on with a limited specific ideological influence on shaping fiscal policies the powers of each side, for example, with regard to introducing or PFM reforms. In the 1980s, Tanzania’s ruling party, the CCM, changes to the budget brought before the legislature by the execu- espoused ‘African socialism’ including collectivization of agricul- tive and obligations to react to proposed changes. ture and of industry. However, in the 1990s, it began swinging While the relationship is rarely without any problems, there is toward a more liberal policy agenda; including widespread priva- a great deal of variation, with problems being much greater in some tization. This privatization process offered many opportunities for countries and in particular time periods than in others. Countries with well-connected individuals and groups to benefit. The 2000s are a Westminster/United Kingdom tradition tend to have relatively therefore seen as a period that combined a trend toward ‘oligarchi- clear rules and veer toward executive dominance. In contrast, sation’ with some populist policies as an effort by elites to main- countries that have adopted United States-style rules (Nigeria and tain some social cohesion and sufficient political support. The the Philippines in this sample) tend to give the legislature exten- original liberal policy agenda of the 1990s coincided with a period sive rights to block executive proposals, and therefore are more of fiscal crisis, and this triggered a number of PFM reforms, such prone to deadlock. The Philippines illustrates that the challenge as introducing an Integrated Financial Management Information can vary over time when cooperation depends on political align- System (IFMIS) and strengthening commitment controls, as well ments between the two branches of government. as setting up an independent revenue agency. Some of this prog- Looking at the setup of MoFs and their functions across the five ress has subsequently slipped back, as documented in Chapter 2 countries analyzed here, a striking aspect is that there are far more and explored in detail in Chapter 5, during the period of greater ‘odd’ features than might be expected. Both Nigeria and the Philip- insider politics-cum-populist policies. Similar to Nigeria in 2015 pines have a particularly fragmented setup of CFAs. In both cases, and to the Philippines in 2010, this policy orientation has then these make it particularly challenging to decisively and coherently given rise to a stronger focus on anticorruption since 2015, which pursue functional reforms. These institutional features need to be has created some renewed attention on PFM (and procurement) carefully considered when developing proposals for PFM reforms. reforms. A key question that will be revisited in Chapter 6 is whether to seek reforms of basic institutional provisions when these have been identified as important bottlenecks. As a central credo of political 4.5 Chapter Summary economy, analysis holds that not every issue that is important to As for other policy issues, existing institutional arrangements are a reform is actually feasible to reform, at least not within short-term crucial aspect for the what and how of PFM reforms. This concerns time horizons (see also Levy 2014). One potential implication is on the one hand, two core issues—PFM-related laws and the insti- to engage more substantially and deeply with legislatures, so as tutional arrangements of CFAs—and on the other hand, the wider to prepare the ground for reforms. However, such an engagement would need to be deep and sustained to be likely to be successful. A potential alternative is to prepare changes, such as new PFM 94. As there was some conflation of the economic difficulties resulting legislation, and then to wait for particular windows of opportunity from the oil-price decline with the anticorruption efforts. The slogan to seek progress, for example, during early stages of a new admin- ‘bring back corruption!’ became quite popular, including a critique of the anticorruption efforts as being part of an international neo-liberal istration, preferably one based on a substantial majority. agenda: Moses Ochonu, Bring back corruption: A critique of neoliberal Intergovernmental relations receive considerable atten- anticorruption rhetoric, available at: http://dirayetu.blogspot .com/2016/09/bring-back-corruption-critique-of.html. tion from the perspective of fiscal arrangements and the quest for The World BAnk  49 decentralization, but they are less widely considered regarding rather than attention to whether funds are overall well allocated whether a certain PFM is feasible, can be effectively rolled out, and and used. is likely to contribute to improvements in service delivery. Whether In the broader picture, as traced in Chapter 3, several of the a particular reform is adopted and gains traction can crucially countries analyzed have experienced shifts in voter preferences in depend on whether there is sufficient buy-in at subnational lev- favor of improving governance, which in turn has motivated politi- els and whether the requisite capacity as well as oversight can cians to provide backing to PFM strengthening efforts. A key limi- be established. This is particularly difficult in federal systems, tation is that voters’ preferences in this regard can subsequently when major changes to intergovernmental relations are under way, again shift toward other priorities, such as delivering immediate and can be themselves highly contested, as in the case of Nepal. economic benefits or being tough on crime, rather than clearly Funds have to flow through subnational levels to reach frontline prioritizing better governance over several electoral cycles. service delivery, so unless improvements can be made throughout The limited engagement of the demand side can also be seen as the expenditure chain, central level efforts at improving PFM and a (so far mostly missed) opportunity to better explain and convey PFM frontline service delivery are more likely to remain disconnected. reforms to citizens, mixed with occasional overstatements about what This disconnect was one of the key findings of the cross-country specific PFM reforms are likely to deliver, as discussed further in analysis. It has potentially important implications because if PFM Chapter 5. While PFM reforms are technical and in part difficult to reforms deliver only a limited or no impact on the ability to pro- clearly understand for citizens, many of the issues, reform concepts vide services, an important political motivation for pursuing such and mechanisms can in principle be explained in ‘plain language’, reforms is lost. in particular to CSO staff who follow governance issues over time Across the cases and over time, the political-technical interface and hence have an opportunity to build an understanding. Espe- has ranged from strong political backing, to lukewarm or volatile, to cially as attention is shifting from deploying institutional forms to deliberately reducing capacity for good PFM. The person appointed improving actual functioning, it can be important to involve CSOs as a Minister of Finance is an important signal setting in many and regular citizens more in the monitoring of actual progress situations; but can also be used as a façade, not followed by actual made. This may also help to keep reforms on course over time. backing for reforms. Ministers who have come up through the The ideological orientation of governments for the cases ranks of a ministry tend to have a stronger grasp of and commit- reviewed has ranged from far left (Maoist) to rather neutral, to being ment to continue an existing reform agenda—as has worked quite influenced by libertarian ideas regarding the role of the state. As is well in Georgia despite a relatively rapid succession of ministers, reviewed in Chapter 5, despite this span in ideological influences, as initial appointees assumed wider political roles. PFM reform agendas have been remarkably similar across coun- The review across the five cases found that the ‘demand side’, tries; and for the time period reviewed, the governments concerned that is, organized civil society, tends to be a limited factor in driv- have been relatively fiscally prudent as observed in Chapter  3. ing PFM reforms, with a gradually intensifying engagement, and an From a developmental perspective, there is a range of ideologies interesting, more advanced experience in the Philippines. In all five which offer some positive aspects for development and poverty cases, there are some specialized CSOs who engage in budget reduction, while also entailing some risks (for example, seeking monitoring and to various degrees also in more detailed expen- an excessive role for the public sector or being too opposed to diture tracking and budget allocation discussions, as in Tanzania having an effective state or insufficient attention to social cohe- and the Philippines, respectively. As the Philippines experience sion on the libertarian side). To avoid being seen as partisan, DPs indicates, civil society capacity to engage effectively and more supporting PFM reforms have typically stayed away from engaging deeply on PFM reforms is built over substantial periods of time. more directly with such ideas. Generally, to truly understand the The Nepal example furthermore highlights the risk that civil soci- overarching policy goals, as discussed in Chapter 3, and the result- ety engagement on budgets can lead to expectations of increased ing specific aims, it is important to consider what ideas or ideology fiscal allocations for example, a particular locality or community, matter to a government. 50  Political Economy of Public Financial Management Reforms How Is Progress on PFM Being Made? 5 Reform Intentions, Steps Taken, rather than providing full-fledged technical analysis. Inevitably, while a number of specific reforms are covered, there are other and Results aspects not included here, for example, subnational PFM. This chapter offers a detailed process tracing of how PFM reforms happened in the five countries reviewed. It first looks at whether 5.1 Stand-alone Versus Embedded PFM reforms were sought in a largely self-contained way, or as part of wider public sector or governance reforms, then at the reform PFM Reforms intentions (typically as reflected in PFM reform plans), and also at PFM reforms have rarely been ‘the only game in town’ with regard the cross-cutting challenge of achieving partial reforms rather than to seeking to strengthen public sectors and service delivery. The as a complete sequence from reform intention to implementation breadth and depth of related reform efforts varies across countries. and routine use. Table 5.1 reflects the related public sector reform areas that were The chapter seeks to follow what has been done relative to ini- pursued in the five countries in parallel to efforts at strengthening tially stated reform intentions, how institutional arrangements and PFM. PFM reforms are considered as ‘highly’ embedded if they stakeholder constellations promoted or hindered change, and what were pursued as part of other major public sector and governance has been actually achieved with regard to reform progress. Follow- reforms, while embeddedness is considered ‘low’ if reform efforts ing the cross-cutting sub-sections (5.1–5.4), the chapter broadly were mainly limited to PFM as such, and if there were no or few follows the budget cycle in addressing the specific reform efforts efforts to achieve related other public sector reforms. from budget planning to budget execution and to external audit Embeddedness—that is, PFM reform efforts being part of a and oversight. wider reform package—was highest in Georgia and in the Philippines, It should be kept in mind that this section is intended to which pursued overall governance and public sector reforms, and was provide a closer look at nontechnical drivers of specific reform lowest in Nepal, with Nigeria and Tanzania as intermediate cases. As efforts to complement technical analysis of each reform aspect, the diverging fates of reforms in Georgia and in the Philippines Table 5.1. Related Public Sector and Governance Reform Efforts Degree of Embeddedness of PFM Reforms Related Public Sector Governance Reform Areas Georgia High Early years after 2004: Revenue administration and anticorruption Later years: Public administration, increasing citizen participation, and introducing results- oriented management Nepal Medium to Low Aim to establish good governance and strengthened social accountability; limited direct link/ relationship with PFM reforms Need to agree on a new constitution and system of intergovernmental relations Nigeria Medium to low up to 2015 Parallel efforts at civil service and payroll reforms Medium to high since 2015 elections Since 2015 elections strong emphasis on combating corruption with direct links to PFM reform agenda; urgency to improve revenue collection from oil and non-oil sectors Philippines High for 2010–2016 Overall governance reforms focused on anticorruption and citizen participation PFM reforms pursued as part of these wider efforts Tanzania Medium to low up to 2015 Some parallel efforts at strengthening results-oriented management (Big Results Now) Increased since 2015 elections Since 2015 elections strong emphasis on anticorruption; links to PFM reforms are not very systematic Source: Authors based on case study information.   51 indicate, greater embeddedness does not necessarily bring bet- at the midpoint between the 2010 and the next elections.96 Modeled ter results. However, there can certainly be positive synergies. In on the Malaysian performance management system, the initiative Georgia, the emphasis on again increasing revenue and on com- aimed to pursue accelerated reforms and results in key economic bating corruption created enabling conditions for PFM reforms to sectors, as well as with regard to revenue mobilization. The overall progress; and once significant achievements were made, this in aim of the agenda was to achieve middle-income status by 2025. turn enabled considerations about other governance reforms in the BRN was in part motivated by the expectation of significant addi- areas of results orientation, civil service management, and center tional public revenue from new natural resource exploitation, and of government policy coordination and monitoring. appears to also have been part of a longer standing effort to appeal In the Philippines, PFM reforms were similarly pursued as part to voters during a period of gradually declining support for the of a wider package of good governance efforts. The incoming admin- long-established dominant party. istration created five cabinet clusters, including one on ‘Good With the October 2015 elections, there was a change in empha- Governance and Anti-Corruption’. Among other measures, it also sis. The new government signaled an increased overall emphasis signed up to the Open Government Partnership (OGP). However, on governance reforms and on reducing corruption and wider rent- progress on the overall governance agenda as well as on the PFM seeking. Similar to Nigeria, the government did not initially set reform agenda ultimately remained limited. Greater institutional out any new broad reform plan. The fiscal trend in Tanzania was, and stakeholder complexity (both in terms of the country’s size and however, rather different. The successive discoveries of significant overall governance structures as well as the CFA setup) are seen on-shore and off-shore gas reserves reinforced international inter- as important contributing factors. For PFM reforms, a key strategic est in Tanzania’s gas fields and infrastructure.97 problem was that the adoption and rollout of a new IFMIS solution In Nepal, the integration of PFM reforms with other public sec- was made a pivotal part, but ultimately stalled (as discussed in tor governance reforms was overall more limited. Laws seeking to Section 5.5). improve governance and accountability were adopted during the In Nigeria, up to 2015, the PFM reforms were not linked to a early postconflict period (Right to Information Act, 200798 and strong agenda of wider governance or public sector reforms, except Good Governance Act, 200899). However, given the need to devise for some links to efforts at better payroll management. The years from a new constitutional framework as well as instability and frequent 2007 to 2015 were overall marked by a slowdown in governance turnover in the government, actual pursuit of a broader good gov- and public sector reforms, following earlier efforts made to achieve ernance or public sector reform agenda remained limited. Build- the HIPC completion point. One important link was to strengthen ing on the legal framework set during the early postconflict years, payroll management and, in particular, to reduce the number of DPs supported various initiatives aimed at strengthening social ghost workers on the payroll. accountability, in particular at local levels. The main attention of Following the election of a new government in 2015 and the the government and other stakeholders was focused on develop- decline in oil prices since mid-2014, the governance and public ing a viable constitution and intergovernmental relations which reform agenda in Nigeria has widened and is being pursued with remained hotly contested. There was some intersection of these greater intensity. In particular, increasing non-oil revenues have two agendas with regard to intergovernmental fiscal relations. become significantly more prominent. The government also sought Across the five cases, there are clearly some important links to reinvigorate a number of the public sector and PFM reforms between PFM reforms and other ongoing public sector and gover- that had been initiated earlier, but that had been left lingering. In nance reform efforts. There can be important synergies between addition to revenue collection and more seriously tackling ghost PFM and other public sector reform efforts, in particular changes workers, improving the management of the oil sector has become to the civil service. Civil service rules, regulations, and (de facto) significantly more prominent.95 Thus, given that the country management influences what options are available, for example, incurs huge losses from poor management of the sector, and that for building capable and effective Ministries of Finance. Moreover, reforms were delayed during good times, the country faces a rather anticorruption efforts, if serious, can have important complemen- daunting agenda of implementing economic governance reforms tarities with seeking to reduce leakages in the use of public funds. while managing a fiscal crisis—both of which are closely linked and influence PFM reforms going forward. Before the 2015 elections, Tanzania adopted the ‘Big Results 96. The initiative was unveiled in February 2013 by President Kikwete. Now’ initiative, developed in 2012 and launched in early 2013, that is, 97. See Roe (2016). 98. See: http://www.moic.gov.np/upload/documents/right-to-information- act.pdf. 99. See: http://siteresources.worldbank.org/NEPALEXTN/ 95. See also http://www.statehouse.gov.ng/index.php/news/ Resources/223554-1296055463708/PoliticalEconomy.pdf; for the speeches/1801-president-buhari-s-2016-budget-address, and http://www specific acts: http://faolex.fao.org/docs/pdf/nep137755.pdf and http:// .reuters.com/article/us-nigeria-politics-idUSKCN0PH1EZ20150707. www.mopit.gov.np/links/susasan-ain-english.pdf. 52  Political Economy of Public Financial Management Reforms When seeking and supporting PFM reforms, it is important to and specific programs and activities, across levels of government, remain mindful of links between PFM and other reform efforts. Analy- and also between existing commitments and new demands. sis of the status and progress made as well as operational engage- Given that institutions and staff tasked with managing public ment tends to be compartmentalized between PFM, civil service, expenditures face these pressures and pursue similar aims across and other reforms, such as establishing results-oriented manage- countries, it makes sense to copy good practices that have been ment. While compartmentalization is useful to promote the neces- developed, rather than seeking to ‘reinvent the wheel’. Good prac- sary technical depth, it can lose sight of crucial links. One key link tices for drafting budgets, for cash management, for accounting is that governments interested in PFM reforms are often at least and auditing, and IT systems to support these functions have been equally interested in pursuing a strengthening of revenue admin- developed and are therefore available to be adopted by countries istrations and of bringing public sector payrolls under control—as interested to do so. each of these measures can contribute some fiscal space, and they At the same time, there are also concerns that this relative are often initiated when such space is urgently needed. Moreover, standardization amounts to ‘isomorphic mimicry’—that is, superficial reform efforts across the range of government functions can affect copying. The most widely cited concern is that this leads to ‘reform the overall fiscal position. facades’ without actually achieving expected improvements Furthermore, looking at PFM reforms in a compartmentalized (Krause 2013; Pritchett, Woolcock, and Andrews 2010; Andrews way heightens the risk that DPs promote expansive PFM reform plans, 2009). Other potential concerns are that overly standardized rather than focusing in a more agile and cross-cutting way on key reforms lead to changes that do not really resolve key bottlenecks bottlenecks to improve the overall public sector and to deliver public (Blum, Manning, and Srivastava 2012; Bunse and Fritz 2012; goods and services. However, it is crucial that such reforms achieve World Bank 2012; Brinkerhoff and Brinkerhoff 2015). Related to tangible results to make them politically (more) attractive and to this is a concern that at least some of the reform approaches are ensure that initial plans are actually followed through to full imple- not sufficiently ‘real’, that is, they are based on ideas about how mentation, as discussed below. Compartmentalization can also PFM systems could or should work, but in ways that may not have increase the risk that various initiatives are not well integrated, been practically achieved in any country. For example, while zero- such as separate efforts at establishing results-based management based budgeting is an attractive practice in principle, in particular driven from the center of government and efforts at program bud- for an incoming administration, it has rarely been (fully) used in geting in the Ministries of Finance. practice; and even full implementation of performance based bud- geting remains rare (Curristine 2005; Curristine, Lonti, and Jou- mard 2007; Robinson and Last 2009; Lienert 2012; Moynihan 5.2 Reform Intentions—Similar and Beazley 2016). Intentions Despite Diverging Problems Further potential concerns are that strong standardization reduces the room for innovations; and that some reforms which work PFM reforms are marked by significant uniformity of reform intentions well in upper middle and high income countries are not fully suited across countries. This holds true with regard to the overall reform for low and lower middle income countries. Fiscal resources in LICs goals, and—possibly even more so—with regard to the reform tools and LMICs tend to be more constrained and more volatile. At the that are being pursued. The standard overarching reform goals are same time, corruption and other mismanagement risks tend to be the trio of (a) fiscal stability, (b) allocative, and (c) technical effi- greater in such contexts, and are often more systematic rather ciency (Schick 1998; PEFA 2016 Framework). Reform tools refers than limited to individual instances. Moreover, as North et al. to the specific institutional mechanisms being introduced that are (2013) have emphasized, introducing institutional models of expected to contribute to these goals, such as MTEFs as an instru- overall well-governed countries into a context of countries with ment (or tool) to promote fiscal stability. great power imbalances and many poorly functioning institutions There are several reasons for the strong similarities in reform entails significant risks of failure as well as of unintended negative aspirations. Managing public finances is an activity that faces in consequences.100 many ways similar fundamental challenges across countries: how Against this background of ongoing debates, reviewing PFM to ensure that (a) public funds do not leak, are used effectively and reform plans as developed for the five countries analyzed reveal a efficiently, and are reported on so as to enable effective oversight considerable degree of similarity. Table 5.2 provides an overview of and accountability and (b) from a fiscal perspective, how to ensure eight specific PFM reform aspects—ranging from those focused that spending does not exceed revenue by too large a margin—in the face of (far) greater needs than can be met through avail- able funds. Also, there is a need to weigh the benefits of different 100. North et al. (2013) use the terminology of ‘Open Access Orders’ spending demands or needs against each other—across sectors and ‘Limited Access Orders’. The World BAnk  53 Table 5.2.  Key Elements of Reform Plans Across Countries Georgia Nepal Nigeria Philippines Tanzania MTEF √ √ (halted) √ √ √ Program Budgeting √ x x √ √ Introducing TSAs √ √ √ √ √ (I)Financial Management Integrated √ √ √ √ Initial introduction System (FMIS) in the late 1990s Introducing IPSAS √ √ (cash basis) √ √ √ Procurement reforms and introducing √ √ x E-procurement platform introduced during earlier Partially in place e-procurement reform wave, e-bidding still to be completed Introducing/strengthening Internal √ √ √ √ √ Audit Strengthening External Audit √ √ x √ √ Source: Authors based on reform plans and statements for the five countries. √ signifies that reform plans mention the instrument as being intended; x = no mentioning of this reform instrument. Note made if an instrument was already introduced during past reform efforts. on upstream budgeting to those focusing on budget execution and countries differ in important respects. Important areas of variation audit—and checks whether these are mentioned in the reform pro- are the track record and rules regarding timely approvals of bud- grams of the five countries covered. With a few exceptions, most of gets, variation in the effectiveness of cash management and of the eight types of reforms are being pursued across all five coun- controls, capacities for PFM, and transparency, as well as in the tries. For a few aspects, selected countries already made progress institutional setup for effectively linking planning, budget prepara- before the reform periods considered for this report, for example, tion, and budget execution. with regard to the introduction of new procurement legislation and Given that major delays in budget approvals have severe knock- e-bidding in the Philippines (introduced as part of a reform wave on effects on budget execution, one might expect to see intended that had started in the late 1990s),101 or the initial introduction improvements in this regard, prominently reflected in reform plans and of an IFMIS in Tanzania. The specific reform attempted least fre- strategies, but this is not the case. Based on the country situations quently is the introduction of program budgeting, which is widely as reviewed in Chapter 4, an emphasis on moving to more timely recognized as a rather advanced reform. budget approvals would be anticipated in particular in Nigeria and While the cases studies covered here mostly have an anglo- Nepal, and possibly also for the Philippines. In a sense, seeking phone heritage or links, the similarity of PFM reform plans is not lim- to pursue an MTEF ahead of improvements to the annual budget ited to this group. Notably, francophone countries in West Africa cycle (combined with instruments for managing fiscal volatility), have similarly planned significant reforms to their systems of pub- appears to be both an ‘error of exclusion’ and an ‘error of inclusion’ lic financial management over the last decade. This has at least in with regard to reform intentions. It may still broadly make sense part been inspired by PFM modernization efforts in France which to develop fiscal plans and projections beyond a single year, but took off in the early 2000s with the adoption of a new organic not to try and establish a true MTEF. For Nepal, this challenge was budget law, the Loi organique relative aux lois de finances (LOLF), realized, at least to the extent that further efforts to develop an adopted in 2001 and fully in force since 2006. In 2009, the West MTEF were halted. However, as is discussed below, such a reform African Economic and Monetary Union (WAEMU) adopted a num- focus has actually been missing in the countries concerned. ber of new directives for PFM reform, which member countries are Other areas where the problem constellations differ significantly required to transpose into national law.102 However, actual prog- between countries is cash management as well as the effectiveness of ress in transcribing them into national law and actual implementa- controls. Good enough cash management is clearly a crucial PFM tion still remains limited. function for supporting effective service delivery. Sector ministries The uniformity of reform intentions as set out in the reform plans and agencies need to have predictability in when and how they is rather striking in light of the fact that while there are similarities can access funds for spending on their service delivery activities, in general PFM tasks, the specific problems faced across the five including funds for operations and maintenance, and funds for planned capital investments. Some countries have already well- established cash management systems and an adequate balance 101. See: http://adbprocurementforum.net/wp-content/uploads/Phil-eGP- between revenues and expenditure commitments, while in others, case-study-Sep2011.pdf. 102. http://www.internationalbudget.org/2013/12/are-public-finance- in the sample notably in Nigeria and recently again in Tanzania, management-reforms-working-in-francophone-african-countries/ cash management continues to be or re-emerges as a challenging 54  Political Economy of Public Financial Management Reforms issue. Tanzania’s PFM Reform Program (PFMRP) IV stands out as and political principals, and hence appear as being pursued some- a reform plan that encapsulates some notions of what has worked what by default. and what has not, and that sets out priorities accordingly, includ- As emphasized at the outset, seeking to approximate or adopt ing the issue of cash management. This appears to be motivated international standards is not a bad idea as such. It would make also by the fact that a number of reform steps other countries little sense for each country to ‘reinvent the wheel’ of medium- have been grappling with over the past 10–15 years were already term planning, accounting practices, or treasury management. started in Tanzania in the late 1990s, thus creating a longer track Current international norms and standards have been developed record of experience with PFM reforms. out of decades of practice and refinement. Moreover, having a While institutional reforms are clearly not an end-goal in them- more standardized way in which budgets and annual accounts are selves, inefficiencies and bottlenecks in the institutional setup—as presented can bring important benefits in terms of transparency reviewed in Chapter 4—are something that one might expect PFM and comparability of public expenditures. reform efforts to address at least selectively. However, this has been Nonetheless, too much ‘isomorphism’ by default is not a good much more rarely the case than might be anticipated. The difficulty idea. Pursuing reforms that do not really target and hence do not of agreeing on institutional changes is likely to be one important alleviate key bottlenecks in PFM systems risks to sooner or later reason. Institutional issues as a causal factor received significant result in frustration. Various stakeholders may lose the appetite for attention with regard to fiscal issues—notably the size of deficits further dialogue and reform efforts, including development agen- and the tendency to incur debts (see for instance, Dabla-Norris cies funding engagement on PFM and political principals in the et al. 2010)—but much less so with regard to strengthening PFM. various countries. Also, reforms that fail to stick due to (a mix of) After advocating for the need to integrate planning and bud- limited political commitment or significant opposition, capacity geting for some time in the past, the current international PFM dis- constraints, and weaknesses in complementary institutions, such course seems to have reduced attention in this regard and seems as good enough civil service management, or an ability to con- to pursue more an ‘adding on’ of additional or new features—such strain corruption, can be a waste of effort, which could rather have as processes for an MTEF or for program budgeting—to whatever gone into more feasible areas of reform. At a minimum, greater institutional landscape exists. PEFA reports include sections on selectivity of what is being pursued at a given period of time would institutional arrangements, but as these are descriptive and are seem important. not rated and as PEFA does not really explore causal relationships, these descriptive sections rarely come into center view in the sub- sequent discussions on reform plans. 5.3 Partial Reforms as a Frequent Importantly, while the uniformity in reform intentions is notable Occurrence and at least in some ways problematic, this is not equivalent to the Apart from the resemblance of reform intentions, a striking feature diagnosis of externally imposed or incentivized isomorphic mimicry. when reviewing PFM reforms across countries is the frequency of par- Three of the five countries covered here are MICs and, for most, tial reforms, combined with considerable confusion in whether reforms aid dependency is limited or low. Tanzania is probably the sin- are or are not (fully) implemented. As will be discussed in detail in gle case with a pronounced tendency toward a donor-driven PFM the subsequent sections, some years into implementation, reforms reform agenda during the period reviewed (with strengthening gov- frequently remain only partially completed. For example, a TSA ernment commitment since 2015). Thus, aid-related conditional- may cover some MDAs and levels of government, but others might ity should not be viewed as the main mechanism (or culprit) for still hold separate accounts in commercial banks, or an MTEF bringing about the observed tendency toward a standardized set of document was produced for some years, but not for other years, reform intentions. or may be continuously produced, but with limited real effects Rather, two reasons appear to be at play and will be explored on annual budget allocations. Different reforms remain ‘partially further in the following sub-sections of this chapter. One is that inter- implemented’ in various ways—with some remaining at prepara- national norms, standards, and professional networks constitute tory stages, while others are only applied to some parts of public a powerful set of drivers toward uniformity and copying. These expenditures. networks have arguably grown denser and more influential over Partial progress is not a bad thing as such, especially as long as the past decade, in particular with regard to accounting standards it is an intermediate stage toward eventually more complete reforms. and internal and external auditing. A second reason seems to be In many situations, a staged or sequenced approach to a particular that absent major fiscal crises, isomorphic reforms are the ‘lowest reform area may be more sensible and realistic than a complete common denominator’ between key stakeholders, including senior overhaul or very rapid rollout. For some aspects, a partial reform is technical staff in central finance agencies, development partners clearly preferable to inaction, and may have been the only feasible engaging in the dialogue on PFM even with limited real leverage, The World BAnk  55 approach, as was the case, for example, with introducing automa- reform aspects below. Chapter 6 outlines some options of how tization of budget execution in Nepal. the problem of partial reforms can be built more explicitly into However, there are several important issues to flag. This reform planning, as well as into external support programs for includes considerable risks of confusion about the actual status of PFM reforms. reforms, risks of stagnation and backsliding that need to be rec- ognized, as well as problems with having partial systems in place where ‘old and new’ need to be operated alongside each other. 5.4 Clarity of Expectations—What Are This can lead to conflict with legal rules, for example, if these were Specific PFM Reform Efforts Expected changed to endorse the new system only. A further issue is the fact that partial reforms may not be ade- to Deliver? quately recognized as such. There is a certain tendency to proclaim A further problem that is noticeable when looking at PFM reform that a certain tool—such as an MTEF, e-procurement, or a TSA—is processes is that there is often some confusion about what they are ‘already in place’. In fact, in most countries, there are many years expected to deliver, and how specific reforms are expected to lead to not just between the initial planning and the start of implementa- these aims(that is, the specific causal chain or ‘theory of change’). tion, but also between the start of implementation and having a At an aggregate level, as reviewed in earlier chapters, PFM reforms complete system in place that (at least largely) works continuously are meant to improve budget credibility and to feed into improved and as intended. If partial or incomplete rollouts are not suffi- efficiency in service delivery, by creating a better use and avail- ciently recognized as such, then attention can shift too quickly to ability of funds. Linked to these overarching goals, PFM reforms the next area, neglecting the process of completing what has been are also seen as a key component of ‘improving governance’ and started, and ensuring that it becomes fully embedded and useful. reducing (opportunities for) corruption. However, how individual Such shifts in attention can also increase risks of backsliding. reforms are expected to contribute to these goals is often not A related problem is that partial implementation can get stuck explicitly spelled out, for example, in PFM reform plans. for extended periods of time. This may happen because a reform is Once a specific PFM reform effort—such as establishing an not fit for purpose or for the context, as well as due to resistance MTEF or contracting and rolling out an FMIS—starts, attention is during implementation. For example, a TSA effort may result in drawn to the goal of and difficulties associated with making the a significant share of commercial accounts of public agencies reform as such happen, rather than the expected impacts. Hence, being closed; and then eventually progress to cover those parts the reform then becomes judged as progressing if the specific of government earlier left out. However, it may also happen that instrument is being put into use. the effort comes to stagnate in a limbo with some agencies suc- There is some tendency to overstate expected likely impacts. cessfully resisting the closure of their accounts. A partial imple- For example, program budgeting can be advertised to deliver a mentation that remains ‘stalled’ is likely to entail a greater risk more strategic use of public funds—but given the challenges with of eventual backsliding. For example, if a number of MDAs are establishing meaningful indicators and links with budgets, this allowed to keep their own bank accounts, then others may eventu- aim is rather rarely achieved. Rolling out an FMIS may be expected ally also seek exceptions, resulting in reopening of accounts by to deliver both much better real-time visibility of how money is many MDAs. In Nigeria, a TSA reform was started in 2010, but by spent, greater accountability, as well as providing a system of early 2015, there were still numerous separate MDA accounts with commitment controls (World Bank, forthcoming). While the former very substantial balances. might emerge as a result, its contribution to accountability will Furthermore, a partial rollout of new systems which were depend on whether and how quickly budget execution reports are intended to be rolled out comprehensively or used consistently can actually shared with the public. The effectiveness of controlling create new problems. For example, if new IT systems are only par- commitments will remain dependent on whether central political tially rolled out or used due to various constraints, but uninten- decision makers are interested in reigning in all spending agencies tionally so, this can create problems when there are no explicit that show a tendency to overspend or do so only selectively; as well provisions for manual or semi-manual backup options. Adopting as on fiscal trends and on whether a broadly realistic budget was new accounting standards which are not consistently used can adopted in the first place. As Hashim (2016, ix) has pointed out: create a host of challenges, including a drop in actual transpar- “Implementation [of FMISs] is, however, often associated with dis- ency when the government’s accounts become less intelligible and appointing results and attribution to higher-level public financial consistent. management objectives difficult to establish.” As reviewed in Sec- Across the five countries covered here, partial and incom- tion 5.5 below, in one country, changes in accounting standards plete reforms are common. This holds for reform efforts across the were billed as ‘expunging corruption’, which clearly is an over- budget cycle. This is highlighted in the discussion of specific statement of what such a specific reform on its own can achieve. 56  Political Economy of Public Financial Management Reforms Mechanisms and efforts to track actual impacts of PFM reforms more fractious support in the two chambers of the legislature.103 still require greater attention and development. To some extent, regu- New or additional pressures on budget preparation can also come lar PEFA assessments offer some check on whether the reforms from fiscal contractions, such as that being experienced by Nigeria that have been pursued over the preceding years are yielding since 2015. results; but these reports still only track parts of the expected Efforts at introducing budget preparation reforms have been results chain. For example, PEFA reports may assess whether made across the five countries, as reflected in Table 5.3. The top reporting is done more regularly and more timely, but do not (and row revisits the status of ‘budget preparation and adoption basics’, cannot) investigate in depth whether the reporting done by sub- in particular whether budgets have been prepared on time and national levels and line ministries is reliable. They also cannot are reasonably credible. Georgia, Tanzania, and in recent years, fully track whether funds reach front-line service delivery units in the Philippines have had timely budget preparation and adoption. a more regular and complete way, something that requires more of However, even among these three, only Georgia has consistently a Public Expenditures Tracking Survey (PETS)-type analysis. adopted reasonably credible budgets, when considering subse- quent shifts in allocations as measured by indicators PI-1 (aggre- gate) and PI-2 (composition) of the PEFA framework (last two rows 5.5 Upstream Budget Planning: of Table 5.3). Nepal and the Philippines have recorded limited Getting the Basics Working, deficits, as reflected in Figure 5.1, while the remainder have seen larger swings, in particular in the wake of the global financial crisis and Efforts at Introducing MTEFs (and conflict in the case of Georgia in 2008). and Program Budgets In terms of realization, MTEFs fall into the category of ‘partial reforms’ in three of the five countries. MTEFs are being produced Deciding on budget allocations is an inherently political process, and discussed to some extent, but are not really used to set expec- in which key political stakeholders participate on recurrent basis. tations about spending beyond the annual budget cycle. This is As Hallerberg, Scartascini, and Stein (2009); Norton and Elson especially an issue for Nepal and Nigeria. Furthermore, across the (2002); and others have emphasized, institutional rules and the five countries, having a functioning MTEF mirrors the degree to dynamics among political actors influence decisively whether the which the annual budget process is running well. Countries where process produces budgets that are (broadly) realistic and adopted annual budgets are delayed—and sometimes delayed for several on time, as well as the actual allocations made. The basic thrust of months—are not in a position to establish functioning MTEFs or PFM reform efforts across many countries has been to strengthen even MTFFs (see also World Bank 2013). links between planning/policy intentions and budgeting and to Tanzania is an interesting case where an MTEF was initially make these links more tangible and visible through the intro- introduced during the reform wave of the late 1990s, with efforts con- duction of program budgeting, and to introduce a medium-term tinuing for the 15 to 20 years since then. A precursor was introduced perspective, which also supports planning and budget allocation even earlier, in the early 1990s, when a Rolling Plan and Forward links (projected to the medium term), as well as supporting fiscal Budget was first introduced (SADCOPAC n.d.; Holmes and Evans sustainability. Relatively less thought tends to be given to how 2003). By 2003, the establishment and use of the MTEF in Tanza- these reform efforts link back with the political dynamics related nia was assessed as being ‘in progress’ (Holmes and Evans 2003). to budget preparation. According to Oyugi (2008), the MTEF had become comprehensive It is also worth remembering that a basically functioning annual and linked planning and budgeting, but at the same time he noted budget process requires quite a few institutional and political fac- that annual budgets remained unrealistic. One of the most recent tors to be well aligned: laws and regulations, systems to develop assessment concludes that while aggregate fiscal discipline has budget plans and staff capable of working with these, concerta- improved, supported by the MTEF, within-year allocations to line tion between stakeholders within the different parts of the execu- ministries remain variable and subject to cash availability; and vari- tive, as well as between the executive and the legislature to move ation between budgeted and actual expenditures remain substan- from stage to stage within prescribed deadlines, and an ability to (SADCOPAC, tial especially for capital-intensive areas of spending ­ reach an agreement about expenditure trade-offs along the way. n.d.). Thus, utilization and embeddedness is substantial for the Changes in the political dynamics therefore remain a risk even Ministry of Finance, but remains in progress and more uncertain in countries where such systems have been well established for for line ministries. Overall, implementing this reform instrument some time. For example, in the Philippines—which in principle has a powerful Congress deliberating on the budget for several months—there is a risk that timely budget approval could come 103. Note that this risk did not materialize following the 2016 elections; under pressure again if a future President has to rely on smaller or and the 2017 budget was passed on time. The World BAnk  57 Table 5.3.  Overview of Budget Preparation Status and Reform Efforts Georgia Nepal Nigeria Philippines Tanzania Budget preparation Annual budgets Annual budgets Annual budgets Late approvals until Annual budgets have and adoption basics adopted on time and are adopted with are adopted with 2009, timely approvals been adopted on time reasonably credible significant delays significant delays since then (but not Aggregate credibility Aggregate credibility Budget credibility is institutionalized) has been reasonable, has been relatively low in the aggregate Problems persist with but worsening; high and across sectors budget numbers and reallocations across overall credibility sectors are substantial; development budgets significantly underexecuted MTEF/MTFF Basic Data and First introduced in Efforts started in 2004 First tried in 1999, MTEF first introduced Directions (BDD) 2002, then withered to introduce an MTEF then re-introduced in the late 1990s Document introduced Efforts restarted in and medium-term since 2006; relatively after fiscal crisis, and since 2005 2015/16 sector strategies, but effectively during requirement for HIPC not seen as having Aquino administration since 2001; still a much traction number of limitations but progressing Program budgeting Piloting started in 2006 Not started Has not been pursued First introduced in 2014 Program-based to date currently ‘performance budgeting initiated informed budgeting’ under PFMRP IV (2012) P1 ratings A A C D B P2 ratings B+ C+ D D+ D+ Note: MTFF = Medium-Term Fiscal Framework. http://www.imf.org/external/pubs/ft/scr/2016/cr1629.pdf. Figure 5.1.  Annual Deficits 2000–2016 10 8 6 4 % of GDP 2 0 –2 –4 –6 –8 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 Georgia Nepal Nigeria Philippines Tanzania Source: International Monetary Fund, World Economic Outlook Database, April 2017. These data are calculated as revenue (including grants) minus total expenditure. 58  Political Economy of Public Financial Management Reforms remains partial even after nearly two decades; bringing some ben- state level, interlocutors reported that they include significantly efits, but also clearly having still important limitations. more expenditure requests in the annual budget than they expect Program budgeting has been pursued as a reform in three out of to be able to actually fund—signaling that spending trade-offs Tanzania— the five countries covered—Georgia, the Philippines, and ­ are not resolved during the preparation process, but deliberately but still remains in early or intermediate stages in each of them. postponed to the budget execution stage to manage political pres- Among the three, Georgia has sought the longest to establish a sures.107 Such a ‘defensive’ mechanism of the executive when system for program budgeting, starting initial pilots in 2006. How- faced with overwhelming demands undermines the aim of having ever, it has been challenging to include meaningful targets and to a realistic budget, a close fit between budgets and outturns, and connect spending to specific results. Ten years into this effort, the being able to have agreed medium-term ceilings. Line ministries government is seeking to ensure that line ministries include more that anticipate receiving fewer resources than allocated are also relevant information in this regard into their budget submissions. less likely to seriously engage with more advanced budget prepara- In the Philippines, the efforts are more recent. The effort can ben- tion reforms. efit from the fact that other aspects of performance management Realistic budgets are only possible if ‘within government’ stake- have been more established as well as the fact that civil society holders are interested in and able to come to an agreement on nec- interest in the government performance has been developing since essary spending trade-offs ex ante and to do so comprehensively for the late 1990s.104 In Tanzania, program-based budgeting was all available fiscal resources. This includes that stakeholders are started as part of the fourth PFM Reform Plan (2012–2016), with willing to make these trade-offs transparent to stakeholders out- individual ministries implementing program-based budgeting on a side the government—such as voters and interest groups–through pilot basis, using a template developed by the MoF.105 a comprehensive and published budget document. For example, In terms of underlying drivers, the following stand out: at an in Nigeria, a large share of fiscal allocations remains outside the aggregate level, budgets have been credible in three of the five case federal budget; including spending on fuel subsidies, transfers, study countries during the period reviewed. Thus, the governments and funding for special programs–thus not being part of the dis- managed to keep overall actual spending in line with expected cussions about trade-offs within the budget envelope. envelopes.106 As discussed in Chapter 3, aggregate credibility Regarding more advanced reforms to budget planning, it appears to have been helped by conditions of substantial growth appears that in particular MTEFs are frequently tried and are rarely in many low- and middle-income countries during this period, and openly opposed by any stakeholders, but it is difficult to make them a general awareness that excessive deficits can be costly, follow- binding. An effective MTEF relies significantly on whether stake- ing painful ‘structural adjustment’ experiences in the 1980s and holders in the budget process are willing to bind their hands to a 1990s. Still, contentious relations between executives and legis- specific preplanned pattern of allocations. In the Philippines, this latures can undermine timely budget approvals as well as aggre- process began to work during the Aquino administration when rela- gate credibility, as has been the case in Nigeria, and also in the tionships within the executive and between the executive and the Philippines. legislature were relatively aligned (see also 2016 PEFA report). In However, political dynamics within the executive as well as Tanzania, although the same political party has been continuously between the executive and the legislature can also be a source of in government and MTEFs have been produced for many years, budget planning that lacks credibility with regard to inter-sectoral the mechanism remains relatively ineffective in particular with allocations. As not all line ministries are politically equal, some regard to sectoral breakdowns, as these are not realistically agreed expect to be able to lobby for additional allocations during bud- among stakeholders. The low-income context may also make it get execution. In Tanzania, politically well-connected ministries particularly challenging; a general scarcity of funds relative to an deliberately under-budget some high priority expenditures, so as expanding role of the state with regard to the delivery of goods and to leverage these for in-year additional demands. In Nigeria at the services makes trade-offs especially difficult. The experiences with seeking to introduce program budgeting suggest that having clear demand for programmatic and performance 104. As argued by Moynihan and Beazley (2016), program or performance budgets are more likely to be meaningful if part of a wider information is crucial. Without clear demand—from parliaments, system of performance-oriented management. citizens, and/or the top political leadership—the bureaucratic 105. See: http://www.mof.go.tz/mofdocs/news/latest%20news/ GUIDELINES%20FOR%20THE%20PREPARATION%20OF%20 default is toward superficial implementation. Managers are likely PLANS%20AND%20BUDGET%20201516%20-INSTRUCTIONS%20- to be reluctant to commit to achieving specific outcomes, and they Web1.pdf. may feel that their degree of control over achieving targets is insuf- 106. A relatively extensive literature looks into some of the institutional and political factors that can lead to excessive fiscal commitments, ficient especially in contexts with cash constraints, as well as due or commitments which stakeholders expect to be altered during implementation (von Hagen and Harden 1994; Alesina and Perotti 1995; Alesina 2010). 107. Interviews in Bayelsa and Delta State, May 2013. The World BAnk  59 to other factors (see also Worthington 2013). In Georgia, program From a fiscal perspective, a TSA is an attractive reform, since budgeting has been pursued with considerable commitment from it promises to ‘free up’ resources sitting idle in the accounts of vari- key stakeholders in the Ministry of Finance; but thus far, neither ous ministries and agencies with commercial banks. In addition to parliament nor nongovernment stakeholders have really taken a providing some fiscal space, a TSA gives the Ministry of Finance strong interest. As a result, it has been difficult to persuade line (or other responsible CFA institution) a tool to see when and on ministries to pay serious attention to the design of indicators and what MDAs spend funds, that is, greater managerial oversight and the relationship between budget allocations and results promised. accountability (see also Fainboim and Pattanayak 2010). More- In the Philippines, in principle, there could be some opportunity over, in combination with automating treasury functions, a TSA for stronger demand in this regard given its more advanced civil can be an important tool for reducing face-to-face interactions society engagement on budgeting; while in Tanzania, effectiveness between treasury officials and suppliers seeking to get paid, thus could emerge in some sectors but less so across the board. reducing a potential source of corruption. At the same time, there are several potential sources of resis- tance, with the main ones being line ministries and agencies and 5.6 Progress and Challenges with other spending entities, including subnational levels of government Regard to Improving Budget Execution on the one hand, and commercial banks handling government funds on the other. Line ministries and agencies typically have an incentive A range of potential PFM reforms evolve around budget execution. against a TSA; including subnational levels and agencies. Having This includes systems for cash management and commitment their own accounts can offer a (legitimate) buffer against delayed controls, for payments, for the management of particular expendi- budget releases, as well as offering greater opportunities for leak- ture streams such as payroll on the one hand, and public invest- age, given the more limited visibility and ‘trackability’ of spending ments on the other, accounting and reporting rules, standards and without a TSA. Therefore, the CFA has a stronger case for rolling practices, and internal audit. Furthermore, while IT applications out a TSA if it can credibly promise109 that cash releases will be cut across budget preparation and budget execution stages, they made regularly and as budgeted. There can also be issues around are often first introduced to facilitate managing the execution of the status of subnational governments, which may argue that they budgets, given the high volume of data that needs to be captured. are autonomous and hence should not be part of a TSA operated The particular reform efforts that this sub-section focuses on by the central government. Apart from resistance from within the include efforts to introduce (a) TSAs, (b) FMIS, (c) IPSAS as a new government, another group of losers from a TSA reform are com- set of accounting standards, (d) the strengthening of internal audit mercial banks operating accounts for MDAs. In the absence of to ensure ongoing checks against leakages during the budget exe- a TSA, commercial banks may hold considerable deposits from cution process, and (e) the introduction of e-procurement systems government entities, including from revenue authorities, providing and related procurement-focused reforms. As for budget prepara- an important source of liquidity. Consequently, banks may lobby tion and approvals, the section also looks at some of the ‘basic’ against losing this source of funds. issues in the expenditure process, in particular with a view to basic The experience with how treasury management has improved management of budget releases and of cash during the year and in- across the five countries illustrates these challenges—as well as sug- year reporting (see also Schick 1998; Welham and Hadley 2015). gests some options for how these can be addressed when considering both technical and nontechnical drivers of change. The experience 5.6.1 Introducing Treasury Single Accounts indicates that (a) a TSA quite typically remains a partial reform The introduction of Treasury Single Accounts has been pursued by either by design or by default and that (b) a step-by-step rollout by Governments of all five countries in the sample. However, progress design may often be the best approach, given the combination of has been mixed. The greatest degree of progress has been made likely points of resistance and limitations to technical capabilities. by Georgia, followed by Nepal. A TSA was initiated but remained In Georgia in 2004, the government was extremely keen to largely stalled in Tanzania and in Nigeria at least up to the changes expand fiscal space on the one hand, and to curtail corruption on the brought about by the 2015 elections. In Nigeria, the government other hand. Since independence in the early 1990s, a network of newly elected in 2015 revived efforts to close the accounts held by treasury offices had been established; while MDAs as well as local individual MDAs in commercial banks.108 In the Philippines, a TSA governments maintained their accounts with various commercial was established for revenues, but it does not cover expenditures. banks that had been established since the end of the Soviet Union. As a first step, the new government established the nucleus of a 108. See various news reports in early 2016, for example, http://allafrica 109. See also the WDR 2017 on the importance of credible .com/stories/201602181441.html. commitments. 60  Political Economy of Public Financial Management Reforms TSA used for revenues. This was followed in 2006 by establishing failed, the efforts restarted in the late 2000s. Initial discussions the TSA for central-level MDAs. This process involved the closure were held in 2008, following an IMF Article IV, resulting in plans of around 12,000 accounts at commercial banks. The TSA went for a new but more gradual attempt. In 2009, piloting commenced hand in hand with the establishment of an e-Treasury system (see for central government and three districts. The process had con- below), the nucleus of an IFMIS. siderable external support—it was funded by an MDTF, and fol- A critical concern of key government officials in the Ministry of lowed IMF advice, guidelines, and technical assistance. However, Finance at the time was that the daily interactions between treasury the actual design was led by Nepali senior civil servants using officials and those seeking payment was very problematic. Suppliers Sri Lanka—that is, a regional neighbor—as an example. Moreover, payments— would often offer or be asked for bribes to receive their ­ drawing lessons from the earlier failed attempt, the MoF and FCGO with knock-on effects on the quality and quantity on what suppli- as the core CFAs, designed a ‘lighter version’ with regard to central ers would deliver, given their expectation of having to set aside controls over payments. Thus, the preaudit principle, which had provisions for bribe payments, and for covering extended periods been part of the previous attempt, was removed, that is, spending of waiting to receive payments. Breaking out of this negative equi- agencies were able to retain sole approval powers within allocated librium was hence quite crucial as part of the initial reform thrust budget lines. District Treasury Comptroller Offices (DTCOs) under to restore basic fiscal processes. the FCGO would simply issue checks upon receiving approved pay- At the same time, the government did not try to close all accounts ment requests from spending agency. In addition, the downstream at once. In particular, despite a limited degree of decentralization, systems (the District Expenditure Control System (DECS) and the local governments were allowed to maintain their accounts with FMIS) were programmed only to the Government Finance Statis- commercial banks. The same was the case for the so-called ‘Legal tics (GFS) level in an effort to leave some discretion and win buy-in Entities of Public Law’—a type of quasi-autonomous and quasi- from district chiefs and spending agencies. CFAs, with external commercial public entity. Many ministries have one or multiple support, also provided extensive training and workshop discus- LEPLs, as the administrative unit with specific responsibilities, sions on the TSA and the FMIS. such as internal academies; large IT departments within ministries Given the political instability in Nepal at the time, strong politi- also have this status. cal backing was not available; while there was some incentive related The rules for local governments and for LEPLs were then gradu- to relative aid dependency, an effective reform team in the Finan- ally tightened over time. Since 2015, a number of LEPLs, as well cial Comptroller General’s Office, and external resources provided to as all local governments are required to execute their payments enable the reform effort. Continuous and very short-term changes in through the e-Treasury system drawing on their subaccounts in the the government rendered opposition from central level line minis- TSA. This more than doubled the number of organizations using tries less strong—as ministers typically held their positions for less the TSA from 415 to 1015, with about 3,000 individual users than a year. The previous unsuccessful experience was utilized, in able to access the system. LEPLs and municipal governments terms of taking a deliberately lighter approach. A key active driver that were brought into the TSA were still allowed to keep excess was the reform team that had been pulled together in a so called funds in accounts with commercial banks and to retain the interest ‘PEFA secretariat’ under the FCGO, which had been set up follow- earned on balances in these accounts. However, they can no longer ing the first PEFA assessment in 2008. With resources provided make any payments directly from these accounts. To use them, by development partners, the reform unit was able to invest in a they must first be transferred to the respective subaccount in the large-scale information and training effort, targeting local lead- TSA and then pay from there. ers across the 75 districts to win buy-in and in training treasury Thus, in Georgia, implementing a TSA was pursued in a step-by- staff on the TSA (rolled out jointly with an FMIS as discussed step approach over a period of about 10 years. It was approached in further below). The treasury worked hard to ensure that it would a way that initially limited the complexity and the number of users issue checks promptly upon submission of payment requests from involved and also included finding ways to respect the relative spending units to win credibility and trust that cash management autonomy of local governments and to maintain some leeway for was on a better footing. MDAs. In parallel, the effort was enabled by the fact that revenues Good progress was achieved in terms of the overall thrust of were rapidly increased, and thus there was a credible commitment the reform, while sustainability and durable institutionalization still by the MoF that spending ministries and agencies could access remain as challenges. Good progress was made with regard to their budgeted funds on time and when needed. closing accounts in commercial banks (involving about 14,000 Nepal followed a route somewhat similar to Georgia in terms of accounts) and having a much reduced number of accounts with a gradual rollout of a TSA, but the reform still remains more partial. the Central Bank (445). Revenue paid by taxpayers to accounts Following an earlier attempt made in the late 1990s which had in commercial banks are largely transferred daily to TSA accounts The World BAnk  61 in the Central Bank.110 Through this, the central government has All MDAs, including those responsible for revenue collection, gained a daily view of the overall cash available, compared to were ordered to use the TSA; and to use any accounts with commercial uploads from districts only on a monthly basis previously; and uses banks strictly as transit accounts, with all balances being swept to the this to steer overall budget execution. Transparency has also been Central Bank-based TSA at the end of each day. Specific guidance increased through publication of a ‘daily budgetary status’.111 was issued by the Office of the Accountant General, which also Trust between the treasury and spending units improved in terms established a help desk for MDAs. A ‘Remita e-Collection Plat- of providing funding as needed on the one hand, and reporting form’ was rolled out for revenue collection. According to published accurately and timely from spending units to the treasury on the reports, at the start of this renewed push, public sector balances other hand. Key weaknesses are that the system still needs to have held in commercial banks amounted to NGN 2.2 trillion (approxi- a more full-fledged regulatory basis; and sustainability beyond the mately US$11 billion) in the second quarter of 2015 (about period for which external funding is available remains to be seen. 2.3 percent of annual GDP).116 By September 2015, around 600 Furthermore, as covered in Chapter 3, TSA and FMIS implementa- out of 900 federal MDAs were reported to be in compliance with tion proceeded alongside increasing fiscal revenues, thus facili- the new regulations; and by November 2015, NGN 1.5 trillion tating credibility. Some sustainability has been proven, as it was (approximately US$7.5 billion) had been transferred into the TSA. possible to rapidly resurrect the system and the IT infrastructure By March 2016, the IMF reported the TSA rollout as completed at following the devastating 2015 earthquake.112 the federal level, including for main revenue entities. Thus, in the In Nigeria, the first round of efforts at establishing a TSA context of a fiscal crisis and a newly elected government, overcom- remained at a halfway stage, but in a more haphazard way. Efforts at ing previous resistance from MDAs suddenly became possible— establishing a TSA were initiated in 2010. Actual rollout started in with very explicit backing from the President for this to happen 2012, with initially 93 federal MDAs beginning to use the Budget (see also Section 4.2). Execution System of the GIFMIS, increasing to about 230 by the However, the renewed effort to roll out the TSA in Nigeria end of 2013.113 However, many of the accounts held by these became an unusually publicly contested case of a PFM reform, in MDAs were not actually closed, and the amount of public funds the context of wider growing contestation. The renewed TSA efforts held in commercial bank accounts actually increased in 2013.114 were significantly covered in the press; with emphasis given to the Also, the TSA did not cover accounts of revenue-related agencies, issue that commercial banks would lose from this change. Press in particular the Federal Inland Revenue Service and the Customs coverage also focused on controversies regarding the setting and Service.115 Given a wider context of a government focused on pre- allocation of fee payments established for banks and IT compa- serving the status quo (as discussed in Chapter 3), and a frag- nies involved, in particular to the company providing the e-Remita mented setup of central finance functions, it was not possible to system.117 The contract pertaining to e-Remita specified a 1 per- go further. cent transaction fee—which in principle would have amounted to The situation changed following the 2015 elections and the around US$100 million—instead of a fixed fee per transaction, brewing fiscal crisis. The renewed emphasis on the TSA was trig- to be charged to transferring MDAs and to be distributed between gered by the fiscal constraints emerging with the oil price drop the company that had designed e-Remita (Systemspecs), commer- that started in the second half of 2014 on the one hand, as well as cial banks (40 percent), and the Central Bank (10 percent).118 As by a motivation of the new government to reduce opportunities for stakeholders began to raise red flags, the terms of the contract leakage on the other hand. The incoming government was elected were adjusted in October 2015.119 The issues raised with regard to on a platform of anticorruption, as well as of being better at get- ting things done, and hence had a dual motivation to seek closer control over resources in a context where there have long been 116. http://www.systemspecs.com.ng/understanding-the-treasury-single- account-tsa-system-things-you-should-know/ concerns about a lack of transparency concerning revenue and 117. The fee is set at 1 percent of TSA transactions, to be divided potential large-scale losses. between SystemSpec (the company providing the e-Remita platform), the commercial banks operating transit accounts used by MDAs, and the Central Bank of Nigeria. See: https://en.wikipedia.org/wiki/Treasury_ single_account#cite_note-15 and http://www.vanguardngr.com/2015/11/ 110. See 2015 PEFA report, ratings and assessment for PI-15. tsa-1-transaction-charges-is-revenue-for-remita-cbn-banks/. See the 111. Available at: www.fcgo.gov.np/publications. Uploads remained details of a senate hearing http://www.elombah.com/index.php/special- current as of October 2016. reports/5624-the-full-senate-report-on-the-abuse-of-the-treasury-single- 112. See http://mdtfpfm.org.np/uploads/files/document/Nepal_PFM_ account. MDTF_Progress_Report_July-_Dec_2015_82.pdf. The progress report 118. A senate hearing organized in March 2016 shed light on some of covers the business continuity achieved for the TSA. the dubious ways in which this contract had been made by the Central 113. Implementation Completion and Results Report for P088150. Bank. 114. IMF CR 16102, 2016; p. 23. http://www.imf.org/external/pubs/ft/ 119. The contestation even led to the creation of a Wikipedia entry under scr/2016/cr16102.pdf ‘TSA’ exclusively addressing the Nigerian issues and perspectives—see: 115. See the IMF’s Fiscal Affairs Department report, January 2013. https://en.wikipedia.org/wiki/Treasury_single_account. 62  Political Economy of Public Financial Management Reforms the TSA were linked to wider resistance against the reforms being and partial control over these accounts, and a float continues. In pursued by the new administration. By spring 2016, there was a late 2013, the government announced that it would accelerate Corruption’— full-scale campaign around the slogan ‘Bring Back ­ the TSA implementation process for public expenditures starting against the governance reforms being introduced.120 in early 2014122 to improve its ability for short-term cash man- In Tanzania, discussions about establishing a TSA had been agement and reduce short term borrowing needs and costs; but ongoing for a number of years, but without real traction. In earlier with limited effect. The government managed to close a number years, a number of accounts were closed, but as the system was of dormant accounts that had not been used in several years and not firmly entrenched, accounts subsequently again proliferated. to recover their balances.123 In June 2015, the Department of The situation became more difficult again in the mid-2010s as Finance issued a new guideline, limiting which banks MDAs may arrears and cash-management problems re-increased, despite the use to maintain balances.124 This was followed in September early introduction of an FMIS discussed in the following section. 2015 by another directive, giving MDAs, SOEs, and local govern- Following the 2015 elections and facing a constrained fis- ments a one year transition period to transfer their deposits to the cal situation, the government made a push toward implementing a government designated banks, considered to be more ‘secure’.125 TSA. As in the case of Georgia and Nepal, this was pursued in Ultimately, the successive directives reflect the fact that the a somewhat incremental way. In early 2016, total balances by CFAs, and the Bureau of the Treasury in particular, were not able MDAs in commercial bank accounts were estimated at TZS 600 to ‘impose’ a TSA on other stakeholders. One of the reasons why billion or around 0.6 percent of Tanzania’s 2015 GDP, that is, sig- resistance in the Philippines seems to have been too strong to nificantly less than in the case of Nigeria, but still sizable. Simi- overcome, is that the government sought to roll out the TSA to lar to Nigeria, there was substantial resistance from commercial all types of entities, that is, MDAs, local governments, as well banks, while the move to a TSA had direct presidential backing.121 Government-Owned and Controlled Corporations (GOCCs) at as ­ Initially, national MDAs and regional administrative secretariats once—in a large and complex country. Such a large-scale attempt were required to move their funds to Bank of Tanzania (BoT) seems to have triggered more resistance than is likely to have been accounts, by the end of January 2016. They were permitted to encountered using a more incremental approach. Moreover, there maintain commercial accounts to meet monthly operating expen- is a close institutional and functional link between establishing a ditures; however, the payment of suppliers of goods and services TSA and an FMIS and significant problems emerged with estab- was moved to be directly through a Tanzania Interbank Settlement lishing the latter, as discussed in the following section. System (TISS). Local governments are expected to move their cash The efforts at introducing TSAs and their varied success are balances to accounts with the BoT from June 2017. a vivid illustration that seemingly highly technical reforms intersect In the Philippines, introducing a TSA had been proposed as one with a number of political and economic interests. Anticipating such of the core elements of the PFM Reform Roadmap to be achieved by interests and identifying options for addressing them has been 2015, but de facto progress remained limited. A TSA was to be rolled crucial for making progress in Georgia and in Nepal. In the latter, out jointly with a comprehensive GIFMIS by 2015. The intention the second attempt at establishing a TSA illustrates a deliberately was to cover all spending agencies, including subnational levels. gradual and partial approach to reforms—which inevitably carries However, by the end of 2016, the system as established since limitations in terms of impact. In Nigeria, the reform process was 2011 covers only revenues. Revenues are paid into one of 19 particularly messy, with fiscal duress and explicit political backing commercial banks designated for revenue collection, and are then eventually emerging as crucial drivers. The incentive for MDAs of transferred daily to the Government’s Common Fund, controlled by more credible cash releases still remains missing due to the fiscal the Bureau of the Treasury and held in the Central Bank. Banks crisis, creating a continuing threat to sustainability. had previously held paid-in taxes and other revenue for several days before transmitting them. To compensate them for the loss of this short-term liquidity, the government switched to paying a fee. On the expenditure side, MDAs continue to use accounts at one of three designated commercial banks. The BTr has only limited 122. See http://business.inquirer.net/153869/single-bank- account-for-govt-launched, and http://www.philstar.com/ business/2014/01/04/1274884/govt-hastens-implementation-treasury- single-account. 123. http://malaya.com.ph/business-news/business/treasury-single- 120. The campaign implied that a rising cost of living was related to account-dbm-flushes-out-p438m-dormant-accounts the anticorruption efforts—rather than to the weakening of the Nigerian 124. http://www.dof.gov.ph/index.php/department-of-finance-issues- economy due to the decline in oil prices. revised-guidelines-on-authorized-government-depository-banks/ 121. See: http://www.thecitizen.co.tz/News/Banks-to-lose-billions-in-new- 125. http://www.businessmirror.com.ph/agencies-given-one-year-to- govt-directive/1840340-3054886-qygkgaz/index.html. transfer-deposit-bank-accounts-to-gfis/ The World BAnk  63 5.6.2 Introducing Financial Management system that subsequently requires continuous efforts—ongoing licensing, hardware and software maintenance, continuous use by Information Systems (FMISs) and training for staff in MDAs—and repeated rollouts of upgraded The introduction of FMIS has been a central pillar of PFM and wider systems, to remain fully functional and to continue to contribute government reforms over the past two decades. Growing computing to a strengthened PFM system. power on the one hand, and growing complexity of governments From a political economy perspective, a number of potential on the other hand made it attractive and increasingly necessary challenges could be assumed to be relevant. Similar to the issue to capture revenue and expenditure data in databases rather than with implementing a TSA, stakeholders in spending agencies may manually. Additional rationales or expectations are that FMISs will have incentives to resist an FMIS due to the greater central con- allow strengthened control over commitments and payments and trol that it creates. These reasons for resistance may interact with thus help to reinforce fiscal stability and probity (World Bank, forth- basic capacity and IT constraints in terms of ensuring that staff coming). However, as Diamond and Khemani (2005) have pointed and systems are available and accessible for all entities meant to out, FMIS reforms frequently stall. In contrast, Dener, Watkins use the system. Preparation periods for FMIS can be lengthy, and and Dorotinsky (2011) note that a strong majority of projects sup- hence entail risks of the process being derailed due to changes in porting FMIS achieve the installation of operational systems, even key driving stakeholders and agencies (that is, shifts in the win- in low-income environments. In addition to some comparative dow of opportunity). Potential supporting factors are the fact that treatments, there are several case studies (for example, Peterson an increasing number of countries operate some form of FMIS, 2006; Joshi, Srivastava, and Nguyen 2015). However, overall, sys- resulting in a growing body of experiences to learn from (both good tematic evidence and discussion about the introduction of FMIS and not so good), and the pressure from increasing volumes and is still quite limited–in part because this represents a reform with complexity of spending. As for other PFM reform aspects, further multiple technical challenges—related to IT, procurement, project important issues concern whether implementation and impact can management, and other factors (Combaz 2015).126 Beyond the be and remain robust in the context of potential ongoing chal- question of whether an FMIS is successfully installed and rolled lenges, such as stakeholders having (continuing) incentives to cir- out, key questions concern how well they are actually used and cumvent controls and oversight mechanisms being created by the sustained and the overall impact on PFM—which has received FMIS. greater attention and analytic coverage only recently (see Hashim Among the five cases included, Tanzania was the earliest to 2016, 12; and World Bank, forthcoming). introduce an FMIS, starting in the late 1990s. The system was imple- Among different types of PFM reforms, introducing an FMIS mented across national level MDAs and successively also to sub- stands out at being particularly costly and complex—which should national levels of government. As Diamond and Khemani (2005, provide an incentive for ‘trying to get it right’. Roughly, the cost of 14–15) emphasize, the introduction of the system had significant an initial rollout of FMIS-type systems in the five countries cov- political backing at the time and was driven by significant fis- ered has ranged from around US$5 million to US$50 million.127 cal pressures and problems with excessive commitments made by Given variations in comprehensiveness and coverage (for exam- MDAs.129 ple, only one versus multiple levels of government), a variety of The Tanzania experience suggests that with a political as well funding sources (the government as well as various contributions as a fiscal interest, and a reasonable sequenced approach, the imple- from DPs), and modalities, it is difficult to establish actual costs mentation of an FMIS can be relatively smooth, even in a low-income comparably.128 The introduction of FMIS also reflects some of the context. The introduction of the FMIS happened during a relatively wider issues that stakeholders need to consider when seeking to reform-oriented period in government, following the first multiparty introduce large-scale IT solutions, such as the need to develop elections organized in 1995 and the election of Benjamin Mkapa technical specifications that provide a good fit with the require- as President. PFM reforms at the time were focused on bring- ments and processes. Furthermore, similar to other reforms, but ing significant payment arrears under control that had resulted even more so due to continuous evolution in technology, intro- from excess commitments.130 Introducing an FMIS was part of ducing an FMIS is not a one-off reform. Rather, it brings in a the first PFMRP being adopted. After a first attempt with a system designed by a national consultant proved problematic, the gov- 126. http://www.gsdrc.org/wp-content/uploads/2015/06/HDQ1229.pdf. ernment decided to purchase a customized ‘Enterprise Resource GSDRC overview paper with useful references. Planning System’, in this case, EPICOR. Once introduced, the 127. Hashim (2016) provides a range of World Bank funding commitments for IFMISs ranging from US$3 million in Cabo Verde to US$231 million in the Russian Federation. 128. For example, the FMIS in Nigeria covers the federal level only, with 129. http://www.mof.go.tz/index.php?option=com_content&id=34 separate systems being established at the state level. For the four other 130. https://extranet.imf.org/depts/fadbank/TAR/Tanzania_Advancing cases, FMISs have covered both national and subnational levels. Public_Financial_Management_Reforms_October_2012.pdf 64  Political Economy of Public Financial Management Reforms system has been in continuous use, and software updates were system as designed was to cover fiscal planning, budget prepara- undertaken. tion, budget execution and financial reporting comprehensively for However, the Tanzanian experience also highlights that suc- all levels of government. The IT project was tendered in 2013, but cessful FMIS implementation does not ensure good fiscal manage- failed, and then retendered and awarded in 2014. However, the ment, as overriding commitment controls eventually re-emerged. The President subsequently did not approve the GIFMIS when he was FMIS’s primary use has been for the budget management and asked to sign off on the recommendation of the DBM’s Special accounting modules (general ledger, accounts payable, accounts Bids and Awards Committee (SBAC). A key motivation seems to receivable). In parallel to the expenditure reforms of the late have been that the scale of the project was overambitious and the 1990s, the government established the Tanzania Revenue Author- President was concerned that it would fail—leading to a costly ity (TRA) in 1996. It managed to pursue increases in revenue col- waste of funds and a potential drag on the President’s record. It lection up until 2007 (see Chapter 3), but subsequently, revenue thus also reflects a somewhat surprising breakdown between the plateaued. Once revenues no longer increased further relative to technical and the political levels to agree in time on a more man- GDP, payment arrears began to re-emerge (IMF 2012). As budget ageable approach. This was a very significant drawback for the releases again became less reliable, FMIS users in spending agen- PFM reform agenda, given the GIFMIS’s centrality to it. cies reverted to overriding commitment controls in the system. Subsequent to the failed attempt, CFAs agreed to pursue a Thus, one of the commitments of Tanzania under its 2014–2017 much scaled-down version, a Budget and Treasury Management Sys- Policy Support Instrument Agreement with the IMF was a pub- tem (BTMS), to be used initially only by DBM and BTr.134 A tender lic information campaign launched as part of the 2016 budget for this much smaller system was successfully completed in late speech, that payment orders to suppliers would only be honored if 2015, and design started in 2016. The system is intended to generated through the FMIS.131 be initially used only by the two main CFAs, the Department of The Philippines provides the most dramatic counter example Budget and Management and the Bureau of the Treasury, with in terms of seeking, but not succeeding thus far, in implementing an actual use scheduled to start in 2017. While starting out with a FMIS—although this was intended as the centerpiece of PFM reforms limited scope, the system is designed to be expandable and can by the 2010–2016 Government. Following the approval of a PFM be rolled out to the spending agencies in principle in the future reform roadmap in early 2011, a PFM Reform Committee, called (requiring procurement of additional user licenses and training). ‘GIFMIS Committee’, reflecting the system’s central status, was The sharp reduction in the scope, strategies, and costs has made formed under Executive Order No. 55 in 2011 in September.132 progress more manageable, but has also left the overall reform The committee comprised senior representatives (commissioner, path uncertain. chairperson, secretaries, and undersecretaries), from the COA, The Philippines’ experience with the GIFMIS indicates that DBM, DoF, and BTr—that is, it brought together key representa- seeking a comprehensive solution is a particularly high-risk approach, tives of the fragmented CFA agencies involved. The GIFMIS project even when pursued by a highly reform-committed government. PFM aimed at phasing out the many dispersed financial management reform plans gave a central place to the GIFMIS as the centerpiece systems and processes in place through the implementation of of the overall PFM reform process, further increasing risks that if a unified system and with the aim to reduce fiduciary risk and the project failed it would also delay related reform efforts such consolidate reporting. Significant expectations were raised that as the TSA discussed above.135 It also shows the risk of pursuing the GIFMIS would—quasi automatically—deliver solution to key a large-scale reform within a limited window of opportunity. Given shortcomings of the PFM system, such as fragmentation, insuf- the constitutional provision of a single mandate and hence a maxi- ficiently credibility of budget execution, as well as problems with mum of six years duration for a single administration, pursuing timely, reliable, and comprehensive reporting on expenditures.133 a reform that takes several years from concept just to contract- In 2012–2013, a comprehensive conceptual design was devel- ing went against a more robust strategy of seeking smaller reform oped, together with functional and technical specifications. The steps that have a more limited span from inception to starting implementation. The detour from investing in a large-scale design and then eventually moving to a much smaller-scale solution had 131. IMF Country Report 2017/13. a considerable cost in terms of lost time and at least partially lost 132. The formal approval of the PFM Reform Roadmap was made in January 2011 through a Joint Resolution (01-2011) of the COA, DBM, efforts and credibility. and DoF, and then in September 2011, Executive Order No. 55 Directing the Integration and Automation of Government Financial Management System was signed. 133. “Interviews conducted by the review team with a range of government stakeholders revealed inflated expectations as to the nature 134. http://www.gov.ph/2015/12/16/new-it-system-public-fund- of reforms which GIFMIS will automatically deliver.” Department of efficiency/ Foreign Affairs and Trade (DEFAT, Australia) review (2014). 135. See also Bysouth, Philipsen, and Belisario 2014. The World BAnk  65 In Nigeria, implementing a GIFMIS was initiated in 2003, during still remains uncertain. Reaching a successful procurement and a period of relatively strong pursuit of PFM reforms as noted in Chap- subsequent rollout of the GIFMIS was a major achievement in a ter 3. At the time, political backing from President Obasanjo coin- complex and fragmented environment with pervasive corruption cided with a strong technical team on the one hand, and incentive such as Nigeria. The contrast with the Philippines indicates that a to pursue reforms to complete the debt relief process on the other more limited scope likely facilitated progress; and that in this case hand. World Bank support for establishing a GIFMIS commenced a combination of a dedicated technical team and external sup- with the Economic Governance Reform Project approved in 2004, port was able to make a difference, even though with significant with the intention of starting GIFMIS implementation by 2007. delays. However, resistance from line ministries remained signifi- However, as for other aspects of PFM reforms, momentum flagged cant, albeit being somewhat weaker than with regard to the TSA from 2006 onwards as the main driving factors peeled away. The discussed above. As stakeholders have pointed out, one conse- fragmentation of the CFA system, which has parallels to the situ- quence has been that the quality of FMIS use, including the data ation in the Philippines, further hindered progress. In clear con- being entered into the system, remains inadequate, requiring fur- trast to the Philippines, the system was only intended to cover the ther effort.138 In addition, connectivity problems and constraints federal level of government, with states being expected to adopt in the effective troubleshooting of program bottlenecks have made their own systems—being spurred to do so by the example set at the system frustrating for users. Thus, while it has been possible the federal level. to set up a system, thus far it only partially delivers the expected Between 2004 and 2009, further preparation continued but impacts of PFM, due to a combination of incentive as well as at a very slow pace, followed by a drawn-out procurement process, capacity constraints. concluded only in 2011. Among CFA institutions, the Office of In Nepal, efforts have been ongoing to use information technol- the Accountant General held the main responsibility. Eventually, ogy for automating PFM processes and for capturing information; but systems specifications were sufficiently agreed to issue a pro- incrementally, following initial setbacks. The incremental approach curement notice, with the full procurement process to contract taken appears to have been a reasonable choice pursued by mid- signature taking nearly another two years. In line with standard rec- level bureaucrats for a situation in which sustained high-level ommendations, in May 2011, the federal government concluded backing for reforms is not available, given continuous political cri- a contract for a customized off-the-shelf solution, provided by a ses and shifts. Nigerian subsidiary of Hewlett Packard, for a cost of approximately For budget preparation, a first attempt to introduce a Budget US$29 million.136 Shortly after, Jonah Otunla was appointed as Management Information System (BMIS) was made in 2005, but met Accountant General for the federal government, replacing Ibrahim with resistance from line ministries. To deal with this resistance, Dankwambo.137 the Ministry of Finance moved to using the BMIS itself for the Once a contract had been signed, implementation moved for- central compilation of the budget, while creating a ‘Line Ministry ward. Design and testing proceeded rapidly, and the system went Budget System (LMBS)’ for spending ministries. By putting ‘Line live in April 2012 for the Budget Execution System. The reap- Ministry—LM’ in the title to purposely make it unique from its pointment of Ngozi Okonjo Iweala as Minister of Finance in mid- own IT system (BMIS), greater buy-in was facilitated. The MoF 2011 brought further renewed momentum to the GIFMIS effort, also modified the LMBS so that it was similar in format to the given her initial involvement in conceptualizing the reform. The line ministries’ manual budgets to facilitate the change for users. original pilots were 93 MDAs, with coverage increasing to 209 These initiatives ensured the line ministries, in addition to the MDAs by mid-2013, 551 MDAs by early 2015, and to all but a MoF, would be ‘winners’ of the reform. Implementation progressed few agencies by 2016. One marring aspect was that similar to the gradually. By 2015, the LMBS was fully functional and the MoF establishment to the TSA involving some suspected malfeasance, began integration with its own BMIS. in May 2015, two officials from the Office of the Accountant Gen- For capturing data on budget execution, a similar step-by-step eral were arrested by the Economic and Financial Crimes Com- process has been pursued. A unique expenditure system was devel- mission (EFCC) on accusations of having established companies oped and rolled out for the districts (District Expenditure Control to win bids related to the rollout of the GIFMIS (NGN 1 billion). System). The FCGO first piloted the DECS/TSA in three districts While there is a sense of achievement against quite a few odds before rolling it out to the remaining 72. In addition, once the new with regard to implementing the GIFMIS, sustained use and impact expenditure system was institutionalized, the FCGO began the roll- out of a revenue system Revenue Management Information System (RMIS) in 2015. At the central level, the FCGO also established 136. See: http://gifmis.gov.ng/gifmis/index.php/implementation-update. The system is based on Oracle. an FMIS into which it compiles expenditure and some revenue 137. Dankwambo had been Accountant General since 2005. He resigned to run for a governorship which he won in 2011 (in Gombe State), and was reelected in 2015, both times on a PDP ticket. 138. Interviews with key stakeholders. 66  Political Economy of Public Financial Management Reforms information and which it uses to generate reports, including func- and on the other hand, the structure of an LEPL which facilitated tional, administrative, and economic breakdowns (see 2015 PEFA paying IT experts at rates competitive with the private sector.140 assessment). Monthly expenditure outturns for ministries and dis- Overall, introducing some degree of automation of treasury tricts are now posted on the ministry’s website. However, data still transactions and of moving to electronic databases for recording reve- needs to be transferred from other systems into the central FMIS nue and expenditure flows seems to be feasible in a range of different to do so, opening potential space for errors; and the coverage still country contexts. Early adoption proved possible in Tanzania, that remains incomplete (for example, it does not comprise debt ser- is, a low-income environment, and eventually progress was made, vice). The strengthened capturing of budget execution data has even if involving substantial delays, in Nigeria and the Philippines. also enabled the establishment of a mid-year budget review pro- Availability of support from DPs has played a role in facilitating cess led by the MoF. financing and technical approaches. In Georgia, the introduction of an FMIS was initiated jointly with However, introducing more ‘integrated’ systems has proven the treasury reforms started in 2004. With support from a World more challenging than introducing more limited and partial systems— Bank project approved in 2006, the original idea was to imple- for technical as well as nontechnical reasons. As experienced in ment a “Customized off the shelf solution (COTS),” to be in place Nigeria and the Philippines purchasing large-scale IT solutions by 2010. The MoF team invited several international companies can involve procurement problems, including risks of corruption. providing relevant programs to present the solutions and custom- The larger-scale and the more complex the system, the lengthier ization they would offer. After being disappointed by the level of the process of preparing specifications, actual tendering, and con- customization being offered, a decision was then taken to develop cluding a contract can become. This in turn risks extending such a system nationally instead.139 This was subsequently done ‘in- processes beyond available windows of opportunity for reforms.141 house’, specifically through hiring a significant number of IT spe- Moreover, as the Tanzania experience demonstrates, suc- cialists into the ministry, rather than outsourcing to a private local cessful introduction of an FMIS is not a guarantee for effective and provider. efficient use of public resources or even more effective control of The creation of the system was done in several stages, start- commitments and prevention of payment arrears. Prevention of pay- ing with a web-based e-Treasury system, then moving to an e-Budget ment arrears and of delays in cash releases is largely dependent system for budget planning and appropriations. At later stages, an on a wider context in which central institutions of the government e-procurement system was developed and eventually integrated; are able to keep spending commitments in line with revenue or and modules for tax and for debt were added. As of 2015, the are able to increase revenue in line with increases in spending system still remained cash-based (that is, recording cash transac- commitments. Having an automated system that has embedded tions rather than accruals), with a plan to transition to accruals commitment controls can be useful, especially to flag aberrant over time (see the following section). The rollout to subnational spending requests to central agencies. However, decisions about levels and to LEPLs took place gradually, alongside the rollout of overspending on certain categories and by certain ministries are the TSA described above. typically more political than technical, and so cannot be effec- For Georgia, the path taken proved viable and avoided costly tively prevented by IT mechanisms. detours as experienced in the Philippines, or the creation of an ini- An implication from a nontechnical perspective for the planning tially fragmented approach as in Nepal, facilitated by several factors. and implementation of FMIS is to consider ‘scalable incrementalism’ in A fundamental one was the general strong commitment to pursu- particular in more complex environments. A deliberately sequenced ing reforms as discussed in Chapter 3. This also created an impe- approach may be more feasible in many situations than a ‘big tus toward seeking fast progress on core problems, which seems to also have fostered a preference for starting the development of 140. As Hashim and Piatti (2017, 53) note, maintaining the an in-house IT-solution, as opposed to beginning the establish- functionality of FMIS depends on technically capable staff, including IT ment of a comprehensive FMIS with a lengthy procurement phase. specialists, and this can be challenging if regular civil service pay scales are low. In terms of capacity, there were two enabling factors: on the one 141. A further key risk is dependency on the vendor: once the hand, an availability of qualified IT experts that could be recruited, procurement decision is made in favor of a particular vendor, the government locks itself into working with a particular system (unless it is willing to re-invest substantially in a different system). Competition among different subsidiaries for widely used systems such as Oracle or SAP partially mitigate this risk; while it may be particularly high for a system sourced from a local commercial vendor (as in the case of the 139. The three invited companies were SAP, Oracle, and Freebalance. In- e-Remita system introduced in Nigeria; or also for the FMIS introduced house development has also been done by a number of countries in Latin in Burundi—not covered here). If vendors are not providing sufficient America. See http://blog-pfm.imf.org/pfmblog/2013/05/fmis-choice-the- support to troubleshoot the system, users will become less inclined to dangers-of-in-house-development-in-low-capacity-countries.html. report regularly and adequately. The World BAnk  67 bang’ approach on the one hand; while it would also be preferable aware of reforms in this area, even more so than is the case for to a more fragmented approach, as it emerged in Nepal, on the other aspects of PFM reforms. other hand. For a highly committed government, a scalable incre- An international standard has emerged in recent years and has mental approach offers the potential benefit of being able to have triggered widespread efforts at adoption across non-OECD countries, a core system in place relatively quickly; while in an environment including the five cases reviewed here. The International Pub- with more limited commitment, taking such an approach may still lic Sector Accounting Standards (IPSAS) have been endorsed be possible to pursue by key public sector managers within CFAs. by the IPSAS Board, which is organized under the International Moreover, it is critical to pay greater attention to how these sys- Federation of Accountants (IFAC), with funding from the World tems once in place actually function and to identify key weaknesses. Bank Group, the Asian Development Bank, and others. The stan- It is very important not to create a false sense that having rolled dards are based on the International Financial Reporting Stan- out an IFMIS means that better fiscal controls and more accurate dards, developed for private sector accounting, and issued by the reporting are in place, when this is in fact not the case. Budget International Accounting Standards Board (IASB).142 The IPSAS execution can remain politicized, with or without automation, in Board started functioning in 1997 and issued the first standards situations where certain ministries are allowed to overspend and in 2000. It has since added additional standards on various areas others are not. The power of an MoF to limit and monitor the (38 standards in total for accrual-based statements and 1 stan- expenditures of the presidency or of the security apparatus, as well dard for cash-based statements), designed to meet the needs of as of otherwise politically influential ministries and agencies is public sector accounting. frequently constrained, even after an IFMIS has been established. The IPSAS Board has issued both accrual-based standards and Constraining overspending from these ministries is ultimately a standards for cash-based public sector accounting. The cash-based political compromise. DPs can make important contributions to IPSAS encourage additional voluntary inclusion of accrual-based ensuring that there is clarity on how well systems are actually information in the principally cash-based financial statements. used, and to avoid reinforcing a sense that ‘reforms are done’ once Given that the IPSAS Board is an international professional body, an IFMIS is in place. not a government or an international organization, its standards do Creating a realistic sense of what has been achieved and why not have a legally-binding status as such; each country ultimately limitations persist is also necessary to assess what further improve- has to adopt its legally binding standards. The board encour- ments can happen and how. The study by Hashim and Piatti (2017) ages that financial statements prepared include a disclosure of underlines that ‘benefits from an FMIS will accrue only if the con- the extent to which IPSAS standards are being complied with; for trol protocols that come with it are diligently applied’ and that a example, as typically there is an extended transition period until ‘credible budget is critical’. Both of these conditions do not hold in financial statements would comply fully with all accrual-based several of the cases discussed here. It is important to be realistic IPSAS standards. about what benefits an FMIS delivers in such situations, and when The key principle is the idea that expenditures should be and how such weaknesses can be addressed, for example, during recorded as an obligation when they are committed, rather than at the ‘windows of opportunity’ that may arise. Ignoring these challenges point when payments are actually made. A further requirement of or a lack of analyzing limitations would be risky and reinforce accrual-based accounts is the recording of assets and liabilities, the tendency of rushing from one reform area to the next, while and recording the depreciation on assets with finite lives (see also neglecting to ensure that what is being changed actually makes a Moretti, Cavanagh, and Flynn 2016 and Khan and Mayes 2009). difference. Particularly important is the coverage of liabilities, including pen- sion obligations and similar liabilities, which are typically not fully 5.6.3 Reforming Accounting and Fiscal captured in cash accounting systems, and potentially also asset transfers. The main challenge is that accrual accounts require Reporting Standards greater capacity and a more specialized and trained accounting Accounting standards have received considerable attention as an cadre and that accrual accounts can also be more difficult to area of reform ‘inside’ governments. The ultimate aim of reform- interpret. ing accounting standards is to improve the quality of financial In particular, the IMF has played a crucial role with regard to information and to enable better decisions on resource allocations assessing the reliability in applying accounting standards, as part of and risk management. At the same time, accounting standards its monitoring of fiscal risks. The fund’s overall concern is to reduce are a highly technical area; so, many potential stakeholders of the risks of unexpected/unknown government liabilities—originally reform are not fully aware of the details, options and choices to make and their respective implications. Wider stakeholders, such as citizens in general or organized civil society, may not even be 142. See www.ipsasb.org and www.ifac.org. 68  Political Economy of Public Financial Management Reforms Definitions: Accounting, Fiscal Reporting, Transparency, BOX 5.1 and Risk Apart from accounting standards, other international state of the public finances to citizens in the form of fis- standards have been issued for fiscal reporting mainly cal forecasts (in fiscal strategy or budget documents), as a way to enable comparability. These are the Gov- government finance statistics (fiscal reports pro- ernment Finance Statistics (GFS) standards, issued by the duced in accordance with statistical standards), or gov- IMF. There have been three versions of GFS standards, ernment financial statements or accounts (fiscal reports each set out in a manual (GFS Manual [GFSM]): 1986, produced in accordance with accounting standards). 2001, and 2014.143 The main purpose of the GFS is to • Fiscal transparency refers to the clarity, reliability, fre- enable macro-fiscal analysis. Adherence to GFS enables quency, timeliness, and relevance of public fiscal report- international comparisons across countries by standardiz- ing and the openness to the public of the government’s ing, for example, the functional breakdown of public expen- fiscal policy-making process. Within this, clarity refers to ditures, and the overall scope of the ‘general government the ease with which these reports can be understood sector’ and the ‘public sector’. GFSM 2014 considers the by users, reliability refers to the extent to which these standards established through the issuing of the IPSAS reports reflect the government’s true financial position, since 2000, as well as other updated standards, such as frequency (or periodicity) refers to the regularity with those on reporting on public debt and for national accounts. which reports are published, timeliness refers to the • Government accounting refers to the concepts, stan- time lag involved in the dissemination of these reports, dards, rules, and systems used to generate the financial relevance refers to the extent to which these reports pro- information used in fiscal reporting. vide users with the information they need to make effec- tive decisions, and openness refers to the ease with • Fiscal reporting refers to the production of summary which the public can understand, influence, and hold information about the past, present, and future state of governments to account for their fiscal policy decision. the public finances for both internal (management) and external (accountability) uses. • Fiscal risks are factors that lead to differences between a government’s forecast and actual fiscal position. • Public fiscal reporting refers to the publication and dissemination of this summary information about the Source: IMF (2012). 143 triggered by the Asian financial crisis of the 1990s. The main in 2007, followed by a new Fiscal Transparency Code issued in initial tool for this purpose have been Reports on Observance of 2014.145 Fiscal Transparency Evaluation (FTE) Reports were Standards and Codes (ROSCs) (covering a total of 12 specific established since then as a new tool to assess the accuracy of fis- areas); typically as specific reports on fiscal transparency.144 The cal information. IMF also issued a Code of Good Practices on Fiscal Transparency As reflected in Table 5.4, fiscal ROSCs have been relatively few, as have been Fiscal Transparency Reports thus far. As noted by the IMF in 2012, efforts in this area had waned since the 143. For GFSM 1986 see: https://www.imf.org/external/pubs/ft/gfs/ early 2000s.146 The fiscal crises experienced by several Euro- manual/1986/eng/. For GFSM 2001 see: https://www.imf.org/external/ pubs/ft/gfs/manual/pdf/all.pdf. For the most recent manual see: https:// pean countries—which especially in Greece had been obscured www.imf.org/external/Pubs/FT/GFS/Manual/2014/gfsfinal.pdf. The general by inadequate fiscal accounting and reporting—triggered renewed government sector consists of resident institutional units that fulfill the attention. The FTE reports cover accounting standards and respon- function of government as their primary activity. This sector includes all government units and all nonmarket, nonprofit institutional units (NPIs) sibilities for their application and oversight, as well as issues of that are controlled by government units. The public sector consists of all comprehensiveness. Again, however, only a few such reports have resident institutional units controlled directly, or indirectly, by resident government units—that is, all units of the general government sector been produced to date—among the five case studies, only for the and resident public corporations. That is, the public sector is defined Philippines. as the ‘general government’ plus public sector enterprises. GFSM 2014 takes into account the standards established through the issuing of the IPSAS since 2000, as well as other updated standards, such as those on 145. http://www.imf.org/external/np/fad/trans/ reporting on public debt. 146. See: http://www.imf.org/external/np/pp/eng/2012/080712.pdf— 144. There are also other types of ROSCs. Fiscal Transparency, Accountability, and Risk. The World BAnk  69 While there are other assessments covering fiscal transpar- seek to hide some form of misappropriation of funds (for example, as a ency and accountability—notably PEFA assessments and the Inter- source for funding election campaigns or additional security expendi- national Budget Partnership’s Open Budget Survey—neither of these tures). In principle, external audit offices would take on the role to goes as deeply into the technical detail of assessing the application assure that accounting standards are accurately applied; however, of accounting and fiscal reporting standards. The 2016 guidance for both their capacity in terms of auditing skills or operating budgets PEFA assessments has one sub-indicator on data integrity of finan- to carry out audits, as well as their independence may also be cial data (PI-27.4), focusing on access to information (read-only constrained. versus changes to records by creation and modification); and exis- Because there are potential nontechnical and incentive issues, tence of a body, unit, or team in charge of verifying data integrity. there are some risks that efforts to adopt (accrual or cash-based) PEFA indicator 29.3 focuses on accounting standards, but only IPSAS have an element of ‘isomorphic mimicry’. Stakeholders within assesses whether international or consistent national accounting the government may have some incentives in this regard; as well standards are being used (with a higher rating for adopting inter- as those involved in the setting and dissemination of international national standards). standards. For technical-level stakeholders within the govern- From an overall concern about government accountability, ment, initiating reform efforts toward adopting IPSAS can involve seeking to adopt internationally accepted standards for government a mix of needing to or wanting to comply with IMF requirements accounting and reporting are positive steps. The rationale for adopt- (for example, for a new program), get involved in various interna- ing standards is that they are based on careful consideration and tional conferences, and so on. Given that accrual accounts are inputs—that is, they distill accumulated knowledge and experi- more demanding to interpret, there may be an expectation that ence. They can also provide legitimacy. In principle, adherence to stakeholders outside of the government would not notice any ‘cre- such standards should reduce the scope for intentional or uninten- ative practices’ used, even if in principle accrual accounts should tional hidden risks; and better fiscal reporting is a public good for be more comprehensive and hence improve transparency of the stakeholders from local to national and to global levels. Using a government actions.149 Thus, the actual quality of implementa- unified set of standards, if applied correctly, should enable better tion should receive considerable attention to ensure that the cross-country comparison of fiscal allocations to different types of expected benefits in terms of more comprehensive accounts and activities and expenditures. greater transparency, comparability, and accountability are indeed At the same time, the highly specialized nature of these reforms realized. may obscure some of the nontechnical and incentive issues involved. Across the five countries reviewed for this study, reform inten- First, there are costs and potentially capacity constraints involved tions and the status are as follows. All five countries are seeking to in arriving at a full statement of government financial and non- introduce IPSAS. For Nepal, it has been most clearly stated that financial assets and liabilities (both ‘start up’ costs to arrive at the intention is to initially move to compliance with cash-based initial valuations, and subsequent costs of continuous application, IPSAS, rather than the accrual-based version. At the same time, all including substantial needs for training or hiring qualified staff), five countries are still at a preliminary stage of this reform effort. and second, governments and specific stakeholders may still have Looking at recent PEFA assessments, for two out of the five incentives to not fully report liabilities for various reasons (such as countries, the Philippines and Tanzania, the accounting standards avoiding a formal breach of debt ceilings),147 or to continue but used have been improved to a greater extent than the quality of 1 4 8 Table 5.4.  Dates of Fiscal ROSCs and Fiscal Transparency Reports Georgia Nepal Nigeria Philippines Tanzania Published fiscal ROSC 2003 2007 — 2002, 2004 2002 Fiscal transparency reports — — — 2015 — Source: IMF.148 147. Adopting IPSAS may also allow some countries to claim a lower debt burden, as it includes provisions for discounting, while ‘traditional’ public sector accounting calculates debt to GDP ratios by looking at the face value of the debt relative to current GDP. This may both be a benefit and a curse (as risks of accelerating debt are always a significant risk). For example, an investor in Greek debt has claimed that it is far lower 149. A recent analysis conducted interviews with financial managers at than headline figures if calculated consistent with IPSAS standards local levels in England where IPSAS had been introduced. While they rather than the currently applied ‘Maastricht’ definition of debt. found impacts on accountability to be strong in general, they did not 148. https://www.imf.org/external/NP/rosc/rosc.aspx?sortBy=CountryName agree that IPSAS could prevent ‘creative accounting’ practices. http:// &sortVal=N; and http://www.imf.org/external/np/fad/trans/ etheses.bham.ac.uk/4692/1/Mangualde13MRes.pdf. 70  Political Economy of Public Financial Management Reforms Table 5.5.  PEFA Ratings for Accounting Standards Used and the Quality of In-year Budget Reports Georgia Nepal Nigeriaa Philippines Tanzania 2008 2013 2008 2015 2012 2010 2016 2009 2013 P 24, dimension (iii) [P 28.3] A A C B C D C Cc C Quality of information of in-year budget reportsb P 25, dimension (iii) [P29.3] B Cd C C C B A C B Accounting standards used Source: PEFA reports; pre-2016 framework. Note: a. Data for Nigeria are based on an unofficial assessment carried out in 2012. b. 2016 PEFA format: accuracy of in-year budget reports; c. both indicators had not been rated in the 2006 PEFA assessment for Tanzania; d. the 2013 PEFA assessment for Georgia states that ‘C’ is the correct score and judged the score given in 2008 as too high, based on a misunderstanding of cash-based IPSAS already being in use. information reflected in in-year budget reports(Table 5.5). In Geor- financial statement, comprising assets and liabilities related to gia and Nepal, in contrast, the quality of information of in-year LEPLs and SOEs. The general commitment was revisited in the budget reports is judged as good or very good, while accounting PFM Action Plan for 2015. The aim is to achieve full compliance standards remain national rather than conforming to international with accrual-based IPSAS by 2020 (PFM Management Reform standards (as required for a B rating).150 Strategy 2014–2017, p. 10). The 2014–2017 Reform Strategy For Georgia, reforms to accounting standards and practices provides a more detailed outline of which IPSAS standards are received rather little attention during the initial post-independence to be adopted in what year (for example, treatment of changes period (that is, until 2003), and also during the initial period follow- in currency exchange rates, of conditional liabilities and assets, ing the Rose Revolution and the commencement of other public sector and so on), with an overall goal of full introduction by 2020. The reforms. Accounting standards were cash-based, focused on rev- reform will also involve the addition of a full-fledged accrual-based enue and expenditure flows, and continued to follow legacies from General Ledger module to the suite of IT-systems for the treasury. the previous system in which public sector accounting staff had Initial progress was made by the end of 2015, including the been trained. This approach was initially sufficient while reforms translation of all the standards into Georgian, as well as initiating focused on ‘big picture’ fiscal risks and opportunities, such as training on the first set of standards to be adopted. At the same time, strengthening the functioning of the treasury and tax policy and a lot of ground remains to be covered in terms of rolling out train- administration. However, the reforms as undertaken enabled in- ing on IPSAS standards, undertaking actual valuation of assets year budget reporting that was judged as reliable already by the and liabilities, as well as of creating the accrual-based General first PEFA assessment undertaken in 2008, within the parameters Ledger module, and making it operational. The country’s limited of the standards used. size generally should facilitate progress, while full completion The focus on accounting standards increased from the late may still be delayed by some years. The real ‘proof’—as for other 2000s, when the international influence on the PFM reform agenda ­ countries—will then be subsequently, with regard to assessing became stronger(following the conflict with Russia and subsequent whether the accrual-based financial statements are adequate and increased external support). As set out initially in the 2009 PFM whether any significant public sector flows, assets, or liabilities Strategic Vision, the intention is to go through a process of moving remain uncovered or not accurately reflected. from cash-based, to modified cash-based, to accrual-based IPSAS For Nepal, the initial intention to reform public sector account- compliant accounting standards. The same strategy also sets out ing and reporting was set out in the 2009–2012 PFM Reform Program the aim of developing a GFS (2001) compliant budget classifica- (PFMRP Phase I), which was developed following the initial 2008 tion and chart of accounts and to develop a consolidated annual PEFA assessment.151 The intention set out in the plan was to improve the rating for the indicator related to accounting standards to a ‘B’ on the PEFA scale; and specifically to develop, adopt, and 150. A ‘C’ rating requires that there is a national standard that is codified, disclosed, and consistently followed. Moving to a ‘B’ rating introduce Nepali Public Sector Accounting Standards (NPSAS). requires that “the majority of international standards have been The next PFM Reform Strategy/Program (PFMRP Phase II) covers incorporated into the national standards. Variations between international 2016 to 2025. The NPSAS were formally adopted in 2009, and and national standards are disclosed and any gaps are explained. The standards used in preparing annual financial reports are disclosed.” implementation started for FY2009/2010. However, one key dif- (2016 PEFA guidance; the sub-indicator related to accounting standards ficulty noted was that there was a lack of full compliance with new is now PI-29.3, as additional indicators have been added to the frameworks). It needs to be noted that the assessment for PI-24 (iii) relies on whether responsible agencies (internal audit, external audit) have concerns regarding the quality of reporting, and therefore as noted 151. See: http://www.capa.com.my/images/capa/ICAN-CAPA_PSAAconf_ above is not a direct assessment. RameshworKhanal.pdf. The World BAnk  71 standards, including due to capacity constraints of accountants government and other interest groups.155 Its seven directorates across ministries, agencies, local governments, and SOEs. Only include one for public sector accounting standards. two pilot ministries, and 14 other central level bodies were compli- The intention of IPSAS implementation in Nigeria is very ambi- ant by 2012 (Ministry of Physical Infrastructure and Transport and tious insofar as it is intended to be implemented across all three Ministry of Women and Social Welfare) (PFMRP Phase II, p. 5). levels of government—federal, state, and local —in a federal sys- For a further 14 line ministries, training was organized, and use of tem in which subnational levels have guarded their relatively fis- NPSAS started in FY2015/16.152 cal autonomy quite closely. A subcommittee of the Federation As for other budget execution reforms discussed above, the Account Allocation Committee (FAAC) was charged in mid-2011 main ‘owner’ of the NPSAS reform efforts has been the Financial with developing a roadmap and supporting the rollout of IPSAS. Comptroller General’s Office (FCGO) which also houses the overall Members of the FAAC Subcommittee include the Office of the PFM reform secretariat. The Accounting Standards Board (that is, Accountant General of the Federation (OAGF) as well as state-level professional association) has supported the development of the Accountants General (one from each of Nigeria’s six ‘geopolitical NPSAS. Achievements were still limited by the time the next PEFA regions’). Also established was a federal-level IPSAS Implementa- assessment was carried out in 2014 (published in 2015), with the tion Committee and an Implementation Committee for state and overall rating for PI-24 (Quality and Timeliness of in-year-budget local levels. An initial implementation date was set for 2012, later reports) as well as for PI-25 (Quality and Timeliness of Annual pushed to 2014 for cash-based IPSAS and 2016 for adopting Financial Statements) remaining at C+; with no change yet in the accrual-based IPSAS (Adamu and Ahmed 2014). rating of the particular sub-indicator (PI-25, iii) on ‘Accounting Moreover, the Accountant General156 and the FAAC Subcommit- Standards Used’—the rating remained at ‘C’, given the still very tee chair157 set expectations for the impact of IPSAS high, arguing at limited use of the NPSAS by 2014. a workshop in 2010 that introducing these would minimize corruption In Nigeria, plans to move to IPSAS standards were initiated in in Nigeria, as well as contribute to attracting foreign investment.158 2009; with the Accountant General being the main champion of the In line with the discussions in Section 5.4 above, both of these reform. In this case, the move toward IPSAS for the public sector claims are clearly exaggerations, as no single reform on its own was closely related to efforts at establishing better private sector can bring about such significant changes; improvements in the accounting practices in Nigeria, as a way to attract greater foreign judiciary are just one of several complimentary reforms needed. direct investment. IFRS—that is, the international standards for the Implementation of cash-based IPSAS has started, but there are private sector—were adopted in parallel. The move toward adopt- still challenges with regard to actual application; while the transi- ing IPSAS was politically sanctioned in July 2010 by a decision tion to accrual-based standards remains uncertain despite repeated of the Federal Executive Council (shortly after Jonathan became announcements. The main legitimizing document is the 2010 deci- the President following the passing away of his predecessor in May sion by the Federal Executive Council. Financial regulations (fed- 2010). Subsequently, a bill creating a new Financial Reporting eral), instructions (state level), and memoranda (local level) were Council (FRC) was passed in May 2011 (that is, immediately after harmonized. Subsequently, the OAGF issued a National Treasury the general elections of April 2011).153 The FRC is established Circular in October 2014, requesting all public sector entities as a federal government parastatal under the supervision of the across the three levels of government to set up their own specific Federal Ministry of Industry, Trade, and Investment; and has a IPSAS implementation committees.159 mandate to issue rules and to regulate accounting, actuarial, valu- Part of the challenge is that the degree of automatization ation, and auditing standards, used in the preparation of financial of accounting systems varies across levels of government (and statements in Nigeria.154 Membership includes representatives of different software is used, and there has not been agreement on how to harmonize). The commitment to starting implementation of accrual-based IPSAS was again confirmed in summer 2015 for early 2016 by the Accountant General (Ahmed Idris), but has not actually happened, given a lack of prerequisites. It is unlikely to 152. Regarding some of the initial challenges experienced with this transition, see: https://synergyaccounting.wordpress.com/2016/11/29/ nepal-public-sector-accounting-standard-npsas-cash-basis- implementation-challenges/. 155. http://www.financialreportingcouncil.gov.ng/index.php/about-us/ 153. The FRC supersedes an earlier institution, the Nigerian member-bodies Accounting Standards Board that had existed since 2003, and in earlier 156. At the time: Jonah Otunla. configurations since 1982. The FRC is loosely modeled on the United 157. At the time: Udo Hilary Isobara. Kingdom’s Financial Reporting Council, which, however, exclusively deals 158. See: https://www.facebook.com/ManufacturingTodayNG/ with corporate governance (including accounting standards). See www posts/408739622522669 [accessed April 30, 2016]. .frc.org.uk. 159. http://oagf.gov.ng/wp-content/uploads/2015/10/STEPS-TO- 154. http://www.financialreportingcouncil.gov.ng/index.php/about-us/fact- SUCCESSFUL-IMPLEMENTATION-OF-IPSAS-ACCRUAL-BASIS-IN- about-frc NIGERIA-BY-2016.pdf 72  Political Economy of Public Financial Management Reforms be feasible before 2020.160 The costs of transitioning to accrual- and adverse audit opinions were issued for several ministries, based IPSAS has been described as very significant; while no compliance with the previous standards (NGAS) had been uneven, aggregate information or assessment on this is available. Based on and not sufficient further evidence of implementation of the the Nigerian experience, Okolieaboh (2014) underlines that costs new PPSAS was available. The assessment therefore argues that can be very significant and efforts should be made to assess likely accounting standards are not (yet) consistently applied. Due to costs when the transition to IPSAS is planned. the institutional setup in Philippines, there is also some problem Overall, there is an impression that Nigeria has sought to do ‘too with the role of the external audit function with regard to assess- much too fast’ with regard to adopting IPSAS; and in some ways ‘put- ing the quality and consistency of accounting standards applied; ting the cart before the horse’ in terms of pursuing tangible improve- given that the COA issues the standards and is at the same time ments in a cost-effective way. While the formation of commissions responsible for external audit. and the formulation of time tables signals commitment, seek- As for several of the other reforms discussed in this chapter, ing to roll out IPSAS across levels of government and with other Tanzania was a relatively ‘early adopter’. Preparation for introduc- PFM reforms (notably completion of the GIFMIS) struggling, true ing IPSAS aligned accounting standards was initiated in the early progress has been difficult. At the same time, costs have already 2000s (PEFA report 2006: 66). Cash-based IPSAS were intro- been considerable.161 Rather than emphasizing the need to shift duced starting in FY2007/08 for the national level, while alterna- accounting standards, from a functional and accountability per- tive standards remained in use for local levels. By the time of the spective, getting a simpler set of standards to be used reliably and 2013 PEFA assessment, local governments had moved ahead by consistently would seem to have been a better route. using accrual-based IPSAS, as did public corporations, while for The Philippines PFM Reform Roadmap for the period 2011 to the central government, cash-based IPSAS still remained in use, 2015 does not discuss IPSAS explicitly, but their introduction has with a target to switch to accrual-based standards by 2016/17.164 nonetheless been pursued. There was a relatively well-defined pro- Subsequently, it was decided to start implementation of accrual- cess of developing the Philippines Public Sector Accounting Stan- based standards for national MDAs from July 2012, with a transi- dards (PPSAS).162 The standards had been developed by a Public tion period until 2017. National efforts to move toward IPSAS Sector Accounting Standards Board (PSASB) that was established were further reinforced by the East African Community’s joint in 2008. The intention is that they will apply to all levels of gov- commitment. The East African Monetary Protocol (EAMP) signed ernment. By COA Resolution No. 2014-003 dated January 24, in late 2013 stipulated a joint transition to accrual-based IPSAS 2014,163 the PPSAS were adopted replacing the previous New standards for all members.165 Member states also agreed to target Government Accounting Standards (NGAS). ‘B’ ratings in relevant indicators of PEFA assessments. PPSAS are broadly consistent with IPSAS guidance and estab- Thus far, while implementation has progressed further than in lish a modified accrual standard. Out of the 32 IPSAS, 28 are the other countries, it still remains early to gauge the impact of IPSAS adopted, of which 25 were already implemented in 2014, with a introduction in Tanzania. The National Audit Office (NAO) reports further 3 implemented since 2015. A revised Chart of Accounts for the fiscal year 2014/15 point out that asset management of was also adopted in accordance with COA Circular No. 2013-002 many MDAs is still not in compliance with IPSAS standards; it dated January 30, 2013. Subsequently, COA Circular 2014-003 also notes a number of other implementation gaps.166 The number dated April 15, 2014 was issued to provide policies and guide- of unqualified audits across MDAs was slightly lower than in the lines on Chart of Account’s implementation (PEFA 2015 draft). previous fiscal year. No PEFA report or Fiscal Transparency Evalu- Training on both new standards was provided, and the Government ation has taken place yet since the introduction of accrual-based Accounting Manual (GAM) was under revision in 2015. accounting standards. The audit office report also underlines the Despite the move to the accounting standards consistent with importance of the capability of the Supreme Audit Institution IPSAS, the 2015 PEFA rates the ‘transparency, completeness and con- (SAI) to assess compliance with the new standards. sistency of accounting standards applied’ as ‘D’, because qualified Overall, the introduction of IPSAS still remains a relatively less known or discussed part of the wider PFM reform agenda for many stakeholders, while the transition is advancing in many countries. 160. Based on consultations with World Bank staff working on PFM reforms in Nigeria. 161. As noted by Okolieaboh, the considerable cost may even create perverse incentives: “Lastly, if the IPSAS implementation turns out to be 164. A key reason why the rating for the quality of in-year financial a huge budget undertaking, it risks becoming an avenue for corruption reporting was kept at C for the 2013 assessment was the fact that thereby defeating the accountability goal; one of its most advertised and expenditure arrears continued to accrue outside the IFMIS. persuasive benefits.” (Okolieaboh 2014: 7). 165. http://www.pwc.be/en/pwc-academy/assets/pdf/ipsasevent/10_ 162. See: http://agap.org.ph/convention/ppt_13/IPSAS.pdf. bernard_ndungu_the_ipsas_journey_in_east_africa.pdf. Members are 163. The 1987 Constitution (Article IX D, 2(2)) provides the COA with a Burundi, Kenya, Rwanda, Tanzania, and Uganda. mandate to promulgate accounting rules and regulations. 166. See: http://www.nao.go.tz/?wpfb_dl=167. The World BAnk  73 Especially at preparation stages, the introduction of IPSAS has Given its importance, some studies have been produced on non- been managed by Accountant Generals or equivalent entities, and technical aspects of procurement reforms. This includes Frøystad has involved professional associations. In contrast, involvement et al. 2010, as well as two notes on procurement reform efforts in of spending units appears to have been rather limited. There is MENA countries (2014). consequently some risk with regard to actual compliance given the Procurement reforms pursued across a number of countries costs and capacity constraints involved. have mainly targeted the following aspects: (a) legal reforms, sig- Furthermore, in sectors and jurisdictions where corruption has nificantly influenced by the UNCITRAL model law as discussed been prevalent, it appears not unlikely that various constraints—in in Chapter 4 and the issuing of related regulations and Standard capacity, as well as in controls and oversight—may be used to mis- Bidding Documents; (b) introduction of e-procurement covering represent expenditures and asset valuations. To ensure that actual various stages of the procurement process; and (c) institutional improvements in transparency and accountability result, in par- reforms such as setting up procurement oversight agencies, typi- ticular in environments with high levels of corruption, some verifi- cally linked to the adoption of new laws. These efforts have also cation that the standards are in fact accurately applied appears as involved training as well as developing a more specialized procure- crucial. As is discussed in Section 5.7 below, internal and external ment cadre. audit functions have a particular role to play in this regard; and The progress on strengthening procurement as diagnosed by strengthening their capacity to audit IPSAS based accounts is an PEFA assessments is rather remarkable given that procurement is important complement. Beyond this, additional attention by non- one of the aspects of PFM (potentially) most significantly affected by government stakeholders as well as through external independent rent-seeking interests. Four of the five countries had achieved at assessments may also be needed, at least until such time that new least a ‘B’ rating (and Georgia an ‘A’ rating) by the mid-2010s; accounting standards and reporting are fully established and well only for Nigeria, greater progress is yet to happen. Both the understood. Philippines and Tanzania had already introduced procurement reforms before the core periods reviewed for each country in this report. In Georgia and Nepal, progress was made between 5.7 Strengthening Public the late-2000s and the mid-2010s, with a particularly strong Procurement 167 improvement in Georgia. In addition to PEFA assessments, reports on ‘Benchmarking Public Procurement’ (World Bank) pro- Reforming public procurement is a central element of PFM reforms, vide an assessment of the quality of rules and regulations from a and in particular with regard to the fact that procurement can account private sector perspective.168 for a significant share of the total leakage of funds (Sanchez 2013). Table 5.6.  PEFA Ratings for Procurement Georgia Nepal Nigeriaa Philippines Tanzania 2008 2013 2008 2015 2012 2010 2016b 2006 2009 2013 PI-19 [PI-24] D+ A C B D+ B B [C+] C+ B [NRc] NR i. Transparency comprehensiveness and competition in D A C B D B B NU B B the legal and regulatory framework. ii. Use of competitive procurement methods. C A C D D B B [B] NU B [NR] NR iii. Public access to complete, reliable and timely C A C C D B B [B] NU B [NR] NR procurement information iv. Existence of an independent administrative NU A NU A D NU B [D] NU NU [D] D procurement complaints system Note: Changes were introduced to the sub-dimensions of this indicator over time, in ways that affect comparability across time and across countries. The sub-dimensions of the 2011 framework are presented and were applied except for the Philippines 2016 assessment. a. Data for Nigeria are based on the unofficial assessment carried out in 2012. b. For the 2016 Philippines assessment, ratings according to the old methodology are quoted, followed by ratings according to the updated methodology in square brackets. The first dimension was changed between methodologies from focusing on the legal framework to focusing on procurement monitoring. c. Reassessment of the 2010 performance made in 2013 based on the new methodology. NU: Not used167; NR: not rated. 167. According to the guidance on the PEFA framework: “In some cases, it may be decided for certain reasons that a particular indicator will not be used. For example, it may be the case that the PEFA assessment is going to be combined with another detailed assessment of the relevant indicator, using a different assessment tool. In all such cases “NU” is 168. Reports have been published since 2015, see: http://bpp entered instead of a score.” .worldbank.org/reports. 74  Political Economy of Public Financial Management Reforms In Georgia, procurement reform was undertaken as part of the multiyear contracting for priority projects. In 2015, the govern- second wave of PFM reform efforts. Following several reports that ment amended the procurement act based on the implementation raised procurement as still weak, reform efforts began in earnest experience of the law. This is, however, not fully compliant to inter- in 2010 but then moved at a fast pace. The new legislation was national standards. A key planned next step is for the PPMO to based on the UNCITRAL framework and in-house work began on merge the individual MDA procurement portals into an integrated designing and implementing a centralized system for procure- procurement portal, which would create a complete e-procurement ment management. Still in the same year, a Georgian Electronic- cycle (from bid publication to awards and contracting), and further Government Procurement (Ge-GP) system was launched. With increase transparency. the e-procurement system in place, the number of tenders adver- The initially created electronic tender process has opened up tised increased from around 2,000 to around 30,000 per year. the submissions of bids, which was previously reported to involve Through the introduction of an electronic procurement platform physical violence to prevent competitors from submitting their bids.169 the Georgian procurement system has become considerably more However, at the execution stage, especially small projects imple- transparent and nondiscriminatory, encouraging competition and mented at the local level still suffer from rent-seeking, collusion, minimizing the risk of corruption. and intimidation. The fragmentation of e-bidding systems by MDAs A key limitation remained that the Law on State Procurement has made it harder to monitor compliance, and central data and 2010 still allowed a number of exceptions which were rather loosely statistics remained unavailable. Furthermore, as the 2015 PEFA specified. In reaction to this issue, the rules for simplified pro- report notes, systematic data on the use of competitive/noncom- curement were tightened in late 2015. In addition, there were petitive methods and whether these have been adequately justi- concerns that having to decide exclusively in favor of the lowest fied or not, are still not available. The report indicates that not only technically acceptable bid could hurt longer-term value for money, competitive processes are used for over 80 percent of tenders; but given that cheaper options often have a shorter life-span or higher also some tenders are deliberately split to remain below thresholds running costs. To enable greater considerations about quality, a that require the use of more competitive methods. two-stage public procurement process for purchases of goods by Furthermore, local stakeholders indicate three key concerns: way of a two-stage electronic tender or a two-stage simplified elec- (a) continuing problems of collusion and intimidation, (b) prob- tronic tender was being prepared. Overall, the reforms achieved lems for many local suppliers to meet the requirements set out in and being pursued further are quite remarkable against the start- the 2007 procurement law, and (c) lengthy processes which—in ing point; with key changes made within a relatively short period addition to budget approval delays and secondary approvals for of five years (2010–2015). capital projects by the Planning Commission—make it very dif- The commitment to international integration–noted as an impor- ficult to execute capital budgets. Between 50 percent and 70 per- tant reason behind the movement to reform procurement in many cent of the expenses are bunched in the last trimester of the fiscal countries—also played a clear role in Georgia. There was a sense ­ year (Krause et al, 2013; PEFA 2015), and under-execution of that the external criticism of procurement needed to be addressed. capital budgets has been around 30 per cent.170 At the same time, as in several areas reviewed above, such as Nigeria likewise adopted new procurement legislation in 2007. the introduction of a TSA, the government still left some initial The Public Procurement Act stipulates the creation of a National wiggle room for favouring preferred companies or for moving pro- Council on Public Procurement (NCPP). However, the NCPP was curement processes quickly for example, during pre-election years never formally constituted and inaugurated. De facto, the Fed- to show progress in construction, through allowing exceptions. On eral Executive Committee has fulfilled some of the responsibilities the issue of procurement reforms, civil society also played some given in the legislation to the NCPP. role in terms of criticizing the uses of simplified procurement (TI For public procurement, a new institution that has been cre- Georgia 2015). ated is the Bureau of Public Procurement (BPP), in accordance with In Nepal, some gains in reforming public procurement were UNCITRAL standards. The BPP supersedes an earlier body, a Bud- made, but substantial concerns still remain. Legal provisions were get Monitoring and Price Intelligence Unit (BMPIU, also known as reformed in 2007 with the introduction of the Public Procurement the ‘Due Process’ Unit) that had been established in 2001 and Act (PPA). This stipulated the establishment of a Public Procure- located within the presidency. The BPP functions as a regulatory ment Monitoring Office (PPMO, under the Office of the Prime Min- body for procurement through issuing guidelines (Procurement ister and the Council of Ministers), and decentralized procurement to MDAs. It was followed by the adoption of new procurement regu- 169. Interviews in Nepal, April 2013. lations. An independent Procurement Review Committee was also 170. The PEFA 2015 report notes that 46 percent of expenses are established to review complaints. In 2010, the government further bunched in the last trimester, but data from the FCGO as well as a report by Krause et al. 2013 show bunching of more than 70 percent in the last reformed the procurement processes by introducing e-bidding and trimester. The World BAnk  75 Guidelines, Standard Bidding Documents). It also provides ‘no reforms, its de facto level of engagement was more limited. One objection’ for tenders above NGN 1 billion for works and NGN reason cited for this is that it was less of a priority as it was the 100 million for goods and services, with approval powers given to legacy of the previous Estrada and Arroyo administrations, so the the Federal Executive Council (FEC). The BPP reported that it had transparency agenda and the GIFMIS stood more at the forefront. brought about cumulative savings of NGN 420 billion between Problems with executing capital expenditures remain substan- 2007 and June 2013 as a result of its review processes. However, tial in the Philippines, and frequent bid failures and weak capacity to it is widely considered that public procurement remains a source procure among MDAs are cited as the main reasons for this. Following of considerable leakage as a result of corrupt practices in contract initial efforts, the contract for the upgrading of Phil-GEPS was awards and management. terminated, and a rebidding was conducted in 2017. The GPPB The post-2015 government remained committed to procurement collects a significant amount of data and analyzes performance on reforms; and has sought to create the NCPP (rather than dropping contracts awarded. However, the data are published on an ad hoc this institutional provision as an alternative option to bring the law basis and not linked to Unified Account Code Structure (UACS) and reality in line). Similar to other countries, there are concerns and the budget. The total value of contracts awarded through com- about the ability of local companies to be competitive and about petitive methods in the most recent fiscal year data was available the time taken by procurement processes; and revisions to the (2014) was 74 percent for 15 agencies, but is likely lower if the 2007 law are under discussion. entire government was considered. While the Government Procure- The Philippines started efforts toward procurement reforms in ment Reform Act provides a protest mechanism, there is no provi- the late 1990s. Initial support of President Estrada in late 2000 and sion for an independent administrative procurement complaints a decisive final push under President Arroyo’s first term enabled body. Moreover, the protest mechanism requires payment of fees the eventual passing of a Procurement Reform Bill in 2003 (see and this discourages potential complaints. Campos and Syquia 2005). The role of civil society was critical to Tanzania has pursued procurement reforms since the early pass the procurement bill through Congress. An NGO specializing 2000s, but also with mixed progress and multiple legal changes made in procurement issues (Procurement Watch Inc, PWI) was created over time. A first UNCITRAL-inspired procurement law was adopted in 2001, which played a key role in generating wide public support in 2004. This act was then replaced by the Public Procurement for the Reform Bill, as well as in monitoring and directing public Act 2011. The latter only became effective in December 2014, attention to subsequent implementation.171 following extensive delays in publishing the regulations. The 2004 The 2003 Procurement Act created the Government Procure- Act created a Public Procurement Regulatory Authority (PPRA), ment Policy Board (GPPB), a policy and monitoring body with the while in addition the 2011 Act provided for a Procurement Policy mandate to manage all procurement related matters together with Division in the Ministry of Finance, with the responsibility to draft internal rules and regulations and a Procurement Manual.172 The and amend procurement rules and regulations. A third body is the GPPB, among others, publishes an annual collection of procure- Public Procurement Appeals Authority (PPAA), established follow- ment monitoring reports on a selection of MDAs and subnational ing the 2004 Act. governments. The Philippines Government Electronic Procurement In June 2016, a further Public Procurement Amendment Bill was System (Phil-GEPS) was piloted in 2001 while the law was being passed. This amendment contains provisions that allow an open drafted and promulgated in parliament with help from the Cana- framework without the need to refer to agreed prices set forth by dian International Development Agency (CIDA) and then was even- government procurement services agency. In this way, the govern- tually rolled out in 2003 and 2004. Phil-GEPS provides a platform ment hopes to provide better value for money in procurement as on which bid notices and awards are published.173 Bid submission opposed to previous procurement legislations. itself (that is, e-bidding) was piloted only later, in 2015, and has According to the Controller and Auditor General (CAG), the not been fully introduced to date. enforcement of procurement legislation remains uneven, and lack of Recent actual progress in further improving the quality of pub- feedback from the regulator to procuring entities may encourage them lic procurement has been limited. While the Aquino administration to consider the reporting requirement to be a formality. The budget remained formally committed to continuing public procurement for goods and services procurement in Tanzania thus remains vul- nerable to misallocation and rent-seeking. In addition, the decen- 171. PWI and other accountability-focused NGOs, such as those tralization of procurement to MDAs has provided opportunities for participating in the Transparency and Accountability Network (TAN), also leakage. There have been many examples of the rigging of ten- benefited from external support, including a media campaign funded by ders for large projects and of smaller contracts being awarded to USAID and initial seed funding for PWI provided by the World Bank. 172. http://www.gppb.gov.ph. companies belonging to government officials or their associates. 173. The number of posted notices grew from 43,000 in 2004 to over At the same time, formal compliance with requirements in terms 600,000 in 2016. The number of bid awards published is less than half of the number of bid notices. of procurement methods used and recorded has been gradually 76  Political Economy of Public Financial Management Reforms increasing. Also, the implementation of public procurement rules 5.8 Strengthening Internal Audit has led to the removal of some members of tender boards (includ- ing MPs, councilors, and MDA/LGA chief executive officers), As reviewed in Table 5.2 at the start of this chapter, internal audit resulting in a substantial loss of rent-seeking opportunities and is another dimension which all case study countries have sought to allowances for those individuals. strengthen. The logic of this is quite clear: internal audit can help Procurement stands out as an area in which all five countries to prevent misuse of funds during the expenditure process. It can have pursued significant legal, institutional, and process reforms also provide ongoing flags for issues that need to be considered by which have brought some improvements. International influences the management. in terms of trade integration and copying of ‘international best From a nontechnical perspective on PFM reforms, it is also easy practices’ has clearly played an important role in the procurement to hypothesize that developing effective internal audit mechanisms is reform efforts that have been made across the five countries. At likely to be challenging. Internal audit can be assumed to require the same time, there has also been a sense that the initial set that managers have an interest in guarding against non-regular of rules adopted was not quite adequate, resulting in subsequent use of funds. Such an interest may not be present in countries further changes. with significant or pervasive corruption, which typically includes Still, important concerns remain, in three respects. One is significant rent-seeking among political and administrative elites. whether the set of formal rules and procedures is actually fit for Furthermore, in societies and organizations that have an estab- purpose (for example, in terms of time periods needed, openness lished culture of strong hierarchies or oligarchic control, internal to smaller suppliers, effects on the national economy and so on), audit staff would likely face dilemmas and restrictions on the type second whether the rules on the books can be implemented across of issues that can be raised. the country and compliance can be reliably monitored (see also As Vani (2010) identifies, modern internal audit is an additional Williams-Elegbe 2015); and a third set of concerns is whether bureaucratic layer to advise management on whether funds are effec- even though rules and procedures have been strengthened, collu- tively deployed and internal control mechanisms appear to be effec- sion can still occur, and illicit gains still be made to the detriment tive. Even in many high-income countries, internal audit functions of an efficient use of scarce funds in the public interest. were only created over the past 20 years. Furthermore, as Diamond The real impact of the procurement reforms in terms of strength- (2013, 377) has pointed out, in contexts where cash releases are ening value for money and of reducing the scope of corruption remains irregular and insufficient, and especially where this is combined as yet little measured, while efforts in this direction are ongoing. with detailed line-item allocations, even well-intentioned manag- These intended impacts of the reforms undertaken are challenging ers may need to circumvent at least some financial regulations to measure—as a World Bank procurement manager noted: “for in seeking to deliver public goods and services. In addition, he many years, procurement indicators have been limited in number raises the issue that the ‘best practices’ for internal audits being and superficial in substance.”174 Most recently, a number of ini- shaped in industrialized countries and modeled on private sector tiatives have been made in this direction, notably a ‘Methodology audits, may not be the best fit for the needs of developing coun- for Assessing Procurement Systems’ (MAPS) led by the OECD.175 tries. He sees a centralized and control-focused approach being Following three pilots, initial assessments using the MAPS are more appropriate, compared to one that is decentralized to MDAs expected to be undertaken from late 2017 onwards. At the World and takes on a more ‘internal consulting’ perspective on how to Bank, efforts are also ongoing to develop and deploy procurement improve processes. performance assessments. Efforts in this direction will be criti- A third point is that—given its ‘in the middle’ role—it can be cal to see whether the ‘formal’ changes in this area which have difficult to establish clearly whether internal audit functions are effec- been widespread, somewhat counter-intuitively so, are matched tive or are just an additional institutional layer created without adding by real ‘functional’ changes in terms of delivering better procure- much value. Simply having a modern audit structure—standards ment results; and how much variation there is in this regard among and manuals, and trained staff—is not equivalent to an increase in countries. expenditure effectiveness and integrity. Most outcome indicators of PFM are the result of several systems working in tandem. When external audit is stronger than internal audit, the former may raise concerns about the latter and identify weaknesses. When both are weak, it becomes more challenging to gauge the actual quality of 174. https://wbnpf.procurementinet.org/post/indicators-are-crucial- procurement-performance-management [accessed February 14, 2017]. internal audit. 175. http://www.oecd.org/gov/ethics/benchmarking-assessment- Across all countries and according to all PEFA reports avail- methodology-public-procurement-systems.htm. Especially indicators 9 and 10 are intended to capture the degree of actual functional progress able as of 2016, internal audit as measured by PI-21 (in the old sys- and use of improved systems. tem and PI-26 in the revised 2016 version), is the second lowest The World BAnk  77 Table 5.7.  PEFA Ratings for Effectiveness of Internal Audit and Sub-dimensions Georgia Nepal Nigeriaa Philippinesb Tanzania 2008 2013 2008 2015 2012 2010 2016 2006 2010 2013 PI-21 [P-26] D+ C+ D+ D+ D+ D+ C C C B (i) Coverage and quality D C D D C D C NU C B (ii) Frequency and distribution of reports C A C C C D C NU C B (iii) Extent of management response C C D D D C C NU C B Note: a. Data for Nigeria are based on the unofficial assessment carried out in 2012; b. while P21 is broadly equivalent to P26 in the 2016 framework, there are some important differences. performing indicator after that measuring the legislative scrutiny Overall, it is quite likely that further progress will be achieved over of audit reports.176 Consistent with the assumption that building the coming years. effective internal audit is difficult in lower income and higher In Nepal, internal audit remained weak, despite being included corruption environments, the countries for which internal audit in PFM reform plans, in 2009/10 as well as again in the second reform is rated as being better performing tend to be either those that plan adopted in 2016. This happened despite the fact that internal have higher income levels and overall less corruption, such as audit is under the leadership of the FCGO, which managed sev- Botswana, Mauritius, and South Africa, as well as some of the eral other reform aspects during this period relatively successfully developmental states, such as Ethiopia.177 (relative to overall country governance constraints) as reviewed Interestingly, across the five case studies reviewed in greater above. However, internal audit does not check the effectiveness detail, internal audit functions have been strengthened in three out of of internal controls; and (as documented in reports of the Auditor five; with particular progress noted for Tanzania. Some strengthening General) follow-up on recommendations made by Internal Audit was also achieved in Georgia and in the Philippines, but to levels Units has been weak and is not systematically tracked. that are still below their overall average levels achieved; while in One apparent reason is that PFM reforms to the extent that they Tanzania internal audit progressed more than PFM performance could be pursued under difficult circumstances, such as Nepal’s, overall. In Nepal, progress has stagnated. For Nigeria only one focused on ‘more basics’ first, in particular improving basic treasury (informal) PEFA assessment is available. However, additional case functions and reliability as reviewed above. Moving to a more ‘mod- study information also does not suggest a significant strengthen- ern’ form of internal audit has ultimately not yet been a high pri- ing of this dimension. ority. As discussed, for the introduction of an FMIS, the political In Georgia, focusing on internal audit started in 2007, while environment of fragmentation and control of political parties at agreeing on an approach for strengthening internal control and audit the local level has also made it difficult to push an internal audit took some time. The initial draft legislation on inspection developed function very far. by the MoF was withdrawn. As for a number of other countries, Weak internal audit functions in Nigeria is largely in line with there was a perceived need to move beyond a traditional control expectations, given a context of very high levels of corruption and and inspection function; in this case as had been developed across also of having yet to achieve progress on a number of other PFM all parts of the former Soviet Union. New legislation eventually reform areas. As identified in the 2013 draft PEFA assessment for took shape through the adoption of a Law on State Internal Audit Nigeria, follow-up on internal audit findings has been very weak; and Inspection in 2010. The reforms have been driven by an EU- starting with the fact that internal audit reports may not include Georgia Agreement on developing a Public Internal Financial Con- actionable recommendations, and that those which are provided trol System, in 2007. In 2011, the MoF assessed the legislative are typically ignored. Given the weaknesses in the external audit framework and initiated amendments to the law and in 2011, a function discussed below, there has not yet been a clear lever to new Law on Public Internal Financial Control was adopted. Signifi- draw greater attention to this issue. cant training was received from the EU and also through the OECD. In the Philippines, internal audit was just beginning to be pur- sued from 2008 onwards. A DBM Circular Letter No. 2008-05 pro- vided guidelines on the organization of the Internal Audit Service 176. See also the IMF blog by S. Vani making a similar point: http:// blog-pfm.imf.org/pfmblog/2010/08/posted-by-sanjay-vani---of-107- and clarified functions and the service’s structure. The Head of the countries-in-which-public-expenditure-and-financial-accountability-pefa- Internal Audit Service or Internal Audit Unit reports directly to the assessments-have-bee.html. 177. No PEFA report is available for China; and the rating for Rwanda is head of the agency or the secretary. Initially, Internal Audit Units ‘only’ a C+ (B+ for Ethiopia). were still rare across the public administration. Given a relatively 78  Political Economy of Public Financial Management Reforms strong external audit function (discussed below), the COA issued even well intentioned managers struggle to follow rules in a con- many qualified or adverse opinions in its external audit reports, text of constant cash constraints. Furthermore, as heads of Inter- criticizing a lack of compliance with internal controls and adher- nal Audit Committees are heads of departments and therefore not ence to financial management rules and regulations. impartial, this leads to questions about the quality of the reports. Following the 2010 elections, attention to strengthening inter- While IAG staff are functionally responsible to the IAG, they are nal audit increased. The Department of Budget and Management administratively responsible to LGAs and MDAs, as these employ issued a new Internal Audit Manual in 2011. By 2016, most the auditors. This makes supervision and discipline of staff by MDAs had audit units in place. Internal audit reports are devel- the IAG difficult.179 Following the 2015 elections of a President oped annually, although in many agencies still with some delays strongly committed to reducing corruption and poor use of funds, due to capacity shortages. There was some shift in COA findings the role of the IAG has been strengthened, and a number of verifi- with regard to somewhat fewer adverse audit findings, but still a cation exercises have been instructed by the President. large share of qualified audit opinions. It remains therefore early Overall, the experience in Tanzania suggests that over 15 to to judge whether the internal audit function becomes truly effec- 20  years, a substantial and more professional internal audit func- tive in increasing the integrity and efficiency in the use of public tion can emerge, while political commitment as well as the relative funds. strength of external audit have been important factors. A contribut- In Tanzania, as for most other PFM reform areas, efforts to estab- ing factor to strengthening internal audit appears to have been lish a modern internal audit function dates back to the late 1990s and relatively well developed external audit (discussed in the next early 2001.178 The 2001 Public Finance Act already required the section)–which has been able to consistently point to weaknesses establishment of Internal Audit Committees in all MDAs; and the in internal controls and compliance, and thus provides a check Public Finance Regulations established a framework for the over- on whether internal controls and audits are functioning. In coun- all internal audit function to provide an independent appraisal of tries such as Nigeria, where external audit still remains very under the adequacy and effectiveness of internal controls. The intended developed, there is no comparable check and stimulus to seek role of Audit Committees is to issue recommendations resulting improvements. Furthermore, since 2015, presidential interest and from internal audit reports, and to ensure and monitor follow-up of backing of the IAG has strengthened the internal audit function. these. An Internal Audit Manual was issued in 2004, and training modules were also developed in the early 2000s, under the leader- ship of the Accountant General. However, capacity still remained 5.9 Reforming and Strengthening low, and by the mid-2000s these systems were not considered as the Role of External Audit effective. Strengthening external audit is a reform that many DPs are keen to By the late 2000s, internal audit still remained weak, but fur- support. The idea is that this is a key mechanism for strengthening ther efforts were undertaken. As reported by the NAO, while most accountability (OECD 2011; GIZ 2013). The number of SAIs has MDAs had Internal Audit Units and Audit Committees, audit units increased significantly over time, with many new SAIs being estab- remained understaffed and ineffective. The revised Public Finance lished following the fall of communist regimes, in Africa, as well as Act (2010) established the position of an Internal Auditor General in other regions since the 1990s. The wave of setting up new SAIs (IAG); thus creating a dedicated leadership role to this ‘bureau- has been followed by many efforts at institutional strengthening. cratic layer’. The first IAG was appointed in 2011. The capacity of External audit potentially plays an important dual role—with the Central Internal Audit Unit was significantly strengthened in regard to ensuring accountability, as well as to stimulating efforts for the following years, with an increase from 5 to 40 staff. The IAG strengthening PFM systems, since at the heart of SAIs mandates is to also established an Audit Report Implementation Unit to be able flag risks and systems’ weaknesses(see also DFID 2005). However, to track audit recommendations and follow-up by MDAs. external audit institutions do not play these two roles alone, but By the mid-2010s, the internal audit function was seen as hav- rather in an accountability system that also involved legislatures ing been strengthened but the actual effectiveness of controls and and executives. Most SAIs do not have direct enforcement pow- compliance with rules still remained limited. The quality of reports of ers.180 Follow-up on SAI findings—or at least the authorizing envi- Internal Audit Units was seen as improving, and follow up on rec- ronment and reinforcement of demands to make changes—mostly ommendations by MDAs was emerging. Internal audit structures rests with national legislatures and with the executive. continued to benefit from support by DPs, and the Institute of Internal Auditors. However, Tanzania also continued to be affected by the challenges pointed out by Diamond (2013), namely that 179. Case study interviews in June 2013. 180. According to the International Organisation of Supreme Audit Institution (INTOSAI) database, out of a total of 90 SAI’s responding on 178. A basic internal audit function had existed since 1961. the question, 33 had the authority to impose sanctions, while 57 did not. The World BAnk  79 Figure 5.2.  Aggregate PEFA Scores by Dimension by Income Group Low income Lower middle income Upper middle income 4 PEFA Dimension Scores 3 2 1 Budget Comprehensiveness Policy-based Predictibility Accounting External credibility and budgeting and control of and scrutiny transparency budget execution reporting and audit Source: Fritz, Sweet, and Verhoeven (2014). As identified in the comparative analysis carried out in 2014, What stands out from Table 5.8 is that it is the last indicator— external audit and oversight is the weakest of the six dimensions that the legislative scrutiny of external audit reports—which is particu- are measured by the PEFA framework;this is true in the aggregate as larly weak. For four of the five countries reviewed in greater depth, well as for each of the more specific income groups (low income, ex post follow-up is significantly weaker than the ex ante review lower middle income, and upper middle income). The average of the budget as measured by PI-27, while for the fifth (Nepal), performance in middle income countries is slightly better than in both PI-27 and 28 are equally weak. The table also reflects that, low income countries, but external scrutiny and audit remains the by contrast, across the countries, some strengthening of external weakest category. As part of this analysis we also explored whether audit offices has happened over time. Thus, while the external external audit is rated more highly in democratic countries than audit function started in all countries from a low level (D+) it has in those that are not. This is based on the assumption that in a tended to improve over time, as reflected for the four countries democratic regime, there are more checks and balances, and so an for which repeat assessments are available. However, legislative institution providing horizontal accountability should be stronger, scrutiny has remained a very weak link, with no recent assessment and parliaments can be expected to play a more effective account- for any of the five countries exceeding a D+. ability role as well. However, contrary to this intuition, there is In Georgia, external audit itself became a source of corruption no statistically significant relationship between the presence of a and political intrigue following independence in the 1990s. The ini- more democratic system and the strength of external scrutiny and tially established Chamber of Control of Georgia was focused on audit as measured by PEFA ratings. inspections and was able to impose fines. In a highly corrupt envi- Therefore, it is of particular interest to look across the case ronment, these powers became instruments for extorting payments studies to gain a closer understanding about why accountability func- from audited entities. Early efforts made by DPs to strengthen tions tend to remain weak. The three relevant PEFA ratings included the institution therefore gained little traction. Eventually, after the in the ‘external scrutiny and audit’ basket measure the following: onset of wider anticorruption reforms, the Chairman of the Cham- PI-26—Scope, nature and follow-up of external audit; PI-27— ber of Control was arrested, tried, and convicted in 2005.181 Legislative scrutiny of the annual budget law (that is, not related to external audit as such); and PI-28—Legislative scrutiny of exter- nal audit reports. The 2016 PEFA framework has introduced some important changes to this basket, including PI-30—External audit 181. For abuse of powers and misuse of funds. The chairman later sued in the European Court of Human Rights over mis-treatment and periods and PI-31—Legislative scrutiny of audit reports, hence removing of unlawful detention and was awarded some damages (US$25,000). PI-27 from this basket is more appropriate. See: hudoc.echr.coe.int/. 80  Political Economy of Public Financial Management Reforms Following the start of the reforms pursued after the Rose Revolu- Overall, Georgia is a clear example of the disconnect between tion of 2003/04, a fundamental reform of the institution still took some a (recently) strengthened SAI and a ‘missing link’ between the SAI time. A new legal basis for a State Audit Office (SAO) was adopted in and the legislature. This is also clearly reflected in the PEFA rat- late 2008. The SAO’s statutes are based on the Westminster Model, ings: PI-26 which rates the quality of the SAI as such improved governed by one person elected by the parliament and account- from a D+ to a B+ between 2008 and 2013. However, indica- able solely to it. The office was formally reconstituted again in mid- tor PI-28—Legislative scrutiny of external audit reports–worsened 2012, shortly before the October 2012 parliamentary elections. over the same period from a C+ to a D+. Having a much stronger From 2009 onwards, the SAO, benefited from substantial sup- SAI should in principle help to eventually move this dimension for- port from international organizations as well as from bringing in new, ward, but only if there is a success with making the appointment qualified staff. Assistance was provided by the EU, German Agency of the Auditor General less politicized. for International Cooperation (GIZ), the EU-OECD Support for In Nepal, there has similarly been some strengthening of the Improvement in Governance and Management (SIGMA), UNDP, SAI as an institution, but in parallel a regression of parliamentary USAID, and the World Bank. Bilateral cooperation has been estab- follow-up. In Nepal, an SAI was first established in the 1950s lished with the SAIs of Sweden, Latvia, Lithuania, the Czech (World Bank and ODI 2013). The existence and role of an SAI Republic, and Poland. Combined with the recruitment of young, was confirmed by the 2007 interim constitution and by the new academically well qualified staff, often with experience from large constitution adopted in 2015. Audit powers are quite far-reaching, international audit firms, the capacity of the audit office strength- and include subnational levels down to districts (but not munici- ened rapidly. However, a change in perceptions of the institution palities) and public corporations with over 50 percent of public has lagged. According to its own management, the SAO struggled ownership. to change its image as a punitive institution, rather than what However, the appointment of a new Auditor General remained it aspires to be: an institution focused on “fostering continuous blocked for six years (2007–2013) due to the failure to reach a improvement in the management of public resources.”182 political consensus–in a politically fraught and unstable postconflict One reason for this is that the original issue that external audit period. Eventually, a new appointment was made in May 2013; is seen as being politicized remained unresolved. The outgoing par- but again when this Auditor General retired, a new appointment liamentary majority decided to appoint a new Auditor General just gap emerged in 2016–2017. The audit cadre is seen as techni- ahead of the 2012 elections for a 5-year term. When parliamentary cally still rather weak; and there are also important limitations due majorities then changed following the 2012 elections, incoming to barely sufficient operating funds. Furthermore, legal provisions MPs saw the SAO as being led by an appointee from the previ- require the Office of the Auditor General to focus extensively on ous government. Therefore, even though the SAO is accountable compliance audits, rather than more on risk-based or performance- to parliament, the relationship has been an uneasy one following focused audits. the shift in majorities in 2012. This has negatively affected the A key limitation with regard to legislative follow-up was the review of and follow-up on audit reports and the overall relation- repeated absence of a functioning legislature, and a preoccupation of ship between the SAO and the legislature. In principle, this should the legislature with agreeing on a new constitution. As reviewed in improve following the 2016 elections and the appointment of a Chapter 3, consensus on a new constitution remained deadlocked new Auditor General which is due in 2017. for five years beyond the original deadline of 2010. Parliament Table 5.8.  PEFA Ratings for Indicators PI-26–28 (‘External Scrutiny and Audit’) Georgia Nepal Nigeriaa Philippines Tanzania 2008 2013 2008 2015 2012 2010 2016 2006 2010 2013 PI-26 [PI-30] D+ B+ D+ C+ D+ B+ B [C+]b D+ B C+ External audit PI-27 [PI-18]c B+ A D+ D C+ C+ B+ C+ C+ B+ Legislative scrutiny of budgets PI-28 [PI-31] C+ D+ D+ D D D D C D+ D+ Legislative scrutiny of audit reports Note: a. Data for Nigeria are based on the unofficial assessment carried out in 2012. b. According to the 2016 methodology, the rating was C+ in 2016, while applying the 2011 methodology, the rating was B. c. PI-27 was moved from the cluster ‘external scrutiny and audit’ to the cluster ‘policy based budgeting’ as PI-18 in the 2016 framework update. 182. Interviews in Tbilisi, April 2015. The World BAnk  81 was dissolved twice, first between 2002 and 2007, that is, the The Philippines has a Commission on Audit (COA), a constitu- later phase of the civil conflict that started in 1996, and then tional body that has an unusual dual role—being responsible for the again from May 2012 until January 2014. In May 2012, the accounting as well as external audit of all government accounts.185 term of the Constituent Assembly elected in 2008 ended without The COA submits the annual financial reports to the President and a new constitution being agreed upon. Eventually in November Congress, but it also submits the external audit report. The 1987 2013, new elections were held. However, the new assembly in Constitution endows the COA the “power, authority and duty to place since early 2014 has still not heard any of the audit reports examine, audit, and settle all accounts pertaining to the revenue that have been submitted by the Auditor General; even though it and receipts of, and expenditures or uses of funds and property, has established a Public Account Committee.183 A new constitu- owned or held in trust by, or pertaining to, the government, or any tion was finally agreed upon by September 2015; but its adoption of its subdivisions, agencies, or instrumentalities” (Article IX  D immediately led to renewed protests. Thus, for nearly 15 years, Section 2). The COA comprises a Chairman, appointed by the there has not been a normally functioning legislature. If there is President for a term of seven years without reappointment, and an overall trajectory toward greater political stability, there could two commissioners, also appointed by the President for the same be some improvement over time. term. The appointments of the Chair and Commission are subject In Nigeria, the external audit function has stayed limited, to the consent of the Congressional Commission on Appointments. despite dating back to independence. Based on available PEFA rat- Apart from a dual role, there is a missing legal link between ings it is the only one in the group still rated at the ‘D’ level (D+ in the external audit function and parliamentary oversight. Apart from a the informal 2012 assessment). The original Audit Act was issued brief set of constitutional provisions, external audit is governed by in 1956, while the current main legal basis are the provisions in a presidential directive rather than a dedicated law. The legislature the 1999 Constitution.184 Efforts have been made to adopt an receives copies of audit reports, but is not required to and does not External Audit Act between 2005 and 2008, but these stalled formally review these, or have any responsibility related to follow- in the National Assembly. Moreover, Auditor Generals have been ing up on the findings of such reports. rotated frequently since 1999, with most staying in office for one The absence of a parliamentary oversight role is partially com- or two years only. pensated by direct follow-up between the COA and the executive. Sub- The timeliness of forwarding external audit reports from the mission of audit reports to the legislature is not legally required. Office of the Auditor General of the Federation (OAuGF) to the Public Therefore, the Philippines Houses of Congress also do not have a Accounts Committees (PACs) in the House of Assembly and the Senate committee dedicated to reviewing audit reports, and regular hear- improved in the late 2000s, that is, there is no longer a backlog of ings on audit reports are not held. Rather than relying on lever- reports that are more than a year behind schedule. However, given age from the legislature, the COA engages in direct follow-up with delays in budget execution and in the Accountant General’s clo- audited entities. This process is relatively strong and systematic, sure of annual accounts, there are still delays in audit reports actu- providing some overall effectiveness to the external audit func- ally reaching the National Assembly. Moreover, while the two PACs tion, despite an odd institutional setup and a lack of effective hold sessions and discuss the findings of audit reports, their rec- parliamentary oversight. In addition, audit reports are being made ommendations need to be approved by the plenary of the respective public. houses of the legislature to complete parliament’s role in budget- Among PFM reforms, strengthening external audit has been rel- ary oversight. This has not happened for any external audit reports atively successful in Tanzania.186 An audit office has existed since since the 1999 transition. This leaves the audit reports ultimately the colonial and post-colonial period. In 2008, a new Public Audit in limbo, and reduces the credible threat of criticism or sanctions Act (PAA) was adopted which strengthened the Auditor General’s to those overseeing spending in the executive for mismanagement independence and powers to make recommendations for legisla- of funds. Furthermore, as OBI reports have raised, audit reports tive revisions and to manage staff; albeit the power to appoint are not made publicly available in Nigeria. Therefore, this avenue the CAG remains (solely) with the President. Further amendments of potential public attention to audit findings and resulting pres- were adopted in 2013. Staff numbers and qualifications have sures on the executive to seek improvements or pursue misuse of funds so far also remains closed. 185. See also IMF (2015, 11). 186. While there was a slight worsening for PI-26 from a ‘B’ to a ‘C+’ rating between the 2010 and 2013 PEFA assessments, the latter 183. See http://www.can.gov.np/. In addition, there is a committee for assessment argues that performance is really unchanged, but on an good government and monitoring. upward trajectory: “Performance unchanged, but MDA follow-up on CAG 184. In contrast to the 1958 Public Finance Bill that is still in force, recommendations is strengthening. The 2010 score of ‘B’ for dimension the 1956 Audit Act is considered as lapsed with the integration of key (iii) on MDA follow-up appears too high.” (PEFA 2013, 110). There is a provisions into the constitution, articles 85–87. clear improvement from the situation in the mid-2000s. 82  Political Economy of Public Financial Management Reforms increased over time.187 CAG reports cover financial accountability parliamentary review have varied—being linked to wider problems and compliance with the PPA and analyze the equity implications with the functioning of legislatures as in Nepal, to the absence of of public expenditures. Performance audits are carried out for such a mandate as in the Philippines, efforts by the executive to selected sectors each year, although capacity is limited and report exercise control in Tanzania, or due to excessive distrust between quality remains poor. Reports are published online.188 successive governments and related appointees as in the case of Development and passing of the Public Audit Act was facilitated Georgia. by the appointment of a professional accountant to the position of The progress made with strengthening external audit offices CAG in 2006 and by Tanzania’s membership in African Organization also reflects one of the potential positive influences of ‘adaptive iso- of Supreme Audit Institutions (AFROSAI) and of INTOSAI. Presidential morphism’. One of the factors that seem to promote greater capacity support under the fourth phase administration, as well as pres- and independence of audit offices is the influence of INTOSAI and sure and dialogue led by development partners that focused on the principles for SAIs it has established and promoted interna- the SAI—played important roles in strengthening the NAO (NOAT tionally. This has been combined with substantial engagement of 2014, 88ff). Controllers and Auditors General have been in office development partners, including twinning arrangements between for relatively long periods; the most recent CAG was appointed audit offices, peer reviews, and other forms of support. in 2014. The current law foresees a fixed five-year term that is Strengthening the legislative oversight role is likely to remain renewable once.189 Within an overall authorizing environment, hard, because this concerns the wider relationship of legislatures and the National Audit Office of Tanzania has also benefited from a executives, but at least some of the countries reviewed could see long-term bilateral partnership with the Swedish Audit Office that improvements. Notably, some review of audit reports by the legis- extended over a 10-year period (2004–2014).190 lature is likely to resume in Nepal, and the relationship between Parliamentary review and follow-up has remained hampered, the audit office and the legislature should improve in Georgia albeit in different ways than in the other countries reviewed. The going forward. Still, as discussed in Chapter 4, the wider rela- legislature has several Oversight Committees that review audit tionship between executives and legislatures can be complex and reports. These include the general Public Accounts Committee, difficult. There is also a risk that in some countries, executives a Public Organizations Accounts Committee, and a Local Authori- have de facto control of legislatures and the review of audit reports ties Account Committee (the former two were merged in 2013). assumes more of a pro forma role; in particular, in countries lean- Extensive hearings on audit findings are being held. The PAC then ing toward more authoritarian forms of government. Two potential drafts its own report and recommendations regarding the audit avenues to pursue are (a) emphasizing the need for audit reports to findings to be tabled before the plenary. One limitation is that this be published and (b) strengthening the follow-up between external process has been lengthy, typically taking a full year. Moreover, the audit and the executive directly, as has been the case in the Philip- power of MPs to begin debating National Audit Office reports and pines—while doing so alongside efforts to expand the involvement the PAC report was reduced through the 2013 amendment to the of legislatures. PAA, according to which parliamentary debate and the approval of the PAC report must only happen once the executive has presented their responses to audit reports, which has in turn been delayed. 5.10 Chapter Summary This indicates that having a truly independent external audit and This Chapter has reviewed the actual progression of PFM reforms oversight function remains a struggle. across the five countries. The initial sections covered the issues of Overall, this review of experiences suggests that strengthening whether PFM reforms are embedded and how reform intentions are audit offices as such has progressed; but that effective parliamentary set out. When PFM reforms are embedded, it is particularly impor- review remains a key bottleneck. The reasons for the weakness of tant for external advisers to grasp relevant connections, rather than looking at PFM in isolation. Section 5.2 emphasized the extent to 187. As of 2014, the office has about 800 staff in total including in which PFM reform intentions have been similar across countries. regional offices; of these, about 600 were professional auditors. See: http://www.policyforum-tz.org/sites/default/files/naovisitpresentation_ Sections 5.3 and 5.4 focus on the issues that (a) reforms 0.pdf. See also PEFA (2013, 107). Recruitment and salary levels are often remain partially implemented for extended periods of managed centrally and not under the control of the Auditor General. time, and (b) that there is a tendency toward expectations that 188. As of mid-2016, the most recent available report is for FY2014/15 which was published in April 2016, that is, 10 months after the end exceed what is typically achieved through PFM reforms in low- of the fiscal year. Reports for previous fiscal years up to 2010/11 are middle-income countries. Both of these aspects have impor- and ­ available and have been posted within similar time frames. www.nao .go.tz. tant implications in terms of more realistic and more deliberately 189. 2008 Public Audit Act, paragraph 6. sequenced approaches to reforms. 190. The collaboration and overall ‘transformation’ of the NAOT is Sections 5.5 reviewed reforms to budget preparation and documented in: http://nao.go.tz/wp-content/uploads/2014/11/2014- TRANSFORMATION-OF-NAOT-BOOK1.pdf. adoption processes, and Section 5.6 to budget implementation, The World BAnk  83 including the introduction of TSAs, of FMISs, and of accounting indicator (for example, the use of program budgets) it can be dif- and fiscal reporting standards. Section 5.7 covered procurement ficult to assess how and how far reforms have progressed. and Sections 5.8 and 5.9 internal and external audit respectively. The sections covering specific reforms also document well the As the ‘process tracing’ within and across countries indicates, degree to which partial reforms have been common. Among upstream implementing specific PFM reforms can be much more surprising and reforms, both program budgeting and MTEFs tend to remain messy than textbook examples. Whether it is Nigeria’s TSA expe- ‘efforts in progress’ rather than completed reforms for extended rience, the Philippines struggle with establishing an IFMIS, or periods of time. Similarly, TSAs and FMISs have remained par- Georgia’s problem with strengthening external accountability and tially implemented in most of the countries—notably in Nepal and oversight, problems can take unexpected shapes, twists, and turns. the Philippines (where the latter still remains pending to actually From a political economy perspective, some patterns are appar- start). ent and consistent with what one might expect in terms of ‘incen- Having many different reform strands that are partially imple- tive compatible’ reforms. One is that trying to roll out new systems mented makes sequencing more challenging. Because most reforms comprehensively can trigger resistance. This was the experience in start but do not end, many reforms end up being pursued (fur- the Philippines, while both Georgia and Nepal took more gradual ther) in parallel. A more clear-cut sequencing is most discernible approaches to establishing TSAs and IFMISs, and these proved for Georgia, where strong political commitment and institutional more feasible and successful. In Nepal, the resulting systems feasibility enabled relatively fast progression of specific reform increase opportunities for managerial oversight somewhat, while aspects, such as a rapid transformation of procurement practices deliberately stopping short of detailed controls. Making credible during a second wave of reforms starting in 2010. promises to honor payment requests promptly has facilitated buy- Furthermore, while the implementation of reforms can be ‘pieced in. The choices for gradual approaches were made by domestic together’, the actual use of reformed systems and their impact remains stakeholders, in the case of Nepal after an earlier failed attempt. difficult to assess and judge. Certain reforms and their impacts are Other reforms seemingly contradict ‘incentive compatibility’. easier to document than others. In principle, whether there is The most notable reforms in this regard is the widespread adop- a TSA covering all MDAs at a specified level of government, or tion of procurement reforms—although public sector contracting whether commercial bank accounts are still in use, is relatively is widely considered a highly corruption-prone area, including clear cut; and in principle also the effects on short-term liquid- being widely used by politicians as a source of election campaign ity management. In contrast, judging whether having a program financing. The introduction of Internal Audit Units even in gover- budget is having a beneficial impact, or whether internal audit nance contexts which seem to emphasize leadership discretion effectively addresses potential misuses of funds is more difficult. also appears as somewhat ‘incompatible’ but has likewise been Overall, most implementation reporting on PFM reforms reflects widespread. One possible reason is that such reforms are seen as inputs, and possibly outputs, but remains silent on outcomes and expedient at least up to a point: procurement reforms are part of impacts (see Figure 1.3 in Chapter 1). efforts at trade integration, and internal audit can serve as a signal Tanzania is a particular case in point, highlighting questions of being serious about PFM reforms—and can be used to keep about impact as well as about sustainability. It had started earlier lower level pilfering in check. than the other four cases covered here with implementing many On the positive side, it is remarkable that PFM strengthening standard reforms since the late 1990s. Actual reform steps were has seen at least some progress even in rather challenging contexts. undertaken. However, the fact that for a number of these, actual After two decades of PFM reforms, all the countries covered have functionalities remain unclear is concerning because results achieved progress in at least some areas. Most notably, even should typically be more visible after a 20-year period. Nigeria made progress toward implementing a GIFMIS and a TSA This is a particular gap in light of the growing emphasis on toward the end of the period covered. considering ‘functional improvements’ rather than forms, as set out It is also notable that detailed accounts of how PFM reforms in the WDR 2017. To understand the degree of functional improve- are actually being implemented are still rather rare. PEFA reports ments and potential remaining bottlenecks, it is critical to know to provide snapshots of many aspects at certain intervals and help to what extent expected outcomes and impacts are actually achieved. track progress; but mostly do not cover how or why reforms hap- As outlined in several of the sections above, this would require pened. Many other reports address particular technical assistance additional efforts by governments as well as by the international aspects. Especially for any aspects not directly covered by a PEFA community. 84  Political Economy of Public Financial Management Reforms Key Findings and Implications for What to Do Differently 6 A s noted at the outset of this report, public financial man- further discussion and also delivers some useful ideas for current agement reforms are being pursued by many countries; and and emerging operational practice. are being supported by a range of development partners. Key findings are the following: While progress has been encouraging in a number of places, it 1. Incremental versus transformational reforms: While some ongo- has also been uneven and has remained behind expectations in a ing PFM reforms are widespread, deep and transformational number of countries. reforms are rare and require an alignment of technical capa- This chapter summarizes the key findings from the synthesis bilities, strong political mandates for reforms, and institutional report. Based on these, some proposals are set out on what could opportunities. The range of countries that have opted to be done differently. The intention is to provide insights from a undertake a PEFA assessment over the past decade—149 in nontechnical perspective for debates about PFM reforms, while total as of early 2017—can be used as a rough gauge of the also setting out operationally and practically relevant suggestions countries with some interest in PFM reforms. At the same of what to do, and what to pay attention to when designing and time, there is only a small number of countries that have implementing PFM reforms. seen clear improvements in indicators measuring overall It is important to keep the strengths and limitations of the meth- ‘government effectiveness’ (WGI) or which show significant odology used in mind when interpreting and using the findings from gains in service delivery relative to spending. Among the five this report. As noted in Chapter 1, the key findings of the report countries analyzed here, transformational change is only are based on a combination of quantitative analytic work and of observed for Georgia thus far. Some gains were also made five in-depth case studies. This combination is considered ‘as in the Philippines and in Nepal. Nigeria and Tanzania both good as possible’, by combining both a review of broad patterns, regressed on overall government effectiveness over the time and a closer understanding of how efforts to strengthen PFM have period analyzed, but are beginning to see progress under evolved in specific countries over a period of about a decade (while their governments elected in 2015. also reflecting on preceding reforms where relevant).191 Additional As is discussed in Chapter 3, Georgia benefited from case studies would be highly desirable to further test the frame- a very strong political mandate to pursue reforms. This was work used as well as the implications presented. It should also be combined with relatively good technical capabilities, and the kept in mind that this report covers a particular time period, that government was committed to using these capabilities and is, roughly the 2000s and 2010s. Fiscal dynamics have been dif- to bringing new, better people into the civil service in several ferent in earlier decades and may again be different in the future key areas. In the Philippines, the pro-reform political man- for many low-and middle-income countries. Furthermore, as PFM date and electoral commitment to reforms also existed, but reforms progress globally, key reform challenges evolve from initial they were weaker and faced both political and institutional adoption and rollout of systems, to actual utilization. blockages. While transformational change did not happen, In conclusion, as raised at the outset, there is no presumption there was some incremental positive change. that this report can deliver a complete set of answers on how to better 2. Reforms are possible even in difficult environments, but are approach PFM reforms, or how to bring about success even in difficult then (far from) linear, and ‘windows of opportunity’ are cru- environments. The intention of the report is to provide a stronger cial to enable breakthroughs. The Nigeria experience indi- empirical basis for some key questions, in a way that stimulates cates that even in most difficult environments PFM reforms can—eventually and incrementally—progress. In this case, 191. An ‘optimal’ design would include a larger number of case studies— a lengthy period of very slow progress from 2006 onwards for example, 15 to 20. However, such a large-scale effort remains beyond was followed by a breakthrough for some reforms following reach for most analytic and evaluation work.   85 the opening of a window of opportunity in 2015. Georgia in expectation that they will deliver ‘basic’ improvements, and 2003 would have also been counted as an extremely dif- to some extent, ‘basic’ reforms and ‘functional improve- ficult case with little progress made since becoming inde- ments’ overlap. However, truly establishing such control is pendent, but experienced a burst of reforms following the very challenging in contexts where leakages benefit impor- 2004 ‘window’. Both in Nigeria and in Georgia, as well as in tant interests. At the same time, reforms that tend to be Nepal—another difficult environment—technical experts at categorized as ‘more advanced’ can be pursued, at least up various levels played important roles in preparing the ground to a point. for eventual greater progress. 5. Institutional ‘starting points’ influence how difficult particular 3. Fiscal dynamics, and PFM and revenue reforms: Cross-country PFM reforms are. The case studies have documented that analysis shows no clear association between fiscal dynamics reforms become significantly more difficult in institutionally and PFM reforms, but the case studies indicate a number of ways more complex settings such as the Philippines and Nigeria, in which fiscal dynamics matter. Fiscal stability and the ability relative to situations in which Ministries of Finance directly to deliver services depend as much, and possibly more, on exercise or oversee a larger number of CFA functions, such the ability to generate revenues as on good PFM practices. A as in Georgia or Tanzania. In addition, the distribution of decline in revenue—due to political crises or loss of natural responsibilities between the executive and the legislative resource revenue as in Nigeria—can negatively affect fiscal plays an important role, including the clarity of roles, and stability and the ability to deliver public goods and services. the probability of a stalemate and delayed budget approvals. In Georgia, one of the main problems before 2004 was the Legal and institutional provisions have gradually dramatic drop in revenue that had occurred in the 1990s, achieved greater attention in PFM assessments, in particular and a key early reform gain was to increase revenue col- in PEFA assessments, but given the systemic importance of lection from very low levels. This facilitated further reforms institutional ‘starting points’ both coverage and implications internally, because civil servants could be reliably paid, bet- still merit further exploration and policy dialogue. Shifting ter ones could be hired, and other expenditures could be institutional structures can be very challenging, and making made on time, enabling improvements in service delivery. In progress is likely to require several years of engagement with Tanzania, a fiscal crunch and growing problems with cash various stakeholders. management had prompted the PFM reform efforts made in 6. Further key institutional issues are intergovernmental policy the late 1990s. and fiscal relations and how they have an impact on expen- At the same time, extreme fiscal pressures can be diture chains and accountability and oversight arrangements. difficult to manage for governments, including those with Across the five case study countries, intergovernmental rela- a reform mandate, in particular, if such a government is tions have posed challenges; and most strongly so in Nepal, seen by many citizens as responsible for an economic cri- Nigeria, and the Philippines. In Nigeria, the federal govern- sis. In Nigeria, revenue collection had been shrinking for ment controls only a limited part of the expenditure chain, over a decade relative to GDP, as became more apparent and has little or no direct oversight over how funds trans- after GDP was re-based in 2014. As oil prices decreased, ferred to states are ultimately spent. While the Philippines— and compounded by other problems affecting the oil sector, in contrast to Nigeria—is a unitary state, subnational levels the outgoing and the new incoming government following nonetheless have considerable power to influence on what the 2015 elections were hard hit by the drop in available and how funds are spent. In Nepal, intergovernmental rela- revenue. The resource crunch has prompted acceleration of tions have been the crucial issue that has made it difficult to certain reforms such as completing a TSA, but the crisis’ agree on and embed a new constitution. Districts have been depth and the difficulty of turning revenue collection around controlled by political parties, and the central government in a complex environment pose risks to the federal govern- only exercises limited oversight over how the dominant local ment’s ability to follow through with its overall reform plans. parties decide to allocate and use funds. Central govern- 4. Basic and advanced PFM reforms: Due to political economy ments have to consider carefully if and how they can extend incentives and constraints, some basic PFM reforms appear control mechanisms to subnational levels—from FMIS, as more challenging than advanced reforms. So-called ‘basic’ TSAs, accounting standards, up to the roles and responsibil- reforms focus on control of public expenditures—good com- ities of central/federal level audit offices. Attention to these mitment control, prevention of leakages in procurement, issues matters in particular with regard to ensuring alloca- in payroll and in non-payroll recurrent expenditures, good tive and technical efficiency of spending, that is, how reli- accounting that reflects reality, and so on. More advanced ably allocated funds actually reach frontline service delivery reforms have been introduced or initiated, in part in the units and are effectively used. 86  Political Economy of Public Financial Management Reforms 7. Identifying stakeholder interests related to PFM strengthening However, there are important external influences. can be challenging and attention is still limited. ‘Functionalist’ Technical and managerial staff in low- and middle-income expectations that stakeholders must want reforms because countries often look around for new models to follow; and for reforms are good for the country and are the right thing various reasons find ‘international norms’ attractive. In turn, to do still persist, and limit attention to the issue of what the adoption of such standards is promoted by a host of pro- stakeholders are really interested in and why. There is often fessionals, international agencies, and consultancy firms. some notion of the background and loyalties of the Minister The growing number of international norms and links, of Finance and possibly some other key staff, but typically and the inducements to adopt them provide both benefits much less about how and why key staff and political leaders and risks. On the one hand, an international standard such would be interested in reforms, or rather in maintaining a as the independence of Supreme Audit Institutions, provides status quo. a benchmark that national reformers can use to argue that Moreover, even when there is an interest, it can be reforms, and reforms in a particular direction, are needed, challenging to identify stakeholder intentions on an ongoing for example, towards their parliaments and presidents. On basis, because such interests are often not clearly stated, the other hand, there are risks that ‘best practice’ interna- and can be difficult to discern. Given that PFM reforms are tional norms distract from simpler options, or options that rarely discussed in the public domain, there is typically lit- for various reasons may be more ‘fit for purpose’. tle public debate that would reveal stakeholder positions or Furthermore, the notion of ‘isomorphic mimicry’ interests. Most governments and individual officials publicly assumes that certain reforms are being imitated without commit to a relatively uniform list of reforms as reviewed seeking to use them in line with the original intention of in Chapter 5. Ex post, divergence between reform com- the given instrument. Such mimicry may of course occur, mitment and concrete actions can be detected, including and in particular in situations where reforms are mainly through data analysis, facilitating analysis of real underlying driven by external leverage, for example, in countries that interests; and stakeholders may also become more willing are significantly aid dependent. However, when managerial to reflect on the motivation of officials and decision makers and technical staff initiate reforms, those are rather cases involved in PFM reforms once they have left office. For ongo- of ‘isomorphic reform’, that is, an effort to ‘copy’ a standard ing reforms, understanding interests involves asking ques- or practice, rather than mimicry. Such copying efforts may tions that can be somewhat uncomfortable in a technical at times be reasonably good fits and deliver improvements, dialogue. while in other situations, copying ultimately fails to achieve 8. Most PFM reforms—not just the introduction of FMIS—take the intended impact. many years from conception to implementation and complete 10. Demand from citizens and civil society has not played a signifi- roll-out, while stakeholders often pursue unrealistic time cant role for PFM reforms across the five countries reviewed, frames. The continuous postponement of reforms—the given their technical and internal features; while citizens can GIFMIS in the Philippines, or IPSAS rollout in Nigeria—can ­ play a crucial role with regard to the overall reform mandate. undermine the credibility of reforms as well as of the stake- The specific reforms pursued across the five countries holders involved. While it may be difficult to determine a mostly only had only limited or no involvement from civil drawn-out time table at the outset, ensuring that there are society; albeit with some greater attention in particular on clear interim provisions, and other ways of building in an procurement reforms. anticipation of ‘partial reforms’ that endure for some time, is Citizens still matter, at least in countries where some crucial. Unrealistic reform plans can be costly; and increase form of competitive elections are regularly organized, as in risks if attention and support move on too early. the case of the five cases reviewed. Citizens voted for signifi- 9. ‘Isomorphism’, that is, copying of standards and systems is cant change and reforms in Georgia in very large numbers, more the result of inducements than being the result of external and to a more limited extent also in the Philippines in 2010, leverage than is being assumed, and there are probably more and in Nigeria and Tanzania in 2015. In particular in Geor- instances of ‘isomorphic reform’ than of intentional ‘isomor- gia, and to some extent also in Nigeria, electoral change led phic mimicry’. In the countries studied in detail here, and to a significant acceleration of PFM reforms. However, the similarly in many other low- and middle-income countries, experiences of Tanzania and Nigeria also indicate that the advanced PFM reforms are being pursued, even though ‘long route’ is an imprecise mechanism; sufficiently specific there is limited or no relevant external leverage in the sense and effective reform agendas, as well as a willingness and of external conditionality or high-aid dependency. ability to take on vested interests, are crucial for politicians to achieve their election promises to improve governance. The World BAnk  87 11. Political economy factors constrain the outcomes and impacts Furthermore, since de facto, real sequencing between of PFM reforms: For PFM reforms, the ultimate issue is not that basic and advanced reforms has rarely happened consistently, reforms cannot progress, but that they do not deliver the full many countries today present a mixed challenge of various level of results or benefits promised. Even in countries where partially implemented bits of both types of PFM reforms. reform commitment was low for extended periods of time, Reformers and development partners should ensure to keep the introduction of new systems eventually progressed. This track of where ‘basic’ functional improvements stand, and has been notably the case in Nigeria where years of very lim- how they can be continuously supported, also with a view to ited progress pursued by a set of technocrats eventually gave avoiding backsliding from improvements that have already way to acceleration at least in some areas. However, in one been achieved. This includes paying attention to whether way or another, all case studies also highlight risks that PFM ‘advanced reforms’ that have been (partially) implemented reforms do not deliver the promised or expected impacts in are contributing to basic functionalities (such as whether a full. In Tanzania, several reforms were implemented already FMIS is actually used for controlling commitments). since the 1990s—that is, have had time to mature—but For various forms of external support, it is crucial to they proved not to be sufficiently robust in the face of continuously seek to develop and sustain the basics along- declining political commitment to good PFM practices. side supporting selected advanced reforms, to make sure Notably, the FMIS in Tanzania, although well-­ that basic control mechanisms become fully institution- established and utilized, did not deliver the range of benefits alized. This can further contribute to lowering the risk of that were hoped for; while in Nepal, the partial automatiza- backsliding, for example, during periods of fiscal stress tion also only provides a limited set of controls. Similar con- or periods of political leadership less committed to good cerns are raised for a range of countries in a forthcoming financial governance. One challenge that may arise for DPs evaluation of FMIS experiences (World Bank, forthcoming). from this is that results frameworks or log frames tend to be Along the same lines, expectations that corruption in Nigeria focused on introducing something new, more than on sus- can be rolled back (primarily) through the reform of account- taining PFM processes and practices such as commitment ing standards—as proposed by the Accountant General at controls or accounting practices that are already in place. the time of launching the initiative—clearly appear greater than is likely to hold true in eventual implementation. • While slow and incremental improvements are common, and transformational progress happens in some cases, risks of Given the key findings as summarized above as well as the backsliding also exist, and need to be considered more explic- detailed findings reflected in the chapters of this report, there are itly. As the quantitative analysis indicates, countries with ini- a number of potential implications for ‘doing things differently’. tially lower PEFA scores saw greater improvements over time Any such efforts are not a hard science, but they do seek to follow than those with already better systems, however, the gains a logic of what is likely to be a good fit in different situations. are relatively small over a 10-year period. At the same time, global indicators also suggest limited aggregate improve- • A perspective on nontechnical drivers for PFM reforms suggests ment due to cases of backsliding; illustrated among the case a need to reconsider and nuance arguments about the sequenc- studies by Tanzania’s experience. Efforts at strengthening ing of PFM reforms.192 Improving ‘the basics’ is critical in PFM systems need to be calibrated to these different ‘oppor- terms of key expected impacts of PFM reforms. However, it tunity environments’. This should include having some in- is also the relatively harder goal to achieve. Including ‘non- built flexibility to change between a more incremental and technical drivers’ into sequencing considerations implies a more fundamental reform agenda, for example, following that it is important to continuously keep some focus on the changes in government. During nontransformational peri- basics, without assuming that these are likely to be com- ods, a deliberate focus on (a) developing a reform agenda pleted before advanced reforms are initiated. In parallel, and (b) on monitoring and seeking to counteract backsliding it is worth keeping an eye on whether some stakeholders risks should be pursued. support ‘advanced’ reforms mostly as a ‘distraction’ in ways that are very unlikely to deliver even incremental tangible • A further crucial implication from the analysis is that windows improvements. of opportunity are very important—but it can be difficult for development partners to effectively provide support to incom- ing, highly reform-oriented governments. The opportunities 192. Note also the implication drawn by Allen et al. (2013, 6): “As a for PFM reforms can vary considerably between one govern- result of the development of political economy analysis, technically based models of PFM development and sequencing—such as the platform ment and the next. A government that is strongly interested approach—appear to have limited applicability in practice.” and motivated to pursue reforms can achieve more in two to 88  Political Economy of Public Financial Management Reforms three years than is likely to be achieved with a government by fiscal trends and ‘opportunities’ such as that offered by not interested in improving governance and the manage- HIPC in the case of the former two countries. In Nepal, suc- ment of public finance over a period of 8 or 10 years. The cessive governments—including those led by Maoists and turnaround in Georgia, and the post-2015 government in Communists—have been very interested in ensuring fiscal Nigeria, and possibly also in Tanzania, illustrate this issue. balance and shoring up revenue collection. One mecha- One important question is how best to use windows of nism is that fiscal crises can give rise to a strong reform opportunity. The experience of the 2010 government in the mandate. Another is that facing growing fiscal constraints Philippines suggests that seeking to pursue a ‘big sweep’ relative to public sector tasks, governments seek to mobilize reform with a long lead time (procuring a GIFMIS) is risky; new sources of revenue and to use existing sources more comparatively, the push to complete an existing TSA reform effectively. Thus, when pursuing or supporting PFM reforms, in Nigeria was closer to a ‘quick win’ approach in a limited considering what the fiscal goals and concerns of the politi- but tangible area. cal leadership are is crucial. Windows of opportunity are also challenging because (a) project cycles are rarely aligned to election cycles and • When governments are motivated to pursue PFM reforms due to fiscal challenges, they are likely to seek reforms both on (b) developing new support within the few months from new the revenue and on the expenditure side, with implications for leadership being in place to the closing of the peak ‘honey- the design of support. To overcome a fiscal crisis typically moon period’ is difficult. Leadership changes and resulting requires measures on both sides; and attention to revenue changes in policy orientation and ‘reform drive’ are difficult trends should be reflected routinely in PFM operations (for to predict ex ante, and key appointments and more detailed example, this seems to have been a neglected issue in policy orientation often only become clearer several months Nigeria for some time before the fiscal situation becoming into a new administration. It can be challenging to then rap- severe). Governments are likely to seek support on multiple idly prepare an operation that effectively supports what the aspects, including debt management, fiscal policy, revenue government seeks to do in time to still use the ‘honeymoon administration, as well as expenditure management. This period’ of such a new government, which extends from ‘the poses some challenges for how (at least broadly) coherent first 100 days’ to a period of two years, hence leaving little support can be provided, given that such a ‘broad front’ is space to prepare and launch a new project. likely to involve multiple development agencies and divi- One option can be to have an existing project in place sions within agencies. Given a growing emphasis on domes- that straddles an election period with sufficient built-in tic revenue mobilization in the World Bank Group, there is flexibility to be adjusted to a new government’s needs and generally greater attention to the revenue side than in the intentions, either through a rapid restructuring and reautho- past. At the same time, seeking to mobilize additional reve- rization process, or through within-project flexibility. More nue once a macrofiscal crisis is ongoing may be challenging ‘advanced’ options such as a pooled mechanism with flex- and involve its own set of nontechnical drivers, for example, ible funding can also be envisaged, but none of the five case with regard to the question of who should pay more. studies offers an example in this regard. For the majority of time outside of windows of opportu- • Addressing stakeholder interests more explicitly. An under- nities, it remains worth pursuing improvements, but progress standing of the motivation and aims of stakeholders, just is likely to be incremental and frustrating to those seeking a as of technical questions, is cumulative. Some relatively faster pace and tangible results. A key potential benefit is simple steps such as increased attention and integrating a to ‘have something ready to move’ when windows of oppor- sense of ‘nontechnical drivers’ into widely used diagnostics, tunity arise and—also importantly—to sustain some of the such as PEFA assessments, could lead at least part of the gains that may have been made during previous ‘windows’. way to better understanding the motivation and incentives For example, in Nigeria, PFM reforms saw an initial burst of key stakeholders. In a number of situations it may also be up to HIPC completion in 2006, followed by an extended possible to organize a small seminar of key staff and to do a slowdown and reversals, and then by a renewed reform push rapid joint mapping of key stakeholders to encourage think- following the 2015 elections. ing through of what reforms are most likely to progress and to bring about real improvements. • Fiscal space is an important motivation of governments, and In this respect, it is important to avoid an overly sim- the links between fiscal trends and PFM reforms should receive ple categorization of stakeholders into ‘reform proponents’ greater attention. PFM reforms in particular in Tanzania, and ‘reform opponents’. Most stakeholders have somewhat Nigeria, and Georgia have been driven to a significant extent complex motivations, for example, they may endorse some The World BAnk  89 reforms, but may be cautious about pushing too far in a • Being clear about institutional arrangements and roles is essen- difficult environment, for example, with regard to procure- tial for assessing bottlenecks and likely difficulties and limita- ment reforms. Or they may seek reforms, but also a good tions in strengthening PFM systems, as well as for identifying business deal for a local IT company. Conversely, stakehold- priorities for engagement on institutional changes. Institutional ers who may not be particularly inclined toward pursuing arrangements including the set-up of Central Finance Agen- PFM reforms may change their stance once they experience cies, the executive-legislative relationship and powers over growing fiscal and cash management problems, or rising dis- fiscal and budget matters, and intergovernmental arrange- content over insufficient service delivery in various regions, ments including those for oversight and accountability, which make progress more urgent. are critical and can enable or pose significant obstacles to In addition, it is relevant to differentiate what interests reform progress. On the one hand, coordination challenges, and motivations are of different categories of stakeholders. challenges of parliamentary approval, or subnational resis- Decision makers such as Prime Ministers or Presidents may tance should truly be taken into account when considering be particularly interested in PFM reforms if they have made reform strategies. On the other hand, it can also be impor- electoral pledges to improve service delivery and governance tant to consider if seeking to change existing institutional and seek to show progress. Senior civil servants and pro- arrangements is essential to be able to achieve tangible fessionals will often be keen to show that they are aligned progress. with the ‘vision at the top’.193 They may also appreciate This gives rise to the issue if one should try to address the greater visibility and exposure which being involved institutional arrangements when these are significantly sub- in a reform process offers. At the same time, they play a optimal. Sub-optimal institutional arrangements are ‘sticky’ critical role in shaping the specifics of PFM reforms; and because they are difficult to change; requiring at a mini- it can be important to discuss and mull over the specific mum an agreement between multiple institutions and the reform approaches they seek to pursue.194 Stakeholders in political leadership, often requiring parliamentary approval. line ministries and at subnational levels are a further impor- For example, for new organic budget legislation or amend- tant group to consider and engage with. Encouraging a more ments, sometimes from two legislative bodies (lower and explicit and franker discussion at various international gath- upper houses), and sometimes requiring two-thirds majori- erings of PFM specialists about nontechnical factors and ties in parliament if a constitutional change is involved, or how these played out in different countries may also be par- also agreement from subnational governments and parlia- ticularly useful to senior civil servants and professionals who ments. Such hurdles can be (near) insurmountable. are seeking to ‘make change happen’. Given that the issues and the level of difficulty of One important aspect for reform stakeholders is seeking approval varies, there can be no hard-and-fast that—also in line with insights from ‘doing development dif- rule—whether to tackle institutional arrangements needs to ferently’ and earlier discussions about rapid results—initial be based on case-by-case analysis. When key PFM-related progress can help to build the credibility of reforms. Espe- laws are truly outdated, poorly designed, or missing key ele- cially at early stages, reform stakeholders may want to be ments, it is worth seeking to update them—even if success careful about reform approaches that entail very long plan- may take some time. Preparing a new law can be done very ning phases, or that are very complex or likely to face strong rapidly if all stakeholders agree and feel a sense of urgency, resistance. Easy as well as effective reform options may not but more commonly takes at least 2–3 years;195 and in many always be available, but stakeholders should consider care- countries a further number of years to achieve adoption by fully if any options can be identified—even if this implies the legislature. initially pursuing a partial reform (see also point 10 below). Updating institutional mandates or structures, such as creating more comprehensive ministries of finance, can in principle be done more quickly through executive deci- sion, but may still require also some legal amendments; 193. Conversely, these stakeholders typically would be reluctant or have and finding political agreement may run against the logic of at best limited ability to ‘swim against the tide’ of top-level leadership preferences. spreading influential positions among multiple parties, fac- 194. This report has not explicitly explored whether particular tions, or ethnic groups. Developing clear ideas and a map of organizational arrangements, such as creating dedicated reform units, have been associated with better results. From a political economy perspective, a particular issue of interest is whether more committed 195. A somewhat drawn-out process of development can be beneficial governments achieve more when adopting certain mechanisms such as to ensure that the provisions of new legislation are consistent with other reform units and/or interministerial steering committees. These are issues existing laws, and to avoid a copy and paste from legislation in other for further exploration in future analytic work. countries. 90  Political Economy of Public Financial Management Reforms what it would take to make a change and what it is expected Related to this, achieving the potential functional to deliver is an important first step. It may also be use- improvement requires attention to what is to be improved— ful to engage in a discussion about the costs and missed which can be neglected when the main focus is on getting a opportunities that are associated with having a fragmented particular instrument in place. For example, while the con- institutional set-up. tracting and rollout of an FMIS can absorb a lot of attention, For development partners, an important implication it is critical to also stay focused on whether the system is is that again, there is a need to engage more broadly with actually used for purposes such as controlling commitments. parliaments on public financial management, not just with PFM strengthening and specific elements of it should regard to budgetary oversight, but also with regard to key be considered, and are more likely to be successful—in aspects of PFM reforms. Parliamentarians are key stake- terms of reducing misuse of funds, as well as in terms of holders when it comes to agreeing to legislative reforms— increasing efficiency in service delivery—when they are whether organic budget laws, procurement or debt-related undertaken jointly with other measures. This holds par- legislation, or reforming laws related to internal and external ticularly true in contexts marked by pervasive corruption. audit. Moreover, in nontransformative environments, expectations should be more constrained, because stakeholders will often • Being clear what the priority functional gaps are and having a find opportunities to circumvent the (full) effect of reforms realistic approach to how these can best be addressed. As the and of specific instruments used. PSMRP 2012 and the WDR 2017 emphasize, it is impor- As part of prioritizing which functional gaps to pursue tant to consider reforms that target ‘functions rather than and which instruments to use, it would also be desirable to forms’. The important issue is not to introduce an MTEF, but develop a more systematic understanding of the costs of various rather to move toward an approach to budgeting that consid- PFM reform packages. As noted in Chapter 5, reform costs ers medium- and longer-term implications of spending com- are rarely explicitly discussed. We have some information mitments. In many countries, the number of functional gaps on the cost ranges of the initial procurement of FMIS. For or limitations remains substantial, so focusing on functional reforms that do not entail one or a few large-scale procure- gaps as such needs to be combined with prioritizing among ments, costs tend to be less clear—for example, the costs of different functional aspects. adopting IPSAS, or of establishing and routinizing MTEFs. As we have seen in the discussion of the experiences, Especially where costs mainly involve consultants, work- both a clear identification of gaps and prioritization are at shops, and similar inputs funded from multiple sources over times missing—reform plans include a long list of areas to a period of time, getting a clear estimate may be tricky. How- be addressed, and there is still a strong focus on introduc- ever, some dedicated analytical effort in this regard would ing a range of tools. One aspect of this is also to guide the be an important complement to existing discussions about discussion with stakeholders about what support might be reform selection and prioritization. needed in a more problem-driven way (that is, identifying what are the most urgent or most widely felt functional • Regarding risks of ‘isomorphic reforms’ and ‘isomorphic mim- gaps), and to jointly consider for which of these solutions or icry’, development partners may need to become more reflec- at least partial progress may be feasible. tive of the international norms they promote. Recent years have Moreover, it is important to be more realistic about seen both a debate on ‘isomorphic mimicry’ on the one hand, what functional improvements a given instrument is likely to and a further proliferation of international norms such as the deliver given continuous challenges. An MTEF may deliver formulation of IPSAS on the other hand. A political economy some sense of fiscal intentions beyond the next fiscal year in perspective suggests that it is most important to have inter- some years, but may have little binding power in many con- national norms for those issues that go ‘against the grain’ of texts and years, particularly when fiscal volatility or political key national stakeholders. In those areas, an international change intervenes. An FMIS is likely to increase timeliness norm can help local reform stakeholders to push for the of reporting and ‘monitorability’, but may not deliver effec- necessity and appropriateness of a particular reform, such tive commitment control, or reduce reallocation between as an independent external audit function, effective internal expenditures and the use of exceptional procedures. The audit, or transparency and competitiveness in procurement. (most) likely functional gains that can be made through the As international norms and standards are promoted, introduction of a particular instrument need to be assessed it is also critical that national stakeholders—accountants, realistically. This requires also some honest stock-taking of auditors, but also NGOs, or staff of parliamentary budget why bottlenecks to certain functional improvements exist. offices—have training opportunities to understand these in The World BAnk  91 depth. This is critical for enabling such stakeholders to have and demand to receive additional funds for example, for a an informed discussion both about how principles and stan- particular municipality or region. In Georgia, overall citizen dards should be adopted and adapted into national systems, and civil society engagement was relatively weak, but inter- as well as to judge whether applications once introduced est in particular aspects such as public sector pay bonuses are done appropriately. Conversely, if standards are widely and procurement was stronger, as these were perceived as promoted, but with little check on how they are applied in important remaining areas in which funds were not consis- practice, this can incentivize mimicry, that is, an imitation tently well managed. Public interest also played an impor- of standards without a real intention or capacity to use these tant role in initial procurement reforms in the late 1990s in toward intended effects. the Philippines. Specialized NGOs that focus at least on some aspects • Given that citizen demand for PFM reforms tends to be weak, of PFM reforms are emerging in many countries, including should reform stakeholders try stimulating citizen engage- the five cases reviewed in this report, offering an opportu- ment? The key finding that citizens and civil society have nity. There is scope to further grow the range of interest not been strongly involved in, demanding, or encouraging and the influence of such groups. Potentially, civil society PFM reforms raises the question whether stakeholders seek- engagement can also stimulate greater interest and support ing to pursue and support such reforms should encourage for PFM reforms among MPs, who—as outlined above—are and stimulate citizen engagement. None of the case studies an important set of stakeholders with a direct influence on offers a strong precedent in this regard, with engagement whether or not PFM reforms can proceed. Thus, even though being relatively strongest in the Philippines. most PFM reforms to date have happened without being Overall citizen expectations are crucial for electing directly demand-driven, developing a better understanding better governments as has happened in Nigeria and Tanzania of fiscal and PFM issues and informed public demand can in 2015, and in Georgia since 2004, creating a ‘long route’ have important benefits both in terms of creating incentives of demand for improving PFM. In principle, because govern- for political leaders and decision makers, and in terms of ments care at least to some extent about citizen expecta- creating a demand to sustain and fully use certain reforms tions, strengthening ‘short route’ demand side engagement such as the introduction of program budgets and perfor- has the potential to contribute to motivating and sustaining mance related information. PFM improvements. Such engagement is most likely to be channeled by • Should reform stakeholders deliberately target partial reforms specialized NGOs such as Transparency in Georgia or BudgIT and unorthodox approaches? As discussed in Chapter 5, many in Nigeria, as it requires building some technical knowledge PFM reforms remain partially completed for many years. of reform issues and options. Engagement by such NGOs There appears to be some benefit to introducing reforms also provides a basis for increased media coverage, which deliberately in an incremental way rather than seeking draws on analysis and commentary. From the side of ‘within all out comprehensive reforms which are at a higher risk government’ reform proponents, citizen engagement can be of failure, as happened with the failed introduction of a facilitated by sharing information, as well as by ensuring comprehensive GIFMIS in the Philippines. Is it advisable that the content and expected effects of reforms are shared to target incremental or partial reforms deliberately for a with CSOs and citizens in an understandable form. range of issues, as a potential ‘end state’ at least for some Moreover, it would be important to describe clearly years rather than as a purely intermediate reform milestone? why reforms are being undertaken and what the expected There is some balance to be struck between the feasibil- benefits are. When this is clear, then citizens can contribute ity of reforms on the one hand, and the impacts that can to holding government to account for actually delivering on potentially be achieved on the other hand. Introducing par- these expected benefits. Conversely, promising to ‘end cor- tial reforms such as a TSA covering only some accounts/ ruption’ through accounting reforms, as was done in Nigeria, sections of government, partial automation, or some aspects may not be helpful for stimulating citizen demand, as it is of procurement reforms such as e-bidding, is more likely to likely to lead to disappointment with regard to the actual be feasible than comprehensive reforms. A key disadvantage impacts, even if the reform as such is implemented. is that the likely impacts also remain more partial. Overall, it Experience in Nepal indicates that if there is an effort is likely that for countries and time periods that are not win- to stimulate citizen demand, it has to distinguish (more) dows of opportunity, deliberately targeting partial reforms is clearly between the demand to use funds well and to allocate a more ‘PE smart’ approach than seeking more comprehen- funds to policy priorities benefiting a majority of citizens, sive reforms that get stuck. 92  Political Economy of Public Financial Management Reforms A related question is whether to consider ‘unorthodox management and integrity and efficiency in the use of approaches’, that is, approaches to PFM reforms that con- funds. As noted in Section 5.6, the veracity of accounting is tradict conventional notions of ‘best practice’. Both Georgia assessed in less depth compared to fiscal ROSCs, and track- and Nepal offer examples of rather successfully introducing ing of the losses of funds as they move between central gov- ‘unorthodox’ approaches to treasury automatization. Some ernment and frontline service delivery units is very limited. may consider ‘unorthodox’ approaches as inherently attrac- Assessments that target certain functional improvements in tive, in line with criticizing ‘best practices’ as typically lack- depth should be undertaken at least from time to time and ing a good enough fit and promoting isomorphic mimicry. for areas that have been identified as key bottlenecks. However, there are also some important caveats. Certain ‘unorthodox solutions’ such as nationally programmed rather • A greater emphasis on sustainability is crucial, as risks of ad hoc interference and backsliding persist, in particular across than ‘off the shelf’ IT systems can be costly and become leadership changes. As noted in the section on key find- failures if national IT capacity is weak, or if the firms or ings, there is a sense that the focus of PFM reform efforts staff to develop these are not selected based on meritocratic tends to continuously shift to new areas, and a tendency to criteria in a weak governance context. neglect efforts at fully embedding and sustaining improve- Thus, while unorthodox solutions can be attractive ments that have been achieved. A range of specific PFM in terms of offering a good fit with specific reform needs ­ improvements—from improvements in budgeting practices and opportunities, they require due consideration of poten- (policy orientation, greater realism of expected allocations) tial risks and downsides and how these stack up relative to to better procurement practices to good accounting, report- pursuing a more standard approach. A significant benefit ing, and auditing—require efforts over a considerable period of considering unorthodox solutions is the opening up of of time to become fully embedded. These include the adop- options relative to exclusively considering a binary choice tion or updating of implementing regulations, training of between leaving things as they are, and pursuing all out best staff or managing staff renewal, ensuring that new processes practice reforms for a given aspect of PFM. are fully understood and routinely followed and so on. Estab- Addressing Obstacles to Sustainability and Impact lishing and fully embedding strong routines and practices is essential for building stronger buffers against potential ad • Tracking what functional improvements are being made and hoc political interference. Reform support that constantly sustained over time is critical in terms of incentivizing real seeks out new areas plays into the hands of political and reforms, as well as for planning further steps. Tracking of administrative leadership that seeks to block or undermine PFM systems and functions over time provides a chance reforms during implementation. to understand what has been achieved and to make cor- Based on these recommendations, there are several specific rections where necessary. Positive assessments can help to implications for operational design that emerge. confirm areas in which progress has been achieved, provid- ing recognition and confirmation to stakeholders involved. 1. It is critical to explore what counterparts are really interested Assessments that target expected outputs and outcomes, in, what they seek to improve or what problem they seek to moreover, are critical to assess to what extent the introduc- solve, and how they believe this can be done. In a number tion of particular instruments or ‘forms’ are actually associ- of situations, governments initially articulate a particular ated with functional improvements. reform tool, for example, the need for an MTEF, but what In this regard the case studies indicate the benefits they are really trying to address is to bind the hand of politi- as well as some current challenges related to repeated PEFA cians to bring additional requests into the budget for exam- assessments. Among the cases, the number of currently ple, at advanced stages of preparation. Clients that really completed national level PEFAs ranges from 0 for Nigeria to care about improvements typically welcome a more in-depth 3 for Tanzania. PEFA assessments can help to pinpoint areas discussion of what the key issues are and the options for of continued difficulty or backsliding as with regard to bud- addressing them. get credibility in Tanzania, as well as to recognize areas of 2. In the same vein, to the extent possible it is important to have progress. At the same time, some of the repeat assessments a discussion on what government counterparts see as potential reviewed also contain potentially over-optimistic ratings, and blockages, and to probe on those that have been observed finalizing ratings can be a contested and negotiated process. in other countries—for example, subnational governments/ Furthermore, there are limitations of PEFA assess- power holders, leaders in line ministries, and so on. Frank ments, in particular with regard to the ‘causal chain’ of discussions on this—in whatever format possible—can be linking PFM improvements to improvements in fiscal very useful to tease out risks and issues to consider. The World BAnk  93 3. Understanding and discussing with the client what reform efforts or being combined with other public sector or service have or have not worked as expected in the past and why. For delivery reforms, and whether to stay upstream, that example, in Nepal, the initial failed attempt at establish- is, typically at the national level, or whether to follow ing an FMIS informed how this reform was approached in the expenditure chain from national levels down toward the second attempt. Past problems can be very informative the front line. Including nontechnical factors such as about potential blockages and challenges; while it is essen- stakeholder commitment in the consideration can help tial to also keep in mind that some dynamics can funda- make these choices in a more robust way. Typically, if a mentally shift. government really knows what it wants, a narrow design 4. Probing buy-in from a range of stakeholders. Probing broader is likely to work best, while starting somewhat more buy-in for reforms is critical in at least two ways. If line broadly may be appropriate when there is less certainty ministries or other key decision makers are opposed or unin- about what will work—but still avoiding overload, and terested in reforms being proposed by Ministries of Finance, staying focused on tangible improvements that can be this may create problems for reforms effectively. Also, spend- monitored. Because many countries find themselves ing units can have important additional perspectives on what now in a halfway world of partially implemented PFM key bottlenecks and issues are. A second aspect concerns reforms that have not quite delivered the improvements more other central agencies and decision makers, as well as originally expected, it will also often be most appropri- senior civil servants within a CFA (and across these in more ate to focus on reform aspects along spending chains, fragmented institutional settings): as Ministers of Finance for example, looking at how funds actually move toward come and go, broader central buy-in is important to be sure frontline spending units, or how reliably spending units that at least some reform commitment and momentum is report on the actual use of funds. Along the expenditure likely to continue beyond individual tenures. chain, it may often turn out that other public sector 5. Encouraging and having frank discussions between teams, aspects are present as key bottlenecks—for example, practice managers, and CMUs about whether the engagement provisions on intergovernmental relations or weak man- is likely to be able to use a ‘window of opportunity’ or is more agement and motivation of civil servants, looping back intervening during a period of relative stagnation. This should to point 4. help to set realistic expectations, as well as to calibrate the In addition, PFM operations must pay greater atten- size and complexity of an operation. It can also offer an tion to the intended results chain, and in many situations opportunity to discuss already at the outset potential points this involves some operational engagement on how funds at which it may be opportune to restructure a project, for move across levels of government to frontline units. Where example, if a window of opportunity might emerge following core systems are a full-blown bottleneck, this may be an election during the implementation of a project. premature, but in most situations, the flow of funds, as 6. We should have honest discussions about what has worked well as of performance information across the expen- elsewhere and to what extent, to clarify and jointly set realistic diture chain is important to consider to ensure that expectations. In the past, such dialogue would often largely operations actually enable an improved functioning of build on the number and types of cases that a team hap- governments. pened to be familiar with from past work. We are gradually (b) Investing in embedding already existing or recently estab- reaching a stage where there is an increasingly useful body lished systems versus breaking new ground: for the past of knowledge available that teams draw on in addition to 20 years, DP operations supporting PFM reforms have their own direct experience. This can be immensely helpful often focused on breaking new ground/ establishing new to be able to draw on a range of options from different coun- systems, such as introducing an MTEF, establishing (for tries, each of which may share certain characteristics with the first time) an FMIS, introducing new accounting the respective client country, for example, in terms of size, standards, or supporting the establishment of Internal level of development, and so on. Audit Units, promoting new procurement legislation 7. Operational design choices and organizational structures, and so on. As diagnosed (a) Broad or narrow, upstream or following the spending in this report, there has been a tendency to move from chain: any PFM reform operation, whether an IPF, one reform area to the next; while there has not always PforR, non-lending technical assistance, or (part of a been sufficient emphasis on ensuring that new systems DPL) needs to consider key design choices—whether are actually well utilized and that utilization is continu- to have a narrow or broader focus, for example, target- ous over time, rather than slipping backwards once ini- ing one or two versus multiple PFM reform aspects, tial support ends. Going forward, it will be important to 94  Political Economy of Public Financial Management Reforms invest far more in actual use of systems that have been Finally, there are several avenues for further analytic work established. While nearly all countries now have FMISs, to consider. This includes at least six areas, and potentially also PFM performance still varies widely, so in various areas, others: actual utilization is lacking. Similarly, legal provisions for procurement have been widely reformed as reflected • First, additional case studies exploring similar sets of actors would be highly desirable to further corroborate, refine, or among the five cases, but still much remains to be refute findings and recommendations set out here. Along done to ensure that new systems are fully established the same lines, it will be valuable to repeat the quantita- and that they are used as intended. Certain expansions tive analysis reflected in Chapter 2 as further observations in functionality may be very difficult given political become available. economy constraints, but it can be important to probe these boundaries, as often at least some further gradual • Second, extending the analysis to how the quality of PFM improvements are feasible. affects service delivery, and conversely how efforts to 8. Monitoring functional improvements. Related to the previous strengthen service delivery might drive PFM improvements points, it is critical to invest more in monitoring whether would be a valuable extension. systems are actually used as intended. The appetite of gov- • A third avenue for further work is to undertake a more in- ernments may vary in this regard, but especially incoming depth look at the relative role of the demand side. In this ‘reform governments’ and governments that have made regard, one strategy would be to scan for experiences where explicit anticorruption commitments to citizens or those that the demand side has had direct relevance and impact on face fiscal challenges, should in principle be keen to get a PFM reforms and to analyze how greater attention and grip on whether the systems that are in place actually func- impact emerged. tion as intended. Such monitoring can also be invaluable in guiding where further efforts should be focused. • Fourth, there is scope to analyze the specific engagement 9. Even if we don’t intend to engage more broadly on public sec- of development partners relative to the nature of nontechni- tor reforms, it almost always matters to understand at least cal context aspects in greater detail; for example, with a the basics of wider public sector issues, ongoing reforms, view to identifying what types of engagements have worked as well as dysfunctionalities, to really be able to target PFM better or worse in high and lower opportunity environments reforms well, and to anticipate various links and challenges. respectively. Public sector wages, problems with managing the payroll, • Fifth, as indicated in Chapter 1, this analysis has not explic- intergovernmental arrangements and other factors have a itly covered the specific change management tools used, but direct bearing on PFM functionalities—from more mun- has focused primarily on the level of commitment, factors dane issues such as being able to employ or contract com- influencing such commitment, and how this has affected petent IT staff, to more higher order issues such as whether reform efforts and progress made. It would be interesting to to have one comprehensive FMIS or separate systems for integrate this with a more explicit focus on the potential dif- national and subnational levels, or how to design program ference made by the intensity and choices made regarding or performance budgets in the context of ongoing wider explicit change management. performance management efforts, for example, led by an Office of the President or Prime Minister or a Ministry of • Finally, for each of the specific PFM reforms covered here, there is scope for further in-depth analysis of actual reform Public Service. progress and impact (also linking to point 2 regarding 10. Raising seeming ‘blind spots’ or risks with clients and with whether benefits for downstream service delivery materialize other DPs engaged on PFM where relevant. For example, in from PFM improvements). It would also be helpful to further situations where organic budget legislation is very outdated, develop methods and measures for (comparatively) assess- it can be important to bring up the issue of why this has not ing impacts. been addressed. In such a case, the point is not to quickly draft a new law, but rather to initiate a discussion, including with legislators as a crucial set of stakeholders. The World BAnk  95 References Adamu, S. A., and A. D. Ahmed. 2014. “IPSAS and Nigerian Pub- Andrews, M., L. Pritchett, and M. Woolcock. 2017. Building State lic Sector: The Challenges of First Time Adopters.” International Capability—Evidence, Analysis, Action. Oxford University Press. Journal of Social Sciences and Humanities Innovations 2 (1): Bain, K. A., D. Booth, and L. Wild. 2016. Doing Development 151–160. Differently at the World Bank: Updating the Plumbing to Fit the Akinyoade, A. and Uche, C.U. 2016. Dangote Cement: An African Architecture. London: Overseas Development Institute. Success Story? African Studies Centre Working Paper no. 131. Avail- Barma, N. H., Kaiser, K., Le, T. M., & Viñuela, L. 2012. “Rents to able at: https://openaccess.leidenuniv.nl/handle/1887/38674. Riches? The Political Economy of Natural Resource-Led Develop- Alesina, A., and S. Ardagna. 2010. “Large Changes in Fiscal Pol- ment.” Washington, DC: World Bank. icy: Taxes Versus Spending.” In Tax Policy and the Economy 24 Blum, J., N. Manning, and V. Srivastava. 2012. Public Sector (35–68). University of Chicago Press, Chicago. Management Reform: Toward a Problem-Solving Approach. Eco- Alesina, A., & Perotti, R. 1995. Fiscal expansions and adjust- nomic Premise Note No. 100. Washington, DC: World Bank. Pov- ments in OECD countries. Economic policy, 10(21), 205–248. erty Reduction and Economic Management Network. Allen, R. 2009. “The Challenges of Reforming Budgetary Institu- Bräutigam, D. A., and S. Knack. 2004. “Foreign Aid, Institutions, tions in Developing Countries.” Working Paper WP/09/96, Interna- and Governance in Sub-Saharan Africa.” Economic Development tional Monetary Fund. and Cultural Change 52 (2): 255–285. Allen, R., and F. Grigoli. 2012. “Enhancing the Capability of Cen- Brinkerhoff, D. W., and J. M. Brinkerhoff. 2015. “Public Sec- tral Finance Agencies.” World Bank-Economic Premise (73): 1–7. tor Management Reform in Developing Countries: Perspectives Beyond NPM Orthodoxy.” Public Administration and Development Allen, R., R. Hemming, and B. Potter, eds. 2013. The Interna- 35 (4): 222–237. tional Handbook of Public Financial Management. Springer. Bunse, S., and V. Fritz. 2012. “Making Public Sector Reforms Allen, R., Y. Hurcan, P. Murphy, M. Queyranne, and S. Yläoutinen. Work: Political and Economic Contexts, Incentives, and Strate- 2015. “The Evolving Functions and Organization of Finance Min- gies.” World Bank Policy Research Working Paper 6174. istries.” Working Paper WP/15/232, International Monetary Fund. Retrieved from: https://www.imf.org/external/pubs/ft/wp/2015/ Byrne, D., and C. C. Ragin (Eds.). 2009. The Sage Handbook of wp15232.pdf (December 12, 2016). Case-Based Methods. SAGE Publications. Allen, R., and Krause, P. 2013. “The Role, Responsibilities, Struc- Bysouth K. A., F. L. Philipsen, and C. R. Belisario. 2014. Mid ture and Evolution of Central Finance Agencies.” In The Inter- Term Review of the Philippines—Australia Public Financial Man- national Handbook of Public Financial Management (98–115). agement Program (PFMP). Final Review Report. Retrieved from: Palgrave Macmillan UK. https://dfat.gov.au/about-us/publications/Documents/philippines- aus tralia-public-financial-managem ent - pr ogr am - pf m p- Andrews, M. 2009. “Isomorphism and the Limits to African Public independent-evaluation-2014.pdf (December 12, 2016). Financial Management Reform.” HKS Working Paper No. RWP09- 012. Harvard Kennedy School, Cambridge, MA. Campos, J. E., and J. L. Syquia. 2005. “Managing the Politics of Reform: Overhauling the Legal Infrastructure of Public Procure- Andrews, M. 2013. The Limits of Institutional Reforms in Devel- ment in the Philippines.” Working Paper No. 70. Washington, DC, opment. Cambridge, MA: Cambridge University Press. World Bank.   97 Collier, D. 1993. “The Comparative Method.” In Political Science: at: http://www.internationalbudget.org/wp-content/uploads/ibp- The State of Discipline II, Ada W. Finifter, ed. American Political paper-how-civil-society-uses-budget-information-12-2016.pdf. Science Association. Dener, C., J. Watkins, and W.L. Dorotinsky. 2011. Financial Man- ———. 2011. “Understanding Process Tracing.” Political Sci- agement Information Systems: 25 Years of World Bank Experience ence and Politics 44 (4): 823–830. on What Works and What Doesn’t. World Bank Publications. Combaz, E. 2015. Implementing Integrated Financial Manage- Diamond, J. 2011. The Non-Technical Context of PFM Reforms. ment Information Systems GSDRC Helpdesk Research Report Available at: http://blog-pfm.imf.org/files/the-non-technical- 1229. Birmingham, United Kingdom: GSDRC, University of context-of-pfm-reform-2.pdf. Birmingham. ———. 2012. “Guidelines for Sequencing PFM Reforms.” Back- Cruz, C. and Keefer, P. 2010. “Programmatic Political Parties and ground Paper 1. Available at: http://www.pefa.org/sites/pefa.org/ Public Sector Reform” APSA Annual Meeting Paper. Available at files/guidelines-for-sequencing-pfm-reforms.pdf. SSRN: https://ssrn.com/abstract=1642962. ———. 2013. “Internal Control and Internal Audit.” In The Inter- Cruz, C., J. Labonne, and P. Querubin. 2014. “Politician Family national Handbook of Public Financial Management (374–395). Networks and Electoral Outcomes: Evidence from the Philippines.” Palgrave Macmillan UK. In Annual Meeting of the American Political Science Association. Dressel, B. 2012. “Targeting the Public Purse: Advocacy Coali- Cruz, C., P. Keefer, and C. Scartascini. 2016. “Database of tions and Public Finance in the Philippines.” Administration and Political Institutions Codebook, 2015 Update (DPI2015).” Inter-­ Society 0095399712460055. American Development Bank. Dressel, B., and J. Brumby. 2009. “Enhancing Capabilities of Curristine, T. 2005. “Government Performance: Lessons and Chal- Central Finance Agencies: From Diagnostics to Reform.” Frame- lenges.” OECD Journal on Budgeting 5 (1). work Paper, IBRD. Curristine, T., Z. Lonti, and I. Joumard. 2007. “Improving Public Fainboim Yaker, I., and S. Pattanayak, 2010. “Treasury Single Sector Efficiency: Challenges and Opportunities.” OECD Journal Account: Concept, Design, and Implementation Issues.” IMF on Budgeting 7 (1), 161–201. Working Papers 1–45. Dabla-Norris, E., R. Allen, L. F. Zanna, T. Prakash, and E. Kvin- Fritz, V., Kaiser, K. and Levy, B. 2009. Problem-Driven Gover- tradze, V. Lledo, and I. Yackovlev, and Gollwitzer, S. March 2010. nance and Political Economy Analysis—Good Practice Framework. “Budget Institutions and Fiscal Performance in Low-Income Washington, DC: World Bank. Countries.” IMF Working Papers, 1–56, 2010. Available at SSRN: Fritz, V., A. P. Fialho Lopez, E. Hedger, H. Tavakoli, and P. Krause. http://ssrn.com/abstract=1583338 2012. “Public Financial Management Reforms in Post-conflict De Gramont, D. 2014. “Beyond Magic Bullets in Governance Countries.” Washington, DC: World Bank. Reform.” Carnegie Endowment for International Peace. Retrieved Fritz, V., S. Sweet, and M. Verhoeven. 2014. “Strengthening Pub- from: http://carnegieendowment.org/files/governanace_magic_ lic Financial Management: Exploring Drivers and Effects.” World bullets.pdf (February 14, 2017). Bank Policy Research Working Paper 7084. De Renzio P., M. Andrews, and Z. Mills. 2011. “Does Donor Sup- ———. 2015. “What Drives Improvements in PFM Performance?” port to Public Financial Management Reforms in Developing International Monetary Fund PFM Blog. Available at: http://blog- Countries Work? An Analytical Study of Quantitative Cross-Country pfm.imf.org/pfmblog/2015/02/what-drives-improvements-in-pfm- Evidence.” ODI Working Paper 329. Overseas Development Insti- performance.html#more. tute. Available at: https://www.odi.org/sites/odi.org.uk/files/odi- assets/publications-opinion-files/7098.pdf. Frøystad, M., K. K. Heggstad, and O. H. Fjeldstad. 2010. “Linking Procurement and Political Economy.” U.K. Department for Inter- De Renzio, P., and W. Dorotinsky. 2007. Tracking Progress in the national Development and the World Bank Institute. Quality of PFM Systems in HIPCs. Washington, DC: The PEFA Secretariat. GIZ (German Agency for International Cooperation) and INTOSAI (The International Organisation of Supreme Audit Institutions). De Renzio, P., and M. Mastruzzi. 2016. How Does Civil Society Use 2013. Supreme Audit Institutions: Accountability for Develop- Budget Information? Mapping Fiscal Transparency Gaps and Needs ment. Baden-Baden: Nomos. in Developing Countries. International Budget Project. Available 98  Political Economy of Public Financial Management Reforms Global Witness. 2015. “Shell and Eni’s Misadventures in ­Nigeria— pemna-t-cop_fmis-study-of-selected-pemna-member_lessons-for- Shell and Eni at Risk of Losing Enormous Oil Block Acquired in other-countr....pdf (July 22, 2016). Corrupt Deal.” Retrieved from: https://www.globalwitness.org/ Keefer, P., and R. Vlaicu. 2008. “Democracy, Credibility, and Cli- en/campaigns/oil-gas-and-mining/shell-and-enis-misadventures- entelism.” Journal of Law, Economics, and Organization 24 (2): nigeria/ (February 15, 2017). 371–406. Government of Nepal. 2008. PEFA An Assessment of the Pub- Khan, M. H., and K. S. Jomo. 2000. “Rents, Rent-Seeking and lic Financial Management Performance Measurement Framework. Economic Development: Theory and Evidence in Asia.” Cambridge Kathmandu. University Press. ———. 2015. Public Expenditure and Financial Accountabil- Khan, A., and S. Mayes. 2009. “Transition to Accrual Account- ity (PEFA) Assessment Nepal PFM Performance Assessment II. ing.” IMF; Technical Notes and Manuals No. 2009/02. Retrieved Kathmandu. from: https://www.imf.org/external/pubs/ft/tnm/2009/tnm0902 Gray, H. and Khan, M. 2010. Good Governance and Growth in .pdf (February 15, 2017). Africa: What Can We Learn from Tanzania? In: Padayachee, V. (ed) Khemani, P., and M. J. Diamond. 2005. “Introducing Financial The Political Economy of Africa. London and New York: Routledge. Management Information Systems in Developing Countries.” Green, D. 2016. How Change Happens. Oxford University Press. Working Paper WP/05/196. International Monetary Fund. Hallerberg, M., C. Scartascini, and E. Stein. 2009. “Who Decides Killick, T. 2005. “The Politics of Ghana’s Budgetary System.” the Budget.” A Political Economy Analysis of the Budget Process CDD/ODI Policy Brief 2. in Latin America. Inter-American Development Bank. Kingdon, J. W. 1995. “The Policy Window, and Joining the Hashim, A., and M. Piatti. 2016. A Diagnostic Framework to Streams.” Agendas, Alternatives, and Public Policies 172–189. Assess the Capacity of a Government’s Financial Management Kochanova, M. A., and C. Caceres. 2012. “Country Stress Events: Information System as a Budget Management Tool. Retrieved from: Does Governance Matter?” (No. 12-116). International Monetary https://ieg.worldbankgroup.org/Data/reports/IFMISDiagnostic Fund. Framework.pdf (December 12, 2016). Krause, P. 2009. “A Leaner, Meaner Guardian?: A Qualitative Holmes, M., and A. Evans. 2003. A Review of Experience in Comparative Analysis of Executive Control Over Public Spend- Implementing Medium-Term Expenditure Frameworks in a PRSP ing.” Discussion Paper/Deutsches Institut für Entwicklungspolitik; Context: A Synthesis of Eight Country Studies. London: Overseas 22/2009: Bonn. Development Institute. ———. 2013. “Of Institutions and Butterflies: Is Isomorphism Hudson, A., and C. Wren. 2007. Parliamentary Strengthening in Developing Countries Necessarily a Bad Thing?” Background in Developing Countries. Final Report for DFID. Retrieved from: Note. Overseas Development Institute, London, April. Retrieved https://www.odi.org/sites/odi.org.uk/files/odi-assets/publications- - from: http://www.die-gdi.de/uploads/media/DP_22.2009.pdf (Decem­ opinion-files/128.pdf (December 12, 2016). ber 12, 2016) IMF (International Monetary Fund). 2014. Government Finance ———, Sweet, S., Hedger, E., Chalise, B. 2013. Operational risk Statistics Manual 2014. Washington, DC: International Monetary assessment of public financial management reform in Nepal: a Fund. review of challenges and opportunities. London: Overseas Develop- ———. 2015. Philippines Fiscal Transparency Evaluation. IMF ment Institute. Country Report No. 15/156. Retrieved from: https://www.imf.org/ Kuipers, B. S., M. Higgs, W. J. M. Kickert, L. Tummers, J. Grandia, external/pubs/ft/scr/2015/cr15156.pdf (December 12, 2016). and J. Van der Voet. 2014. “Managing Change in Public Organiza- ———. 2017. World Economic Outlook, April 2017: Gaining tions: A Review of the Literature (2000–2010).” Public Adminis- Momentum? World Economic and Financial Surveys. Washington, tration 92 (1): 1–20. DC: International Monetary Fund. Lewis, P., and M. Watts. 2015. Nigeria: The Political Economy of Joshi, Suhas, Arun K. Srivastava, and Minh Van Nguyen. 2015. Governance. “FMIS: A Study of Selected PEMNA Members: Lessons for Other Levy, B. 2014. Working with the Grain: Integrating Governance Countries.” Public Expenditure Management Network in Asia and Growth in Development Strategies. Oxford University Press. (PEMNA), IMF. Retrieved from: http://blog-pfm.imf.org/files/ The World BAnk  99 Lienert, I. 2012. “Preconditions for Programme and Performance- sites/resultsfordevelopment.org/files/resources/Ngozi-CSO_Paper_ based Budget Reform.” In Budgeting for Results: Moving Towards Revised_Version.pdf (December 12, 2016). Performance Budgeting. Pretoria: CABRI. Oyugi, L. 2008. “Experiences with Medium-Term Expenditure ———. 2013a. The Legal Framework for Public Finances and Framework in Selected Southern and Eastern African Countries.” Budget Systems. In The International Handbook of Public Finan- SEAPREN Working Paper 7, Southern and Eastern Africa Policy cial Management (63–83). Palgrave Macmillan UK. Research Network. ———. 2013b. Role of the Legislature in Budget Processes. In Padayachee, V, ed. 2010. The Political Economy of Africa. The International Handbook of Public Financial Management Routledge. (116–136). Palgrave Macmillan UK. Pant, B. and B. Subedi. 2006. “The HIPC Initiative, MDRI and Lijphart, A. 1971. “Comparative Politics and the Comparative Nepal: A Re-examination.” Economic Review 20. Kathmandu: Method.” American Political Science Review 65 (03), 682–693. Nepal Rastra Bank. Poplawski-Ribeiro, M., T. Baunsgaard, M. Villafuerte, and C. Rich- Peterson, S. 2006. “Automating Public Financial Management in mond. 2012. “Fiscal Frameworks for Resource Rich Developing Developing Countries.” John F. Kennedy School of Government Countries.” International Monetary Fund. Working Paper No. RWP06-043. Matsuda, Y. 2014. “Strengthening Local Service Delivery in the Pretorius, C. and N. Pretorius, 2008. Review of Public Financial Philippines: The Use of Political Economy to Craft Bank Opera- Management Reform Literature. London: DFID. tional Strategies.” In The World Bank’s Experience 233. Prichard, W., and D. K. Leonard. 2010. “Does Reliance on Tax Moore, M. 2004. “Revenues, State Formation, and the Quality of Revenue Build State Capacity in Sub-Saharan Africa?” Interna- Governance in Developing Countries.” International Political Sci- tional Review of Administrative Sciences, 76 (4): 653–675. ence Review 25 (3): 297–319. Pritchett, L., M. Woolcock, and M. Andrews. 2010. “Capability Moretti D., J. Cavanagh, and S. Flynn. 2016. “Implementing Traps? The Mechanisms of Persistent Implementation Failure.” Accrual Accounting in the Public Sector.” IMF; Technical Notes Center for Global Development Working Paper (234). and Manuals No. 2016/06. Retrieved from: http://www.imf.org/ Rakner, L., L. Mukubvu, N. Ngwira, K. Smiddy, and A. Schneider. en/Publications/TNM/Issues/2016/12/31/Implementing-Accrual- 2004. The Budget as Theatre: the Formal and Informal Institu- Accounting-in-the-Public-Sector-44121 (February 15, 2017). tional Makings of the Budget Process in Malawi. DFID Report. Moynihan, D., and I. Beazley. 2016. Toward Next-Generation Robinson, M. 2009. The Politics of Successful Governance Performance Budgeting: Lessons from the Experiences of Seven Reforms. Routledge. Reforming Countries. Washington, DC: World Bank. Robinson, M. and D. Last. 2009. “A Basic Model of Performance- North, D. C., J. J. Wallis, S. B. Webb, and B. R. Weingast, eds. based Budgeting.” Washington, DC: International Monetary Fund. 2013. In the Shadow of Violence: Politics, Economics, and the Problems of Development. Cambridge University Press. Roe, A. 2016. “Tanzania—from Mining to Oil and Gas.” Wider Working Paper 2016/79. Helsinki: UNU-Wider. Norton, A., and D. Elson. 2002. What’s Behind the Budget?: Poli- tics, Rights and Accountability in the Budget Process. Overseas SADCOPAC (Southern Africa Development Community Organ- Development Institute. isation of Public Accounts Committees). 2012.“Medium Terms Expenditure Frameworks: International and Regional Experience.” OECD (Organisation for Economic Co-operation and Development). Reflection and Discussion Paper for the Budget Analysis Group. 2011. Good Practices in Supporting Supreme Audit Institutions. In Enhancing Policy Dialogue in the Context of Poverty Reduction Paris: OECD. Available at: https://www.oecd.org/dac/effectiveness/ Budget Support to Mozambique. Retrieved from http://sadcopac Final%20SAI%20Good%20Practice%20Note.pdf. .pac-networks.org/file/view/MTEFs+050312.pdf (July 22, 2016). Okonjo-Iweala, N. 2012. Reforming the Unreformable: Lessons Sanchez, A. 2013. “The Role of Procurement.” In The Interna- from Nigeria. MIT Press. tional Handbook of Public Financial Management (312–335). Pal- Okonjo-Iweala, N., and Osafo-Kwako, P. 2013. “The Role of grave Macmillan UK. Civil Society Organizations in Supporting Fiscal Transparency Schick, A. 1998. “Why Most Countries Should Not Try New in African Countries.” Results for Development Institute Draft Zealand’s Reforms.” The World Bank Research Observer 13 Paper. Retrieved from: http://www.resultsfordevelopment.org/ (1):123–31. 100  Political Economy of Public Financial Management Reforms Shelley, L., Scott, E. and Latta, A. 2007. Organized Crime and Leadership and a Bureaucratic Organizational Structure,” Euro- Corruption in Georgia. New York, NY: Routledge. pean Management Journal, 32 (3): 373–382. Smoke, Paul. 2015a. “Rethinking Decentralization: Assessing Von Hagen, J., and I. Harden. 1994. “National Budget Processes Challenges to a Popular Public Sector Reform” Public Administra- and Fiscal Performance.” European Economy Reports and Studies tion and Development 35 (2): 97–112. 3 (1994): 311–418. ———. 2015b. “Managing Public Sector Decentralization in Wehner, J. 2006. “Assessing the Power of the Purse: an Index Developing Countries: Moving Beyond Conventional Recipes” Pub- of Legislative Budget Institutions.” Political Studies 54 (4). lic Administration and Development 35 (4): 250–262. 767–785. ISSN 0032-3217. Retrieved from: http://eprints.lse .ac.uk/15564/1/Wehner_Assessing-the-power-of-the-purse_2006 Stalwart Report. 2015. Understanding the Treasury Single Account .pdf (December 12, 2016). (TSA) System—Things You Should Know. The Stalwart Report. Retrieved from: http://www.systemspecs.com.ng/understanding- Wehner, J. and de Renzio, P. 2013. “Designing Fiscal Institutions: the-treasury-single-account-tsa-system-things-you-should-know/ The Political Economy of PFM Reforms.” In The International (September 4, 2017). Handbook of Public Financial Management. Palgrave McMillan. Subedi, M. 2014. “Afno Manchhe: Unequal Access to Public Welham, B., and S. Hadley. 2015. The Capabilities of the Sierra Resources and Institutions in Nepal.” Dhaulagiri Journal of Soci- Leone Ministry of Finance and Economic Development. London: ology and Anthropology 8: 55–86. Overseas Development Institute. Tamang, S., and C. Malena. 2011. The Political Economy of Social Wildavsky, A. B. 1986. “Budgeting: a Comparative Theory of the Accountability in Nepal. Kathmandu: PRAN Unpublished Report. Budgeting Process.” Transaction Publishers. Retrieved from: https://openknowledge.worldbank.org/bitstream/ Wilhelm, V. A., and P. Krause. 2008. Minding the Gaps: Inte- handle/10986/17661/840720WP0REPLA0Box0same0box0 grating Poverty Reduction Strategies and Budgets for Domestic PUBLIC0.pdf?sequence=1&isAllowed=y (December 12, 2016). Accountability. Washington, DC: World Bank. Therkildsen, O. 2000. “Public Sector Reform in a Poor, Aid- Williams-Elegbe, S. 2015. “A Comparative Analysis of Public Pro- dependent Country, Tanzania.” Public Administration and Devel- curement Reforms in Africa: Challenges and Prospects.” Upphan- opment 20 (1): 61. dlingsrättslig Tidskrift 1: 11–32. Therkildsen, O., and United Nations Research Institute for Social World Bank and EU (European Union). 2008. Joint World Bank- Development. 2001. Efficiency, Accountability and Implementa- European Commission Public Financial Management Assessment. tion: Public Sector Reform in East and Southern Africa. Geneva: United Nations Research Institute for Social Development. World Bank. 2004. Report on the Observance of Standards and Codes (ROSC)—Corporate Governance Country Assessment Jor- Tilley, H. 2013. Unblocking Results Case Study: Rural Water in dan. Washington, DC: World Bank. Tanzania. London: ODI May. http://www.odi.org.uk/sites/odi.org .uk/files/odi-assets/publications-opinion-files/8411.pdf. ———. 2010. Tanzania: 2009 Public Financial Management Perfor- mance Report on Mainland Tanzania. Washington, DC: World Bank. Tusalem, R. F., and J. J. Pe-Aguirre. 2013. “The Effect of Political Dynasties on Effective Democratic Governance: Evidence from the ———. 2012. The World Bank’s Approach to Public Sector Man- Philippines.” Asian Politics and Policy 5 (3), 359–386. agement 2011–2020: Better Results from Public Sector Institu- tions. Washington, DC: World Bank, Public Sector and Governance Vani, S. 2010. “Internal Audit in the Public Sector: Under- Board, February 3. developed and Underused” IMF PFM blog; available at: http://blog-pfm.imf.org/pfmblog/2010/08/posted-by-sanjay- ———. 2013a. Beyond the Annual Budget. Washington, DC: World vani---of-107-countries-in-which-public-expenditure-and- Bank. Retrieved from http://www.imf.org/external/np/seminars/ financial-accountability-pefa-assessments-have-bee/comments/. eng/2013/fiscalpolicy/pdf/brumby.pdf (July 22, 2016). ———. 2012. “Has Global PFM Improved in the Last Decade?” ———. 2013b. Georgia: PEFA Assessment 2012. Washington, IMF PFM blog; available at: http://blog-pfm.imf.org/pfmblog/ DC: World Bank. 2012/09/has-global-pfm-improved-in-the-last-decade.html. ———. 2013c. Transforming Central Finance Agencies in Poor Van der Voet, J. 2014. “The Effectiveness and Specificity of Countries: A Political Economy Approach. Washington, DC.: Change Management in a Public Organization: Transformational World Bank. https://openknowledge.worldbank.org/handle/10986/ 15806. The World BAnk  101 ———. 2013d. Tanzania: Public Expenditure and Financial ———. 2017. World Development Report 2017: Governance and Accountability Assessment: Mainland Tanzania. Washington, DC.: the Law. Washington, DC: World Bank. World Bank. ———. Forthcoming. “Lessons from FMIS Implementation. A ———. 2015. Managing Change in PFM System Reforms: A Synthesis of Evaluation Findings.” Washington, DC: World Bank. Guide for Practitioners. Washington, DC: World Bank. Avail- Worthington, R. and I. Lienert. 2013. Performance and Programme-­ able at: http://siteresources.worldbank.org/PUBLICSECTORAND based Budgeting in Africa: A Status Report. Pretoria: CABRI. GOVERNANCE/Resources/285741-1303321730709/FMIS_ Managing_Change_May2015_v1.1.pdf. You, J. S. 2014. Land Reform, Inequality, and Corruption: A Com- parative Historical Study of Korea, Taiwan, and the Philippines. ———. 2016. PEFA 2016: Framework for Assessing Public Finan- Retrieved from http://piketty.pse.ens.fr/files/You2014.pdf (Febru- cial Management. Retrieved from https://pefa.org/sites/default/ ary 14, 2017) files/PEFA%20Framework_English.pdf (July 22, 2016). 102  Political Economy of Public Financial Management Reforms Annex Data Source and Methodology 1 Variable Description Data Source GDP per capita GDP per capita, PPP (constant 2011 international $) WDI database Growth (per capita) GDP per capita growth (annual %) WDI database Population Population, total WDI database SIDS United Nations Educational, Scientific, and Cultural UNESCO website Organization (UNESCO) SIDS Member States and UNESCO SIDS Associate Members  Resource Resource rents % exports; Resource rents % total revenue. Poplawski-Ribeiro, Marcos, et al. 2012 Dummy variable for countries that rely heavily on oil/mineral revenue (>20% of total revenue) Tax Tax revenue as a percentage of GDP IMF FAD’s database Aid Net ODA received (% of GNI) WDI database Growth shock The total number of years in the past five years where per WDI database capita GDP growth has declined by 2 percentage points or more Fiscal shock The total number of years in the past five years where IMF World Economic Indicators database government primary balance proportion of GDP has declined See also Kochanova and Carceres 2012 by 1 percentage points or more Political stability/fragility Political stability World Bank’s Worldwide Governance Indicators Regime Country ratings from 1995 to 2016 Freedom House Programmatic parties GOV1RLC; GOV2RLC; GOV3RLC; OPP1RLC Cruz, Keefer, and Scartascini (2016) "Database of Political Institutions Codebook.” 2015 Update (DPI2015)   103 Annex OLS and FD Robustness Checks 2 (1) (2) (3) (4) (5)   PEFA PEFA PEFA PEFA CPIA-13 GDP per capita (log) 0.2378*** 0.2386*** 0.2388*** 0.2374*** 0.2586*** (0.0403) (0.0399) (0.0398) (0.0402) (0.0402) Growth (per capita) 0.0276** 0.0277** 0.0300** 0.0263** 0.0434*** (0.0115) (0.0115) (0.0119) (0.0119) (0.0133) Population (log) 0.0411* 0.0418* 0.0433* 0.0402* 0.0744*** (0.0232) (0.0235) (0.0235) (0.0234) (0.0257) SIDS −0.2223* −0.2215* −0.2213* −0.2188* −0.0509 (0.1225) (0.1226) (0.1220) (0.1235) (0.1146) Resource −0.3386*** −0.3412*** −0.3378*** −0.3369*** −0.3773*** (0.0911) (0.0886) (0.0895) (0.0898) (0.0973) Fiscal shocks 0.0032 (0.0320) Growth shock (Lagged 4 years) 0.1488 (0.0975) Growth shock (Lagged 5 years) -0.0921 (0.1459) Observations 262 262 262 262 229 R-squared 0.34 0.34 0.345 0.342 0.272 Note: Standard errors in parentheses, clustered at the country level. Observations weighted inversely to number of PEFA assessments. No influential observation was dropped. *** p < 0.01, ** p < 0.05, * p < 0.1   105 (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES AVG_D_CPIA-13 AVG_D_CPIA-13 AVG_D_CPIA-13 AVG_D_CPIA-13 AVG_D_CPIA-13 AVG_D_CPIA-13 AVG_D_CPIA-13 AVG_D_CPIA-13 GDP per capita (% change) 0.1132 0.3807* 0.3692*** 0.5166*** 0.5298** 0.4550*** 0.3624* 0.2603** (0.2275) (0.2001) (0.1220) (0.1748) (0.2514) (0.1221) (0.2066) (0.1187) Population (% change) −0.1903 −1.4343*** −0.8809* −1.3223*** −1.3131*** −1.4219*** −1.4068*** -1.3775*** (0.4217) (0.4032) (0.4981) (0.3922) (0.4508) (0.4060) (0.4036) (0.4202) Resource (dummy) −0.0294** −0.0276** −0.0312*** −0.0298*** −0.0318*** −0.0245** −0.0231** -0.0301** (0.0127) (0.0116) (0.0114) (0.0109) (0.0115) (0.0111) (0.0113) (0.0119) SIDS (dummy) −0.0244** −0.0301*** −0.0307*** −0.0343*** −0.0281** −0.0256** −0.0279** -0.0312*** (0.0111) (0.0114) (0.0115) (0.0107) (0.0123) (0.0112) (0.0132) (0.0117) Initial CPIA-13 −0.0573*** −0.0572*** −0.0589*** −0.0585*** −0.0604*** −0.0575*** -0.0513*** (0.0074) (0.0083) (0.0071) (0.0076) (0.0073) (0.0071) (0.0085) Initial GDP per capita (level in log) 0.0084 (0.0060) Initial regime type −0.0015 (0.0020) Tax (% point change) 0.0115 (0.0087) Aid (ODA) (% point change) 0.0088 (0.0067) 106  Political Economy of Public Financial Management Reforms Regime (Freedom House, % change) −0.0656*** (0.0228) Programmatic parties (% change) −0.1345 (0.1759) Political stability (% change) 0.0016 Observations 132 132 128 119 121 128 117 130 R-squared 0.073 0.369 0.394 0.443 0.411 0.408 0.382 0.354 Note: Robust standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1 Annex Framework Used for Developing the Country Analysis 3 I.  Introduction: Overall Aim economic power holders and between these and citizens, insti- tutional legacies and the nature of the bureaucracy, as well as and Key Assumptions interactions with the international community, and the macro- This case study protocol was used to guide the country assessment of economic and fiscal situation (and its impacts on the political non-technical drivers and how they have influenced PFM reforms. The calculus). protocol helped to facilitate case studies for the global study on PE drivers in PFM. Where suitable, the framework could also be used to generate a background note for a Public Expenditure Review II. Scope (PER) or other PFM assessments. In terms of focus, case studies should seek to do the following: The PE of PFM case studies should address both why reform (a) embed the discussion of PE aspects of PFM reforms in key progress in certain areas has been possible, and in others not, as well features of the wider context as outlined in Section III below, as how robust implemented reforms have been in the light of (likely) (b) develop an overview of the status of the PFM system in terms continuous PE challenges. PFM reforms are inspired by the belief of key features as per Table 1; (c) hone in on two to three spe- that budgets (how funds are allocated and managed) play a crucial cific PFM features/reform areas, typically selecting at least one role in driving wider change in public sector performance (and (more) successful and one unsuccessful area, and trace underly- in turn wider development results), and that governments have a ing drivers and reform processes in ways that capture incentives general incentive to get the most out of available funds (that is, of key stakeholders. For some cases, a ‘wholesale reform’ such as to seek allocative and technical efficiency). However, especially the adoption of a new organic budget law may be an appropriate the latter of these assumptions may not hold true as presumed. focus—but this should then be ‘dissected’ in terms of specific For example, seeking additional aid may be a politically more areas covered by the law. The selection of a focus should be made attractive strategy than pursuing cost savings through improved in consultation with the core study team and the country team/ technical efficiency and strengthened controls against leakages public sector experts working on a particular country. which require significant mobilization of political capital to suc- Case studies should combine a retrospective and a prospective ceed. Nonetheless, numerous governments pursue at least some view: explaining recent PFM system performance and reform efforts PFM reforms—even if this can be considered irrational. A cogent (of the past three to five years),196 and identifying implications for analysis therefore needs to be able to account for problems with the upcoming period (circa 12 months to 24 months). Depending PFM reforms as well as for successes. on the issues at hand, forward looking recommendations should Methodologically, it is not expected that one single political focus on further discreet reforms, or the full implementation of economy driver—such as the centralization of powers over spend- existing reforms, or ensuring that existing reforms actually develop ing, or the nature of the party system—is the dominant explanatory their intended intermediate and final outcomes and impacts (see variable across countries, or that all countries broadly classifiable as Graph  1). Grounded in the analytic evidence, they may caution clientelist or Limited Access Orders display a similar quality of their against certain actions (for example, because they are not feasible PFM systems. While many cases are likely to be broadly classifiable or not likely to achieve hoped for outcomes due to constraints), as limited access orders (in contrast to open access orders), and as well as propose adjustments, mitigating measures, or proactive as displaying some degree of clientelism, they may still vary con- engagement to stretch opportunities for reforms and their impacts. siderably in terms of the quality of their PFM systems. Rather, it is expected that countries display a variety of constellations among 196. A longer time-frame can be selected if appropriate and feasible, and a number of factors, including constellations among political and while still maintaining also a forward looking focus.   107 Graph 1:  PFM Results Framework Case studies: Country choices and experiences of PFM reforms: reform outputs, intermediate and final outcomes achieved (with a focus on selected PFM results); and mapping of underlying stakeholders, incentives, and dynamics; with a view to develop (a) a systematic as well as practical way of mapping PE factors relevant for specific areas of reform and relevant ‘monitoring indicators’; (b) guidance on whether and how typical obstacles, pitfalls as well as opportunities can be addressed proactively. Inputs to PFM Intermediate Final Reforms Outputs Outcomes Outcomes Impacts Government Changes in laws, Transparency and Fiscal Improvements in inputs, including rules, and comprehensiveness discipline state capacity institutional setup procedures for PFM reform Links to policy, planning, Strategic Improvements in coordination Improved and delivery allocation of service delivery information resources DP support to PFM systems and Control, oversight, and Improved reform efforts business accountability Operational management of processes efficiency in budget support Harmonization and Cost-savings (reduction in public alignment of DP Changes in idle balances, short-term spending support people, skills, and borrowing) organizations Complementary DP inputs Changes in incentives and controls Source: Adapted from Lawson and de Renzio (2009). Section III discusses the questions to be addressed in further While a perspective on stakeholder incentives is particularly criti- detail and outlines the evidence to be collected. It also indicates cal for (3), it should also form part of the storyline throughout. what issues should be clarified in an initial pre-mission report. It is 1.  Country Overview important to establish key institutional features and reform prog- 1a.  Overall Country Fiscal Dynamics and Trends ress before the mission, so that data and information collection What has been the overall level of spending (absolute and in per- during a mission can concentrate on addressing questions about centage of GDP) over the past five years and what dynamics are motivation and incentives (that is, the underlying drivers). Gen- expected to take place over the coming 5–10 years? erally, all factual questions should be addressed, while analytic What is the level of revenue (absolute and as a percentage questions are suggestive and should be selected depending on the of GDP), and what is the share of tax revenue? For resource-rich specific focus selected and the country situation. countries: what is the percentage of natural resource revenue (or Section IV provides an outline of the final report to be produced; other windfall revenue)197 relative to total revenue (indicative if and the reference section contains key materials that teams con- exact data cannot be found)? sult prior to embarking on the work (all references are available Declining or expanding revenues are a decisive impulse for from the core team). politicians. Across countries, we also want to understand better whether reforms are likely to happen during fiscal expansion or III.  Issues to Be Covered contraction periods. Questions to be addressed by case studies are grouped as fol- To be covered in an initial desk-based report,with possible lows: (1) key fiscal and political country features relevant for PFM additional information to be collected subsequently in country. reforms, (2) PFM system status and overall reform agenda, (3) spe- cific PFM reform aspects and underlying PE drivers, and (4) guid- ance to help inform an overall story line and recommendations. 197. For example, payments for over-flight rights, fishing rights, internet domains, and so on. 108  Political Economy of Public Financial Management Reforms What are the ratings for PEFA indicators PI-1 and PI-2 for How important is success against these goals for ‘political sur- recent years (or calculate these using the PEFA methodology)— vival’/re-elections/political legitimacy? that is, aggregate fiscal discipline and variance in expenditure What is the nature of the political party system in the coun- composition at the broad functional level? Are there fluctuations try? (are parties institutionalized and do they play a role? If not, in PI-1 (or 2) with the electoral cycle (such as pre-election over- how are interests aggregated?) How stable/unstable is the party spending?) Have recent actual fiscal deficits been close to ex ante system, and how fragmented or concentrated is it? Are there pro- planned/approved deficits? grammatic parties or are parties largely personality driven (Keefer Whatever the given resource envelope is, are governments and Vlaicu 2008)? managing within that or not; and are they keeping sector alloca- How important are illicit and oligarchic interests in the tions in line with ex ante allocations or not? country (and are they separate from or part of mainstream politi- cal parties)? (such as drug cartels [illicit], but also insiders who Existing PEFA data to be covered in an initial desk-based may have benefited from privatization and similar deals and who report;most recent years to be covered in country. dominate the economic sphere) What types of demands with regard to service delivery (and What is the country’s level of public debt (absolute and in possibly state reform) are being raised by civil society, organized percentage of GDP) and the dynamic of debt accumulation? Has labor or other ‘demand side’ actors, if any? Is there effective the country received debt relief in recent years? (and if so, what ‘voice’? has been the pace of renewed debt built-up?) What is the level of aid dependence (in percentage of GDP Descriptive information should be included in the desk- and percentage of annual total public spending)? What percentage based report based on publicly available information and of aid is reported on budget versus off-budget? existing analytic work;the assessment of the government’s Are PEFA scores and broader indicators of state capacity and general level of commitment should result from discussions in integrity aligned (all rather good, rather mediocre, or rather poor) country. or are they pointing in different directions (for example, Afghani- 2.  PFM System Features stan has good PEFA scores but nonetheless high state instability 2a. Central Finance Agency Responsibilities, Authority, Capabilities, and high levels of perceived corruption)? and Relationships with Other Key Institutions198 To be covered in the desk-based report, based on existing How are the Central Finance Agency (CFA) functions organized? data and information sources (for aid: OECD aid statistics Is there a unified ministry (comprising finance, planning, treasury and monitoring reports + other sources if appropriate); for functions)? Or are multiple ministries and departments respon- state capacity and integrity: WGI government effective- sible for CFA functions? What is the formal administrative weight ness and Control of Corruption scores, supplemented of the ministry of finance relative to other ministries? (refer also to with BTI and other sources as appropriate. Brumby and Dressel 2009) What is the formal and informal rela- tionship between institutions covering CFA functions (or among 1b.  Country-level Political Trends major departments within a unified MoF, especially if these enjoy Recent governments (going back 5–10 years) and expected a degree of autonomy)? upcoming changes in government if any (brief description and Are there laws/regulations covering all aspects of the PFM characterization of governing parties and key leaders) system? What are the key features and what is the quality of the Brief overview of the intergovernmental structure. Is the sys- legal framework governing PFM—in terms of being clear, cover- tem of government centralized, undergoing a process of decentral- ing all relevant aspects (budget preparation process, budget ization, or decentralized? What is the balance of power between execution, reporting and accounting; covering public investment the center and the regions, and is that stable or shifting? management as well as recurrent expenditures); what administra- Is there a clear system for dividing up spending to regions tive model and time period is the legal framework related to (for and if so, what rules is it based on (for example, population, exist- example, traditional francophone versus post-LOLF Francophone; ing infrastructure)? Are there ad hoc decisions made (in general 1950s Westminster/anglophone, and so on) or in addition to regular allocations)? If the case is a resource-rich Are there significant off-budget allocations, such as via spe- country, how are revenues from royalties and other main resource- cial funds that are off budget, SOEs, and so on? Are these under related taxes divided? some supervision/monitoring by CFAs or removed from comprehen- How committed is the government to related goals such as sive oversight? improving overall governance and/or improving service delivery? 198. This section draws on Brumby and Dressel (2009). The World BAnk  109 What are the major trends of the country’s PFM system in last Basic information on the organization of CFA functions and of the legal framework and its institutional legacies 3–5 years (up to 7 years if relevant)?199 For countries with repeat should be covered in the desk-based report. PEFA assessments: what areas have improved versus deteriorated? Is budget accountability relatively good or rather weak? If the If recurrent and planning functions are both housed in an latter, is strengthening accountability part of the reform agenda? Is MoF: how small/large is the unit working on public investment it compatible with the way in which the overall system of govern- planning (staffing number of public investment unit relative to ment operates? overall staff of the MoF)? PEFA indicators related to external audit and parliamentary What is known about the quality of staff in the MoF (and follow tend to be the lowest scored, so limitations in this area are other CFAs if separate)? Is it a high-quality, relatively stable cadre? common. At the same time, donors are devoting further resources Or are there major issues with turnover, and/or with quality? What to strengthening SAI capacity—but there is as yet some uncer- is the extent of political appointees (formally and de facto)? tainty over whether and where this can succeed. This point should be evidenced with data from the MoF’s HR To what extent are laws on the books (including recent department on turnover rates and if possible, data on the number reforms) actually being implemented? (for example, fiscal respon- of degree holders among core ministry staff (that is, excluding tax sibility legislation, budget calendar prescriptions . . .) and customs departments) relative to total core staff. Actual extent to which recent reform plans have been How does the MoF (and other key CFA functions if separate) implemented relate to key stakeholders: the top executive, parliament, citizens, regions/subnational levels of government, sector ministries/depart- For the desk-based review, most recent PEFA assess- ments, and donors? (reporting and accountability relationships; ments should be used to compare the country to regional and income group peers (data for country groupings is see also Brumby and Dressel 2009; Dabla-Norris et al. 2010). available from the core study team). It should be noted What is the nature of the relationship between the MoF and if the PEFA is more than three years old (and significant line ministries during budget planning, execution, and reporting? changes are likely to have taken place in the meantime). What areas are in greatest need of improvement (from the perspec- Existing reports and materials (PERs, IMF FAD reports, tive of MoF vs the perspective of line ministries)? OBI assessments, etc.) should be used for a preliminary Does parliament play a significant role: (i) with regard to filling in of Table 1. More detailed information, in particular on recent/ongoing reform efforts is expected to be collected in shaping the budget ex ante and (ii) reviewing budget execution country. ex post? (see also Wehner 2006) Are there thresholds for changes during budget implementation that require parliamentary approval 3.  PE Aspects of PFM Reforms (virement limits, approval for supplementary budgets)? Are sup- 3a.  Overall Reform Agenda Setting, Stakeholders, and Commitments plementary budgets frequently used? Has the government made PFM reforms a central part of its What are institutional veto points and other veto opportuni- reform agenda (for example, as set out in a national development ties to block or derail PFM reforms? strategy)—within the past 5–7 years, and currently? ­ There are many assumptions in the literature that a more Has it set out any specific PFM reform plan or just a general centralized MoF and a limited, more oversight oriented role of commitment (for example, as part of a general commitment to parliament are more propitious for effective reforms [for example, improving governance or strengthening the public sector)? Diamond 2011]; case studies should establish the degree of cen- Does the President/Prime Minister ever talk publicly about tralization and relative ‘autonomy’. the need to reform public financial management? Do the main political parties or specific politicians hold Formal rules should be reviewed and mapped for the ini- views on PFM reforms? Are these publicly announced/known (for tial desk-based report. Specific information and information on informal rules will need to be collected in country. example, in speeches, party platforms), or mostly discussed only within government circles/between politicians and senior bureau- 2b.  PFM Status and Reform Areas crats? Do opposition parties have any particular views? What is the status quo of the country’s PFM system and reform Who have been the ministers of finance for the past 5–7 years efforts (using Table 1)? What are key strengths/areas of good (and duration of service)? What is their background (more techni- performance, and areas of poorer performance (drawing on most cal, more political, domestic/international)? Are ministers seen as recent PEFA assessments and other analytic reports)? How does the country compare to its regional and income group (based on 199. This can include time-frame/perspective up to 10 years if relevant most recent PEFA assessment)? and feasible. 110  Political Economy of Public Financial Management Reforms Table 1:  Main PFM Features and Reform Areas Status Agenda Decision Implementation Part of Under Government Approved Implementation Is the Reform Reform (by whom (since when and Having Recommended (y/n) Commitment (y/n) and led by which Fully Expected (by whom?) (and when?) when?) agency?) Implemented Effects? Cross-Cutting Extension of the comprehensiveness of the budget (that is, covering ODA flows? SOEs? and so on) Budget Planning Basic: ensuring regularization of the annual budget process Clarity/completeness/relevance of the budget classification and correspondence with the Chart of Accounts in use MTFF/EF/BF Program budgeting (and performance management) Budget Execution Basic: restoring cash management and reducing/eliminating arrears Introduction of TSA Reform of Chart of Accounts Improving the regularity, timeliness and informational value of reporting (between spending agencies and CFAs) Introduction of FMIS components or of full-scale IFMIS Strengthening of internal controls and audit Budget Accountability Improving transparency—various criteria (timeliness, level of detail, accessibility, and so on) Strengthening external audit (SAI and parliamentary oversight) Enabling civil society oversight (a) reform advocates or more neutral? (b) possessing significant The desk-based report should list recent ministers of influence—or mainly as technical experts? finance and to the extent available key points about their What have been the main reform proposals for PFM (domes- background; and PFM reform proposals, in particular tically and made by key external advisers—the IMF, the Bank, and those made by the IMF and the World Bank (by other key other 1–2 main donors) (focus on the past 2–3 years)? Map which donors if available), and reflect whether PFM reforms are ones that were put on the agenda have been implemented and mentioned in major documents such as PRSPs or other types of national development strategies. which ones have not, and to what extent (using the table as a National consultants should in particular contribute to the guide). question of whether and how political leaders have publicly addressed PFM reforms. The World BAnk  111 3b.  Process Tracing and Leadership Motivation/Incentives Graph 1 for the expected causal chain: reform motivations and To address these questions, teams should hone in on 2–3 reform conditions, inputs, outputs, outcomes and impacts). areas of particular interest; including 1–2 areas that failed or remain significantly incomplete, and 1–2 that were successful. The desk-review should record any information on these How did the specific PFM reform agendas emerge? Were issues that is available from existing reports;while most of the issues will need to be explored in-country. they developed in reaction to a particular crisis/challenge (such as a fiscal crisis)? Or in reaction to donor demands (and if so: Are there losers from the PFM reform components that the program triggers? Debt relief triggers?) and/or after an initial PEFA government has committed to/has sought to implement? Who are assessment? the losers and what are they losing? Are the losses licit or illicit Are there any discernible links to (a) elections and changes (for example TSA reforms entail the closure of previously legal in government, (b) external incentives such as debt relief or major accounts held by ministries and department; while certain reforms donor rounds, (c) changes in the person of the minister of finance aimed at strengthening control can threaten previously illicit rent- (or other key ‘reform champions’ for PFM), (d) major public pro- seeking from budget leakage); or primarily threatening discretion tests or similar events? by higher level politicians (regional governors, national MPs, min- Were the specific features of the agenda set by domestic isters, PM and President)? stakeholders, or were they strongly influenced by external ‘script- Are there particular winners from PFM reform components: ing’? (for example, were international consultants significantly politicians gaining greater control over other politicians and/or the involved in the drafting of reform plans? Or were these developed administration? Technocrats? MoF or conversely, line ministries? mostly within the public administration?) The central government or the regions? Has there been enthusiasm for particular types of reforms Is there public/general expert knowledge about winners and (such as program budgeting)—and what explains that enthusiasm? losers of PFM reforms, or is that at best known by insiders if at all? Does the level of enthusiasm influence actual implementation? Did the MoF seek good coordination with line ministries on For most cases, this information will need to be collected PFM reforms (especially those related to budget planning, IFMIS in country. Initial discussions should record any views and implementation, and internal controls and audit)? Have there been insights by the national consultant as well as by Bank public specific demands for PFM reforms from line ministries, and have sector, FM and CMU staff (national and international); to be these been reflected in the reform agenda adopted? cross-checked with views and insights from other stakeholders. How long were reforms on the agenda before being adopted 3c.  Demand-side Factors and implemented? (or failed to be implemented despite commit- Do NGOs raise PFM reforms as an issue (and since when have they ments made)? done so)? Are NGOs involved in any PFM related oversight function With regard to the selected reform areas that were more and (for example, of public procurement/investment execution)? less successful: what is seen as the likely explanation? What is the Do private sector stakeholders hold any particular views on likelihood of current reform proposals to be adopted and imple- PFM reforms (in particular if public investment management or mented, and to have expected impacts? procurement are selected as an area of focus)? If relevant, are Has there been an effort to establish with some clarity there relevant divisions in views between different parts of the whether the reforms sought are likely to result in improved service private sector (for example, ‘insiders’ vs ‘everyman’; large versus delivery? (and to what extent current service delivery bottlenecks small businesses)? are grounded in PFM short-comings specifically?) For example, are What is the assessed level of budget transparency (using there plans to undertake PETS type assessments or similar moni- OBI and PEFA data)? Do NGOs or others (for example, academics) toring efforts to see whether service delivery is facilitated by PFM use the available budget information? improvements? (see also Graph 1 for the expected causal chain: What is the role of MPs and parliament—supportive, neu- reform motivations and conditions, inputs, outputs, outcomes and tral, or negative? What are key incentives for MPs and parliament? impacts) (for example, support the president without question? Demand To the extent that some PFM reform elements have been more funds for their region?) implemented (please distinguish between reforms under imple- mentation, and reforms that are fully implemented), is there any The desk review should summarize existing information, evidence that these have the expected effects (for example, does in particular drawing on the Open Budget Index and PEFA the introduction of a TSA result in better liquidity management/ information (PI-10 and PI-26–28);further information and less need for short term borrowing)? If not (or not to the hoped for validation should be collected in country. extent), are there any specific obstacles to effectiveness? (see also 112  Political Economy of Public Financial Management Reforms 3d.  Bureaucratic Incentives and Inertia Given political economy dynamics, as well as institutional To what extent are PFM officials seen as supportive of reforms; features and capabilities: uninterested; or opposed? (in central agencies and in the line min- istries)? Are there incentives for bureaucrats? Or do the reforms • What is the degree of opportunity in this case and period in time? bring risks? If possible: are there differences between senior and managerial staff and general staff; or by staff in different func- • Have recent PFM reform initiatives gone in the right tional roles (for example, treasury staff versus budget department direction? staff)? • What are three to four points to note for approaching PFM reforms going forward (in terms of choice of reform area and/ The desk review should cover existing information;the or strategy for implementation)? Are there any contextual issues should be covered in further depth in country. changes that would make PFM reforms more likely, or those 4.  Questions to Guide the Development of an Overall Story Line and under way more robust? Implications • Looking forward, distinguish: Overall, do key national level politicians and their core support- reforms that are on the agenda but unlikely to be fea- iii.  ers have incentives to pursue (or facilitate the pursuit of) PFM sible given identified PE blockages, reforms? What are the incentives? Are PFM reforms politically salient (and has this changed over time)? improvements that are on the agenda that are likely to iii.  To the extent that there are incentives to pursue PFM be feasible (and their expected robustness given PE reforms: is this leaning towards seeking tangible impact? Or more challenges), towards ‘just enough’/window-dressing/de jure reforms? iii.  changes that are important for a better functioning of What explains the status of selected PFM reform efforts to the PFM system but that are currently not clearly on the date? agenda combined with an exposition of whether and how Use Graph 1 to guide the story line regarding: reform moti- these could become politically attractive. vators, reform inputs/efforts undertaken, outcomes and impacts achieved. The World BAnk  113