WPS6174 Policy Research Working Paper 6174 Making Public Sector Reforms Work Political and Economic Contexts, Incentives, and Strategies Simone Bunse Verena Fritz The World Bank Poverty Reduction and Economic Management Network Public Sector Governance Unit August 2012 Policy Research Working Paper 6174 Abstract Supporting effective public sector reform is a major public sector reforms, but that at the implementation challenge that the World Bank and other agencies and stage, political costs frequently outweigh potential gains; stakeholders have been grappling with. It is increasingly and hence reforms are abandoned or left to wither. recognized that political economy factors play a crucial Real breakthroughs have been achieved in countries role. However, beyond this broad proposition, specific experiencing major structural shifts and those having questions arise: What country contexts are more/less political leadership committed to higher-level goals. propitious for public sector reforms and what reforms are The review of operations shows that successful projects likely to succeed where? And can more explicitly taking are significantly more widespread than the literature political economy challenges into account help to pursue would lead to assume. Furthermore, it provides tentative public sector reforms even in less propitious contexts? evidence that investing in understanding political This paper addresses these issues in two ways: first, economy drivers has been associated with better it draws on the existing literature to identify key project performance. Key implications are the need propositions about factors that can trigger or facilitate to differentiate between country contexts more clearly public sector reforms, and those that tend to work ex ante, concentrate more on reform implementation against (successful) reforms. Second, it investigates the during windows of opportunity that are typically of experience of World Bank public sector operations over limited duration, and design reforms with a clear plan of the decade 2000–2010. It finds that governments in engagement with stakeholder incentives. many developing countries face incentives to initiate This paper is a product of the Public Sector Governance, Poverty Reduction and Economic Management Network. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank. org. The author may be contacted at vfritz@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Making Public Sector Reforms Work: Political and Economic Contexts, Incentives, and Strategies Simone Bunse Verena Fritz1 Keywords: public sector reforms, political economy, incentives, public financial management, civil service reforms, development operations JEL classification: P16, P48, H11, H83 Sector Board: Public Sector Governance 1 Simone Bunse directs Georgetown University‘s Oxford Graduate Program in International Management and is adjunct faculty at Georgetown‘s Public Policy Institute. Verena Fritz is a Senior Governance Specialist at the World Bank. The views reflected in this paper are solely those of the authors. Contents I. Introduction .......................................................................................................................................... 4 II. Political Economy Dynamics and Public Sector Reform: Lessons from the Literature ......................... 6 II.1 Political Economy Factors Hindering Public Sector Reforms .............................................................. 7 II.2 Political Economy Factors Facilitating Public Sector Reforms .......................................................... 10 II.3 Public Sector Reform Implementation and Sustainability ................................................................ 17 II.4 Summing up: PE factors favoring and hindering public sector reforms and their frequency across countries ................................................................................................................................................. 21 III. Findings from the World Bank Portfolio Review............................................................................. 23 III.1 Public Sector Reform Efforts: Spread and Scope ............................................................................. 23 III.2 Frequency and Types of Political Economy Problems: Patterns Across Countries and Operations 26 III.3 Attention to Political Economy Drivers: Evolving Practices and Potential Pay-offs ........................ 33 IV. Making Public Sector Reforms Work: Preliminary Conclusions ...................................................... 37 IV.1 Summary of Key Findings: Public Sector Reform Trajectories ......................................................... 38 IV.2 Implications for a Research Agenda and for External Support to Public Sector Reforms ............... 40 References .................................................................................................................................................. 43 2 List of Acronyms AFR Africa (Sub-Saharan) CSR Civil Service Reforms DPL Development Policy Loan EAP East Asia Pacific ECA Europe and Central Asia FY Fiscal Year GAC Governance and Anti-Corruption GDP Gross Domestic Product GNI Gross National Income GPE Governance and Political Economy HIPC Highly Indebted Poor Countries HR Human Resources HS Highly Satisfactory HUS Highly Unsatisfactory ICR Implementation Completion Report IDB Inter-American Development Bank IEG Independent Evaluation Group IL Investment Loan IMF International Monetary Fund LAC Latin America and Caribbean LIC Low Income Country MDAs Ministries, Departments, and Agencies MENA Middle East and North Africa MIC Middle Income Country MS Moderately Satisfactory MTEF Medium Term Expenditure Framework MUS Moderately Unsatisfactory OECD Organisation for Economic Cooperation and Development PAD Project Appraisal Document PEFA Public Expenditure and Financial Accountability PFM Public Financial Management PSG Public Sector Governance PSMA Public Sector Management Approach QAG Quality Assurance Group S Satisfactory SAR South Asia US Unsatisfactory WBI World Bank Institute WGIs Worldwide Governance Indicators 3 I. Introduction Academics and practitioners alike are increasingly recognizing the influence of political economy variables in public sector reforms; and this recognition is also reflected in the World Bank‘s recently adopted Public Sector Management Approach (World Bank 2012). However, beyond this overall proposition, important questions remain: What country contexts are more/less propitious for public sector reforms? What types of reforms are most likely to succeed in what contexts? And can more explicitly taking into account political economy challenges and opportunities help enable and sustain public sector reforms even in less propitious contexts? This paper contributes to addressing these questions in two ways: firstly, reviewing the existing literature, it sets out and organizes key propositions on factors that can trigger or facilitate public sector reforms, and those that are assumed to work against (successful) reforms. Secondly, building on these propositions, it investigates the experience of World Bank public sector operations over the past decade. It explores to what extent operations have succeeded relative to country conditions and dynamics, as well as whether and how political economy risks are recognized and managed by operational teams.2 While a diverse field, the common denominator of political economy is a concern with the interaction of political and economic factors – and in particular with individual and group interests, incentives, and actions, structural factors, and institutions. Political economy analysis is used to understand the context and dynamics in which development efforts and interventions take place; and especially the incentives of various stakeholders that come to bear on reform initiatives, such as improving public sectors. Public Sector Reforms is an umbrella term that includes several sub-streams, notably public financial management (PFM) reforms, civil service reform (CSR), decentralization, revenue management reforms, legal and judicial reforms, as well as anti-corruption (see also IEG 2008 and World Bank 2012). According to the Bank‘s 2012 Public Sector Management Approach: ―Public sector management (PSM), reform is the art and science of making the public sector machinery work. It is about deliberately changing the interlocking structures and processes within the public sector that define how financial and physical resources and people are deployed and accounted for.‖3 The intention is generally to achieve greater capacity, efficiency, integrity, or responsiveness (and typically some combination of these) within the public administration, in the expectation that this will contribute to realizing other development goals. Key findings of the paper are the following. A review of the literature on public sector reforms, including the literature covering a range of country experiences suggests that many countries – both LICs and MICs – face contradictory drivers both in favor and in opposition to public sector reforms. Incentives for initiating public sector reforms include globalization, regional integration, or aid dependency. However, socio-economic and political dynamics often run counter to reform attempts, especially when these threaten existing ‗limited access orders‘ whereby elites enjoy privileges such as access to jobs, public sector contracts, and regulatory 2 The review of the Bank‘s experience has included project documents and ex -post ratings, findings from a review by the Bank‘s Quality Assurance Group (QAG), and interviews with selected task teams. See section III. 3 World Bank (2012): 3. 4 distortions. Consequently, public sector reforms are initiated in a relatively wide range of countries, while the conditions for real success in implementation are less frequent. As a consequence, many reform attempts become stalled or remain at best partially implemented; but there are also instances of more substantial reform success. Against this discussion in the literature, the paper reviews the World Bank‘s experience with public sector reform operations during the past decade. It finds that Bank operational support is spread across a wide range of countries. Significantly, only about one quarter of these operations is implemented in aid-dependent countries. However, rather at odds with the finding from the literature that the conditions for real progress are relatively rare, the analysis finds a relatively high success rate of about 75 percent (operations rated as moderately satisfactory or better). Even for investment operations, which typically involve greater challenges than development policy operations4, the share of operations rated at least moderately satisfactory is still 70 percent. This appears to have three explanations: (i) a significant share of public sector operations has been implemented in middle-income countries, including a number of countries with strong incentives for reforms; (ii) contrary to what might be expected, the performance of operations at the ‗low end‘ of pre-existing governance conditions, specifically in post-conflict countries has been relatively good (but also uneven); (iii) a share of operations target the initiation of reforms, de jure reforms (legal changes), and technical changes that do not entail a reduction of rent- seeking opportunities or opportunities for shoring up political support, and are therefore relatively easier from a political economy perspective. Especially given point (iii) and to an extent (ii), there is also a degree of disconnect between operational success achieved and actual change in the transformation of public sectors and of governance, as has been noted by IEG 2011.5 This ‗disconnect‘ is to some extent inevitable – public sector reforms will more frequently be initiated or implemented partially, than be truly transformational. That said, the analysis finds some weaknesses in current operational practice – in terms of more realistic assessment of risks and of reform feasibility ex ante, of seeking to deliberately improve or stabilize the ‗authorizing environment‘ for reforms, and of focusing on gains that are tangible and sustainable, rather than driven by notions of international best practice – which contribute to the disconnect. The paper‘s findings lead to a number of propositions for external reform support. First, support to public sector changes needs to reflect key dimensions of country context more explicitly in operational design. Especially engagement in less propitious contexts should address operational feasibility and monitoring reform impact explicitly. Second, in intermediate and more challenging contexts, political economy analysis can play a more routine role in identifying options for a ‗feasible stretch‘ from the status quo to an improvement in public sector performance that is likely to be effective and to stick. 4 Development Policy Operations define a set of prior actions that have to be met before an operation is approved. Project outcomes are ex post measured against a separate, but typically closely related set of indicators. Funding is disbursed as budget support to governments once prior actions have been met. In Investment Lending Operations, all expected outcomes have to be achieved during implementation. 5 In contrast to the 75 percent of public sector operations rated as moderately successful or better, overall progress on public sector reform was only observed in 47 percent of completed and reviewed country programs (IEG 2011: 8). 5 Third, the scale and timing of any reform needs to be aligned with the limited windows of opportunity which may exist. In less favorable environments, where public sector reforms are often abandoned due to elections or the departure of reform champions, specific steps using limited windows of opportunities may still be feasible at certain times. Although such ‗quick wins‘ do not necessarily spark immediate wider public sector reforms (as shown by Ghana‘s experience – see textbox 2), they can provide impulse for further change, as well as bring improvements in the ‗here and now‘. Fourth, reforms should seek to target challenging issues, but in ways that take political economy grounded challenges and limitations into account while seeking to ‗stretch‘ these. The need to show operational success should not lead to external support shifting towards easier issues such as reform design, de jure adoption of reforms, or technical changes that do not effect deeper change. Rather, reform incentives and support should focus significantly on the actual use of improved systems and processes, and on monitoring of results. Fifth, in unfavorable reform contexts, monitoring and fostering support from key political leaders and among veto players is crucial, if challenging. In this context, a clearer understanding of the role and importance of citizen demand for public sector reforms and how to promote it more effectively would be useful. So far, the literature offers little conclusive guidance on this issue. Sixth, approaches for aid dependent countries with strong political disincentives – including those where structural incentives to reform are so weak that they have disengaged from seeking better public sector performance – need to be strengthened. To make this happen may require enabling more innovative operations that may (initially) be small in scale, or include scope for trial-and-error in order to test potential opportunities to stretch the boundary of feasible reforms. Finally, the focus of this paper is primarily on the political economy dynamics surrounding public sector reforms rather than on issues with the content of these reforms discussed elsewhere (World Bank 2012, Andrews 2009 among others). Nonetheless, the paper and its findings are fully aligned with the emphasis that flexibility and innovation in public sector reform content is essential to enable greater feasibility as well as sustainability and impact of such reforms. II. Political Economy Dynamics and Public Sector Reform: Lessons from the Literature Ideally, the literature should provide insights into the specific political economy issues which condition the success of public sector reforms, motivate reformers, and signal their level of political commitment to PSRs. Further, it should establish under what circumstances reforms are more or less likely, the differences between different sub-streams of public sector reforms (for example public financial management (PFM) and civil service reforms (CSRs)), and between different initiatives within these sub-streams (for example seeking to introduce policy-based budgeting versus trying to strengthen merit based hiring practices). Unfortunately, to date most research is not sufficiently systematic and comparative to allow clear and generalizable lessons. Furthermore, explicit political economy perspectives on public sector reform challenges are still rare. This holds particularly true for work on public sector reforms in LICs and MICs. Much of the literature on public sector reforms in developing countries consists of descriptive single case 6 studies. Conceptually, a frequent focus is the suitability of New Public Management-inspired reforms for lower income countries, rather than political economy dimensions. In addition, not many studies disaggregate or specifically analyze the different types of public sector reform, their costs and benefits, and the particular political economy challenges they may present or alliances they may generate. As Nelson (2007, p. 25) argues, ―reform is often discussed as if it was a coherent entity. In reality, reform is an umbrella label for a range of diverse actions, having in common only the intent to increase equity, quality, and/or efficiency of service provision.‖ Finally, there is a gap between research covering states and state-building at the macro level, more recent discussions of political economy and social contexts of development (North et al 2009 and 2010), and the literature addressing public sector reforms in particular (with a few more explicit connectors such as Fukuyama 2004, 2011; Grindle, 2004; also O‘Neil 2007). Furthermore, there has been little investment in comparative data on public sector systems and performance to date (see also World Bank 2012); while recent initiatives are hoped to gradually change this situation.6 Nonetheless, the literature offers an opportunity to systematize proposed reform drivers, group them according to their assumed impact on public sector reforms, and develop a tentative conceptual framework. Drawing links to the broader literature on political economy aspects of reforms in general provides some confidence that the range of factors raised merit attention (many of them recur for other reform issues) – although it is not possible to test their relative weight or eliminate those that may matter less. As databases on public sector reform efforts and actual performance improve, it will be desirable to test these propositions, their effects, and relative weights further in future. II.1 Political Economy Factors Hindering Public Sector Reforms Much of the literature that employs political economy perspectives is skeptical as to the success and sustainability of public sector reforms given the weak political incentives to pursue them. Explanations include general reasons affecting many ‗public welfare‘ oriented reforms, as well as some specific to public sector reforms. The overall literature on political economy and reforms points to several PE factors with potential negative impacts on a range of public welfare enhancing reforms. These include politicians‘ concern with votes, staying in power (campaign contributions, loyalty)7, or acquiring personal wealth (Keefer 2011). In addition, politicians are unlikely to back reforms which are uncertain to work and yield results prior to subsequent elections. However, at times they may also promise reforms to bolster their legitimacy or win elections – without being able or willing to spend sufficient political capital to make them happen. 6 For example, efforts are being made at collecting actionable governance indicators and making these readily available (http://www.agidata.info/site/). 7 The former may be more important in semi- to fully democratic countries, while the latter matters more in semi- to fully authoritarian regimes where elections – if held – tend to be uncompetitive. For example, campaign contributions may be less important if politicians can use state-owned media for their political campaigns, but the loyalty of editors of these media then matters. 7 A somewhat different perspective – but with similar implications – is that by North et al. (2010) who posit that many LICs and MICs function as ‗limited access orders‘ with elite groups maintaining an equilibrium that excludes non-elite groups from many opportunities as a way to restrain violence (to which dissatisfied elite groups might resort). Public sector reforms introducing meritocracy and reducing opportunities for rent-seeking are at odds with such an equilibrium – hence the low incentives to pursue them. Moreover, in such an equilibrium, the uneven application of rules – in ways that privilege some while holding back others – is often crucial to the continuation of rents; and this in turn requires civil servants that are at a minimum acquiescent to such practices. In the literature on public sector reforms specifically, five factors are recurrent themes: (i) the fact that public sector reforms combine significant short term costs with only long-term prospects for meaningful gains which makes it unattractive for many governments to make the initial investment (time horizons); (ii) the degree of clientelism that connects the political level and the public administration and the extent to which the public sector (and specific sub-areas) serve purposes of rent-seeking and the maintenance of clientelist networks (related to the presence of limited access orders); (iii) resistance to change or bureaucratic inertia and status quo bias at the level of the bureaucracy, (iv) frequently limited demand for reforms due to a limited development of civil society and of an entrepreneurial class separate from the state; (v) and specific institutional incentives of political decision-makers stemming from electoral and party systems as well as political regimes which can make it more or less feasible to initiate public sector reforms. The long-term nature of public sector reforms is highlighted by Pritchett and de Weijer (2010). The authors estimate that significant improvements in the functional capability of the state take 10 to 20 years.8 For most countries this is clearly beyond the life-cycle of a single government. This implies that several governments would have to persevere with a common thread of reform efforts and incur related costs, while the benefits from improvements will accrue to future decision-makers. Clientelism can be an important tool for maintaining power and, as emphasized by North et al. (2010), potentially peace. In a clientelist system, politicians and senior bureaucrats act as patrons dispensing jobs or PS contracts to networks that in turn support the politician or political party (Kitschelt 2011 and 2007). Strong reliance on clientelism - found in many developing countries - increases the cost and reduces the incentives for pursuing public sector reforms (Geddes 1991, see also IDB 2006; O‘Neil 2007). In his essay on the political economy of public sector reform in Jordan, El-Said (2000, p. 8) describes ―parasitic relations‖ between the private and public sectors and how ―family members often open business related to [state] agencies in which their relatives are prominent.‖ In many African LICs, clientelism (variously mixed with party-focused politicization) is also an important dimension (van de Walle 2003; Posner 2005; Thomson 2010; Eisenstadt 1973; O‘Neil 2007). Here, rent-sharing among elite groups and clientelist networks permeating the public and private sectors are often linked to ethnicity, religion or region (Devarajan, Khemani, Walton 2011). As many public sector reforms in LICs and MICs (for example in the areas of public procurement, tax administration, or audit offices) ultimately aim at 8 See also Fatile and Adejuwon (2010). 8 limiting the scope of personal enrichment of politicians and their crucial clients, including bureaucrats, ‗privileged‘ corporate interests, they are likely to be opposed (O‘Neil 2007; Campos and Pradhan 2007). Reshaping administrative structures and personnel practices involves changing bureaucrats‘ incentives to make enhanced performance individually rational. However, much research – primarily on bureaucracies in developed countries – has shown that ―old habits die hard‖ (Jacobson 2003; Polidano and Hulme 1999; Hannan and Freeman 1984; Blau 1955; Crozier 1964; Merton 1940). While less research has been carried out on middle and lower income countries, bureaucratic interia or resistance is also assumed to be present. For example, Harun (2007, p. 368), exploring public sector accounting reform in Indonesia, argues that senior officials‘ familiarity with the old accounting system and the widely shared opinion that ―what worked before would continue to work in the future‖ was a key reform obstacle. Where public sector workers are organized in trade unions, they can have particular clout in opposing reforms. The presence of bureaucratic resistance in turn implies an increase in the necessary political capital that leaders have to invest to implement public sector reforms. It is likely that there is significant variation in the nature and degree to which bureaucratic inertia presents an obstacle to public sector reforms, given variations in public sector staff turnover and job security, and possibly also political regime characteristics. Additional reasons for bureaucratic resistance in non-OECD countries include experience with retrenchment during the structural adjustment period which continues to give public sector reforms a negative ring in many countries (see Textbox 2 on Ghana). Furthermore, in countries where civil servants rely strongly on informality to get a ‗decent salary‘, they may be particularly distrustful of a tightening of rules. In Indonesia the dependence on non-salary, quasi-legal or illegal forms of remuneration of bureaucrats and members of the executive undermined reform attempts of more accurate monitoring and reporting of financial operations - they were seen as a direct threat to their income levels (Harun 2007). Demand for public sector reforms – whether from civil society or the private sector – tends to be described as weak, even in countries where opportunities for demand-side engagement would appear to exist. Harun (2007) argues in the Indonesian case that ―the public has yet to become used to the idea that it can elect a new government from time to time, and that it therefore needs to monitor incumbent governments‘ financial operations.‖ Clientelistic linkages between the public and the private sector described above mean that many private sector entrepreneurs will have an interest in the status quo rather than demand reform. Similarly, Sozen and Shaw (2002) find that demand-side pressures did not play a significant role for public sector reforms in Turkey. Despite such a weak articulation of citizen demand for public sector reform, however, the new Turkish government coming to power in late 2002 adopted a sweeping set of reforms in response to structural pressures, incentives, and opportunities – including a recent fiscal crisis, the need to win legitimacy and EU approximation, and the opportunity offered by a single-party government with a comparatively strong electoral mandate. Finally, parts of the literature explore the effects of certain institutional structures or political regimes on public sector reform attempts. Schwarz (1994) shows how the power relationship between the executive and the legislature matters when adopting public sector reforms. The 9 literature on public sector reforms in OECD countries tends to argue that in countries in which single-member districting created an absolute majority for the governing party, the chances of adopting significant PFM reforms where higher (e.g. UK, New Zealand). Geddes (1991) explores how different electoral and party systems affect legislators‘ incentives to initiate CSRs, focusing on merit-based hiring in Latin America.9 She finds that open lists of proportional representation reduce the probability of reform because ―it makes patronage a valuable resource to incumbents in their struggle against challengers within their own party.‖ Cruz and Keefer (2010) argue that countries without strong programmatic political parties are less likely to pursue public sector reform10 given the shortage of the accountability structures this implies which could prompt politicians to be more sensitive to public sector management and performance issues. While programmatic parties exist in a number of middle income countries, they are rather rare in low income environments. The argument that implementing reforms in autocratic regimes may be easier than in democratic regimes (given the limited opposition forces, and less dependence on electoral cycles) derives from the literature examining economic reforms (for a more detailed discussion on the reasons see, for example, Armijo, Biersteker, and Lowenthal 1994; see also Haggard and Kaufmann 1992). In the case of civil service reforms, including transparent and systematic personnel policies, however, Gehlbach and Keefer (2011) find that autocrats confront a tradeoff. On the one hand, such reforms ―allow the autocrat to make credible commitments to reward public officials who undertake efforts that contribute to regime goals.‖ On the other hand, they ―also make it easier for bureaucrats to organize against the regime in case it reneges.‖ Some non- democratic regimes such as Vietnam (Painter 2003) or China (Moon and Ingraham 1998) have undertaken significant public sector reforms (building on long if variable historical legacies of public sector meritocracy as noted by Fukuyama 2011); while other authoritarian as well as hybrid regimes have either not engaged in reforms, or have followed a pattern of partial or start- stop reforms. Despite, these numerous disincentives and obstacles, public sector reforms are on the agenda in many countries – and show progress at least in some. When then do such reforms become nonetheless feasible? II.2 Political Economy Factors Facilitating Public Sector Reforms The literature highlights several positive ‗macro‘ drivers, as well as institutional variables, and more proximate political economy reform facilitators. In contrast to the obstacles discussed above which are mostly linked to stakeholder interests, time horizons, and social structures, the enabling factors are often structural. However, the more proximate positive drivers also center around stakeholder constellations. 9 Electoral rules have also made a difference in countries, such as New Zealand, which embarked on major reforms in the 1980s. As Schick (1996) argues in case of economic management reforms with the first past the post-electoral arrangements then in effect, one party typically had a majority in Parliament and did not have to dilute its reform program to gain coalition support. Interestingly, voters then opted to change electoral rules in the country in the 1990s to more proportional representation, resulting in successive coalition governments and more compromises in reform efforts. 10 To support their finding, the authors draw on the ratings of 511 World Bank PSR loans in 109 countries which were systematically higher in countries with programmatic political parties. 10 Macro-level drivers seen as potentially supportive of public sector reforms include recent democratization, fiscal pressures, economic transformation, globalization and/or regional integration, and regional contagion. Recent democratization is assumed to be a potential trigger for PSRs (although not guarantee their implementation) because it creates or renews electoral competition and demand for improved service delivery and fighting corruption, the observance of meritocratic principles in HR management, and greater participation from civil society (Schneider and Heredia 2003, Peters and Savoie 1995, Kiragu 2002, Shin and Ha 1999, Brinkerhoff 2000). The experience of a number of countries in Latin America, South East Asia, as well as Central and East Europe comes to mind. In these settings, democratization has frequently been combined with other macro-drivers such as economic take-off or regional integration. However, recent democratization is seen as having variable effects. In some regions, such as East and West Africa recent democratization has not been associated with significant progress on public sector reforms – which, as Cruz and Keefer (2010) have argued, may be due to the relative absence of programmatic parties in these regions. The introduction of competitive elections can also create temptations to use the bureaucracy for electoral ends (Schneider and Heredia 2003) or increase opposition to aspects of reforms for electoral gain (Bangura and Larbi 2006). This suggests a need to better understand the way in which and the conditions under which democratization may act as a positive reform incentive. Fiscal pressures figure prominently in the public sector reform literature; in particular ‗first generation‘ downsizing reforms. According to Schneider and Heredia (2003) fiscal crises best explain the timing of reform. Similarly, Suleiman and Waterbury (1990) argue that public sector reform is ―begun and sustained in response to [fiscal] crises.‖ This seems particularly plausible when fiscal crisis goes hand in hand with aid-dependency. Many developing countries have faced high public debts, and operate with tight budget constraints relative to the range of public services they seek to supply. Public sector reform is hence seen as crucial not just to improve service delivery but to prioritize and manage the limited resources more strategically (Kiragu 2002). Part of the argument is further, that crises can break the power of vested interest groups creating opportunities for change (Brinkerhoff 2000). In a successful case like Brazil (see Textbox 1) or Turkey, the experience of fiscal crisis is seen as an important motivating factor at early stages. Subsequent economic take-off and fiscal expansion, however, seems to have aided their sustainability. Although crises may be a trigger, they do not automatically translate into successful reforms. Effective implementation requires deep institutional changes, which do not necessarily accompany crises (Tomassi 2003). Some authors argue that rather than increasing political will for reform, crises can undermine leaders‘ willingness to engage in difficult changes as they are forced to focus on their short-term survival and immediate steps to deal with the crisis (Kulibaba and Rielly 1994, Brinkerhoff 2000). They stress stability as a condition to embark on reform instead. Clearly, there are instances in which fiscal crises have failed to trigger significant reform progress. As is discussed below, macro-level triggers are more likely to result in reforms if they are matched by propitious facilitating political and wider stakeholder constellations. 11 Intensive economic growth (and the shifts in interests and pressures accompanying it) appears to be a potentially powerful driver for public sector reforms. Brazil, China, and India have all been undertaking a range of public administration reforms in recent years – albeit often unfinished – that have feedback links with their economic development (Gaetani 2003).11 Exploring PFM reforms across African countries, Andrews (2008, pp. 25-27) finds that relatively fast growing, African LICs feature in the group of countries with the highest Public Expenditure and Financial Accountability (PEFA) PFM performance scores.12 Stevens and Teggemann (2004) argue that a country‘s economic trajectory and public sector reforms are ―inextricably intertwined.‖ While economic crisis may trigger reform, economic resources are needed to implement them and reform opportunities can open up with economic growth. Once growth has become rapid, some of the typical constraint associated with a limited access order – in which elite groups have strong incentives to focus on the maintenance of privileges rather than pursuing profits in competitive markets – are relaxed, and the political cost of pursuing public sector reforms decreases. A more specific theme in the literature is incentives linked to globalization. Schneider and Heredia (2003) find that ―in the context of increased international competition and volatile capital flows, capitalists have strong incentives to push for administrative reforms that reduce transaction costs (ports, customs, courts), level the competitive playing field (monopoly and trade regulations), and enhance transparency (budgets, international reserves, monetary policy).‖ The restructuring of the global economy has been the common background of public sector reforms in many developing countries (Bangura and Larbi 2006, Kiragu 2002, and Bertucci and Jemiai 2000). Governments seeking to improve the export potential of their countries beyond unprocessed commodities face a pressure to reduce public sector dys-functionalities at least in selected areas such as tax and customs administration. A final positive macro reform facilitator is cross-country contagion and imitation of reforms. Public sector reform has spread across regions in East Asia, Latin America, much of Eastern Europe, and the former Soviet Union (1997 World Development Report). Neighbors‘ example and ―the threat of being left behind‖ can be a powerful driver. 13 In Sub-Saharan Africa, Ghana pioneered the introduction of autonomous agencies in an attempt to improve public service delivery. After transforming its Rural Water and Sanitation Department to an agency, other countries, including Uganda, Tanzania, Kenya, and Zambia, followed suit (Kiragu 2002). Toonen (2001) argues that globalization is associated with a speedy dissemination of ideas which creates domino effects in reform adoption. Likewise, Garcia-Zamor and Khator (1994) argue that with globalization governments become more influenced by international environments and more sensitive to global issues, including the modernization of the state (see also Suleiman and Waterby 1990). 11 See also: http://www.cbs.dk/en/Research/Departments-Centres/Institutter/ARC/Hoejreboks/Arrangementer/2011- 09-23-09-00-00-Globalization-and-Public-Sector-Reforms-in-India-and-China [Accessed: June 1, 2011]. 12 Andrews (2008) attributes this to the HIPC process and the strong targeting of and incentives for PFM reforms in a number of LICs, combined with the opportunities offered by economic growth in a ‗top -league‘ of LICs with higher growth. 13 DiMaggio and Powell (1983) first proposed the concept of isomorphism whereby one unit in a population is forced to resemble other units that face the same environmental conditions.‖ In their competition for political power and institutional legitimacy, reforms may move up political agendas. 12 Others lament such domino effects, arguing that they leave little room for innovation and homegrown responses to public sector problems (Ingraham 1996, Hood 1996, Pastor and Wise 1999), and lead to the adoption of ‗isomorphic facades‘ of institutions that are unable to improve the real functioning of the public sector (Andrews 2009). Thus, especially when the distance between international trends and local conditions is wide, contagion can feed into the syndrome of reform initiation followed by failed implementation that has been discussed above. The more specific pressures and institutional reform facilitators include donor pressure, domestic legitimacy concerns, legacies of earlier reforms, and post-crisis re-building efforts. Public sector reforms can be initiated as a result of pressure from international financial institutions (IFIs) by making lending conditional upon it or continuously advising to undertake such reforms. In poor, aid-dependent countries multi- and bilateral donors are frequently major reform stakeholders (Therkildsen 2000 and 2001; Baimyrzaeva 2010).14 General developmental goals to sustain government legitimacy can also drive reform. Public sector reform in non-democratic regimes, including China and Vietnam come to mind in this context. In Vietnam, public sector reform momentum was driven by elite state actors‘ concern about continued public support (Painter 2003) and about violence sparked by citizens aggrieved by corruption and maladministration. The need to sustain legitimacy is particularly relevant as an explanation of why governments that are in a comparatively strong position (rather than facing a fiscal crisis or pressures from democratization) may still in some instances chose to adopt public sector reforms. Legacy explanations tend to search for broad similarities between the administrative past and the present of countries to account for the importance of the legacy (Meyer Sahling 2009). Fukuyama (2011) draws a number of long historical legacies (stretching several centuries or even millennia) to explain the rise and persistence of certain aspects of public administrations – e.g. of an impersonal, meritocratic civil service –, while also underlining that their prevalence can wax and wane depending on the political order and government of the day. Morgan et al (2010) argue that countries with a history of reform may have a greater willingness to experiment with public sector reforms than other countries – in their case focusing on the specific experience of Tanzania. Finally, post-conflict situations can provide a critical juncture for reform. Rwanda‘s post conflict reconstruction effort, for example, included a vast array of public sector reforms (Musoni 2003). Whether such a critical juncture is used to improve the public administration, however, depends on a political leadership pursuing the creation of an effective state (Leftwich 2000). The fact that post-conflict situations can offer opportunities for reform is also reflected in the experience of World Bank public sector operations discussed in part III. 14 The importance of donor ‗prodding‘ is also widely seen as a factor by development pra ctitioners and government officials in particular from more aid dependent countries. See , for example, ‘Africa must own its public sector reform agenda‘ (April 11, 2011), available at: http://today.gm/hi/news/3278.html [accessed: June 1, 2011]. If reforms are predominantly donor driven and government ownership is weak or absent, implementation is likely to suffer. However, as Andrews (2008) argues with regards to the HIPC process and support for PFM reforms, donor efforts can also help moving reforms forward, especially if governments have a strong incentive to engage. At the same time, donor efforts are frequently affected by coordination weaknesses that can limit their impact (see, for example, de Renzio 2011). 13 Some of the macro-level drivers discussed here – in particular, globalization and international contagion – affect most developing countries; while others are more limited to specific countries or time periods – e.g. sustained rapid growth, a past legacy of meritocratic and relatively non- corrupt public sectors, or acute fiscal crises. The following section turns to the specific stakeholder constellations that can be seen as facilitating whether broader macro-level drivers in fact translate into public sector reforms. II.2.1 Proximate Factors: Policy Entrepreneurship and Windows of Opportunity The more ‗proximate‘ reform factors which emerge in the literature focus on concrete stakeholders, their motivations and ability to move public sector reforms forward (‗policy entrepreneurs‘), and the time (windows of opportunity) when reforms take place (see Panizza and Philip 2005).15 They are connected to wider drivers as discussed above – although the linkages between broader contexts/pressures and proximate factors that translate these into concrete actions are not always explicitly addressed.16 According to Kingdon (1995), policy entrepreneurs act within three simultaneous ‗streams‘ in the agenda-setting process: (i) the identification of problems; (ii) the proposing of specific policies or policy alternatives; and (iii) politics (political changes or actors direct attention to certain agenda items). At certain times a ‗coupling‘ of these streams takes place. A problem comes to the fore, policy proposals exist, and the political climate for decision emerges opening a ‗policy window‘ for successful adoption or policy change. In the area of public sector reform, such windows of opportunity are most likely when structural and dynamic factors align:  for example, when reform pressure increases due to a change in fiscal conditions or to processes of regional or global integration;  when such pressures are combined with changes in government that reduce the stakes that new office holders have in the public sector status quo; and  when ideas for public sector reforms and policy entrepreneurs close to political leaders exist able to shape reforms and move them along. As such, policy entrepreneurship is not only about individual leaders with a certain vision or agenda, but also about process, mobilizing resources and coalitions in support of their reform goal. Success depends on ―the structure of constraints and opportunities [such policy entrepreneurs] face in the institutional and political context; the interests, strength and nature of the political opposition; the strategies they adopt; the networks they exploit; and the manner in which their tactics and communications are ‗framed‘.‖ (Leftwich 2011, p. 6).Heredia and Schneider (1998), Nelson (1994), Gaetani (2003), Bangura and Larbi (2006) all stress the importance of political entrepreneurship by key change agents to overcome reform obstacles. 15 On a discussion of the distinction between more general/background/distant and more proximate causes see Kitschelt (2001). The challenge with proximate causes is that they raise t he questions of ‗why did these causes arise in this situation?‘ -- and their linkage with other drivers (structural as well as individual interests). 16 Most of the literature on policy entrepreneurs draws on examples from OECD countries and from Latin America; while it is also related to the idea of committed reform teams with some degree of ‗insulation‘ from particularistic interests that were seen to be crucial to some of the post-communist reforms in Eastern Europe. For a more detailed discussion of PSR in Central and Eastern Europe, see Hesse (1997) or Dimitrova (2002). 14 Heredia and Schneider (1998, p. 4) emphasize the ―salience of strategic choices and calculations of state reformers in shaping the political costs of different types of reforms‖, while (Bangura and Larbi 2006, p. 285) highlight the need to "encourage and support coalitions around common issues of interest to key stakeholders in reforms". Gaetani (2003) finds that in Brazil under the first Cardoso administration (1995-1998), Minister Bresser-Pereira and his political entrepreneurship played a crucial role in convincing the government of the need to reform its civil service, shaping public opinion, influencing Congress, and getting unlikely constitutional amendments approved. Text Box 1 provides an overview of Brazil‘s reform experience. Similarly, looking at PFM reforms in Africa, Hedger and de Renzio (2010) argue that ―a particular type and quality of leadership by politicians and technocrats may be needed to capitalize on the reform opportunities […]‖ (see also Andrews 2008 and Hedger and Kizilbash 2007). Text Box 1. Administrative Reform in Brazil: Political and economic drivers, policy entrepreneurship, and reform progress Brazil adopted a major public service reform in 1995. Reform objectives combined the need for fiscal adjustment (by staff reductions; wage ceilings; and alterations to the retirement system) and making public administration more efficient and focused on citizens‘ needs (by strengthening the ―strategic core of state‖ while at the same time creating ―autonomous agencies‖ and ―social organizations‖ controlled through management performance contracts (Bresser-Pereira 1999, p. 131)). Especially the initial reform period was inspired by the notions of New Public Management to overcome rigidity in favor of greater flexibility and greater managerial autonomy. Reforms proceeded despite enormous political obstacles, including: the need for constitutional revision to implement the reforms; resistance from lower and middle civil servants, unions, leftist political and interest groups; a political elite in fear of losing privileges; and opposition from the judiciary. The reforms responded to fiscal crisis, a crisis in Brazil‘s bureaucratic system (stemming from the military regime); the transition to democracy; and economic globalization. They were adopted in an encouraging economic context of 4.4% of GDP growth in 1995. GDP growth subsequently averaged 2.3% during the government of President Fernando Henrique Cardoso (see Giambiagi and Ronci 2004). Luiz Carlos Bresser-Pereira, Minister of State Reform, was the main intellectual and political architect of the reform. His policy entrepreneurship was crucial to push these problems into the spotlight, convince the government, shape public opinion, influence Congress, and get the constitutional amendments approved. His resources included his technical grasp of the issues and political skill; closeness to the President; and the political capital the government had gained after bringing to an end a decade of unprecedented high inflation and per capita income stagnation. Key elements of Bresser‘s reform strategy included:  identifying and neutralizing opponents while seeking out political allies (including the President, business people, mayors and governors, senior civil servants, and in Congress);  encouraging a national debate on the reforms to shape and persuade public opinion;  building on existing social demand for reform/dissatisfaction with public services;  denouncing rent-seeking behavior of the opposition;  generating some immediate positive results which did not depend on legislative approval (implementing information technology to manage the payroll; tackling procurement) to gain credibility;  the strategic design of the reform (breaking away from Brazil‘s previous cen tralization, decentralization, centralization cycle by strengthening the core and simultaneous agencification); 15  negotiating concessions (Congress and Judiciary);  designing a media strategy centered on the reform‘s objective of increased efficiency and clie nt- citizen focus, new HR policies, yearly recruiting for stats careers, and defending national interest; and  disseminating New Public Management ideas through Brazil‘s main School of Public Administration ENAP. While the literature is divided on the reform dynamic after Bresser‘s departure in 1998, the positive reform results have been measured (Echebarría and Cortazar 2007) in terms of the decline in the public sector wage bill (from 50% of federal expenditure in 1995 to 35% in 2001), the average wage increase of public officials (21% in 2001), and the rising percentage of federal officials with a university degree (from 39% in 1995 to 63% in 2001). In addition, substantive policy change was achieved in organizational arrangements and personnel legislation (Gaetano 2003). The bureaucratic merit index (IDB 2006) which evaluates the degree to which ―effective guarantees of professionalism in the civil service are in place‖ and civil servants are ―effectively protected from arbitrariness, politicization, and r ent- seeking‖, ranks Brazil amongst the region‘s top performers. Importantly, factors that facilitate public sector reforms are likely to co-exist with those that running counter to them. The range of countries and of reform attempts discussed in the literature suggest that, although public sector reforms are frequently not politically attractive as such, they may nonetheless be initiated in response to crises or other pressures (see above). This implies that, reform pronouncements or initiation may be more frequent than real commitment to it (see the case of Ghana, Text Box 2). This is confirmed by the portfolio review below. The following section turns to discussions about reform implementation specifically. Text Box 2. Public sector reform in Ghana: Fluctuating political commitment and policy reversal Ghana has been reforming its public sector for over two decades. Reform efforts have included ‗first generation‘ reforms aimed at reducing the cost of the public sector, as well as ‗second generation‘ reforms that have sought to improve the quality of the public service, and to strengthen its capacity and performance. The most significant reform drivers have been economic and fiscal crisis; a concern about bureaucratic and institutional weaknesses; and donor pressure (Antwi et al 2008, Hutchful 2002, Stevens and Teggemann 2004). Facing negative GDP growth, Ghana pioneered World Bank-/IMF-sponsored structural reforms (cutting government employees and the wage bill) in 1983 under J.J. Rawling‘s military rule. But the interventions had little direct link to service improvement, some undermined capacity, and any gains were subsequently reversed (Kiragu 2002, Hutchful 2002). Ghana‘s second wave reforms thus focused on service delivery and capacity building (Antwi et al 2008). By the 1990s, Ghana‘s GDP growth had turned positive again. At this stage, key reform objectives included the following (Antwi et al 2008):  developing performance improvement plans (PIPs) and performance-linked reward schemes;  greater responsiveness to public and private sector needs and support of the national development agenda; and  increasing efficiency, transparency, competency, accountability, and cost effectiveness. By the end of 2000, 85% of all ministries, departments, and agencies (MDAs) had completed PIPs (Antwi et al. 2010). However, while some studies report marginal capacity improvement, the reforms did not 16 change organizational culture (marked by inertia, low motivation, and perceived corruption), nor improve service delivery or accountability (Kiragu 2002, Adei and Boachie-Danquah 2003). The main reform obstacles were:  lacking support from the Ministry of Finance (MoF) (no reform budgets and incentives for better performing MDAs (Antwi et al. 2010, Hutchful 2002));  resistance by senior civil servants (Hutchful 2002) and ineffective implementation (the body in charge of coordinating reforms lacked political authority);  the change in government in 2001 following elections (first orderly transfer of power from Rawling‘s National Democratic Congress (NDC) to the New Patriotic Party (NPP)) and the distrust with which the new government viewed the reform programs of the old;  the perception that the reforms were ―outside driven‖ and their association with redundancies and retrenchment, and  lacking continuous stakeholder engagement after initial wide consultation. In 2003 the Kufuor government presented its own approach. Its commitment towards a renewed civil service reform effort showed itself in the creation of a public sector reform ministry and putting Paa Kwesi Nduom, a former US educated partner at Deloitte & Touche, in charge in 2005. Despite being from the Convention People‘s Party (CPP), Nduom was close to the President. His main emphases were pay reform, retraining programs, performance, and dealing with sub-vented agencies and his reform strategy included:  Quick wins to win over stakeholders and show the public results.  Reforming sub/vented agencies first (given easier hiring and firing) making them financially independent from the government.  Negotiating with the unions and holding national and regional stakeholder consultations.  Using a sophisticated communication strategy involving the press and unions to bring civil servants on board. Nduom‘s presidential ambitions, however, created political jealousies and reform opposition by other ministers. After leaving his post to become CPP‘s presidential candidate in 2007 the reform momentum slowed. In some areas, such as pay reform, progress continued. A fair salary Commission had been created by then which was pressing ahead with their mandate. In January 2009 NDC opposition candidate, John Evans Atta Mills narrowly won Ghana‘s presidential election. Public sector reforms, including civil service reform remained on the agenda. However, the change meant yet again a new approach. One of the first changes included the replacement of the ministry of PSR with a secretariat reporting to the President giving it greater political weight and allowing less room for inter-ministerial competition. However, the Public Sector Reform Secretariat has weak technical capacity, there is a need to clarify the roles and authority for reform management and implementation of the various units across the government who share some responsibility in this area, and broad consensus and support for it is still lacking. Given the weak implementation arrangements it is doubtful that the government considers public sector reform a high priority. II.3 Public Sector Reform Implementation and Sustainability The factors that make initiating public sector reforms unattractive from a political economy perspective hold even more strongly for their implementation. Implementation (and non- 17 reversal) requires investments of political capital for uncertain gain. A substantial share of the literature on public sector reforms in LICs and MICs discusses instances of reform announcements and attempts that were not or only partially implemented (see Textbox 3, for a discussion of the Tanzanian experience. See also Heredia and Schneider 1998; Schneider and Heredia 2003; Geddes 199; IDB 2006; Kiragu and Mukandala 2003; de Renzio 2011). Even for developed countries many analysts argue that ―reforms are usually ‗incomplete‘, ‗half-way implemented‘, ‗only paid lip service‘ or are being underscored, but not implemented at all‖ (Toonen 2001, see also Brunsson 1989; Pollitt and Bouckaert 2000; March and Olsen 1995). Text Box 3. Public sector reform in Tanzania: from expectations to stagnation As Ghana, Tanzania has a history of civil service and PFM reform of over two decades – with mixed success. Supported by the World Bank and others, significant improvements in public management and accountability were achieved under President Mkapa‘s administration (1995-2005) – the first to be elected under multi-party rules (although the system remains one of uncompetitive pluralism with its dominant ruling party Chama cha Mapinduzi (CCM) since independence). While Tanzania was initially hailed as a best practice example (Morgan et al 2010) enabled by a more established sense of nationhood compared to many other SSA countries, since 2005 reforms have stagnated and some elements reversed. The reform rationales in the early 2000s were addressing inefficiency, ineffectiveness, and widespread corruption. By subscribing to public sector reform, Mkapa further hoped to establish positive donor relations and enhance his international credibility. At the time of his election the country was facing major challenges in service delivery, lacked revenue to fund services, and donors were reducing support. Economically, the reforms were initiated during a period of GDP growth (averaging > 4% for over 20 years), but continued high levels of poverty. The Public Service Management Department in the President‘s Office (PO-PSM) led the reform effort. Morgan et al (2010) credit the PO-PSM‘s leadership, organizational positioning, capacity and strategy for the reform progress under Mkapa‘s administration. Centralized coordination seems to have aided PFM reforms in particular. Other authors (Issa 2010), however, find that the high centralization of the reform implementation ―weakened the ownership‖ of the reforms in the implementing institutions and stress the need for more demand driven reform instead. The CSRs initiated in the early 2000s were to improve service delivery by streamlining ministries and improving management systems. The reforms focused on (Issa 2010):  performance improvement systems (2000-2004) and management culture (2005-2008); and  quality improvement cycles (2009-2011). A Public Service Act was adopted in 2002 (later amended), and so were various regulations allowing for competitive meritocratic recruitment and promotion. However, implementation slowed or was reversed under the Kikwete government (since 2005). Patronage remains pervasive. Similarly, the country‘s pay and incentive policies failed – political leaders have not hesitated to increase public sector salaries before elections. PFM reforms were to address financial malpractice and to enhance the institutional environment for aid. Important progress in expenditure control and greater efficiency in the allocation and execution of resources was achieved when leadership both by President Mkapa and from within the MoF was forthcoming. Key successes included:  an improved computerized integrated financial management system connecting all ministries 18 and regional sub-treasuries to a central system at the MoF;  a Medium-Term Expenditure Framework (MTEF) monitored by parliamentary committees;  the adoption of international accounting standards;  greater independence of the National Audit Office; and  increased compliance in the area of procurement. Nonetheless, efficient use of public resources still lacks behind. Procurement has remained susceptible to political influence and corruption. In addition, accountability remains weak and audits are rarely acted upon. The political economy factors which help explain the limited progress and sustainability of the civil service and (to a lesser extent) PFM reforms include:  inconsistent/absence of leadership from the top, particularly after 2005 and lack of government commitment to reduce bureaucratic discretion and corruption;  weak parliamentary oversight and government accountability and lack of pressure from opposition parties;  weak demand for performance management tools by managers (absence of incentives) and for good governance by civil society;  political rivalries: despite being from the same party, Kikwete wanted to distinguish himself clearly from his predecessor and set new priorities;  the government‘s perception that the political capital needed for effective reform implementation is larger than benefits of the reforms; and  the donor driven nature of the reform process. While there is little expectation that comprehensive civil service reform is feasible under the current government, a window of opportunity for more limited sectoral reform may open with increasing public demand for greater democratic governance, growing oversight, a stronger media, and Tanzania‘s membership of the East African Community. However, some public sector reforms are seen through. In such cases, political commitment to reform seems to have been grounded in a wider goal and could be sustained over time. Consistent political commitment is emphasized by Sarker (2006) in the case of Singapore, for example, where systematic reforms along the New Public Management model started in 1989. Singapore‘s leadership was driven by the twin goals of excelling in meeting the public‘s needs while at the same time ensuring that the state remains the guiding force in economic management (see also Cheung 2003). This type of motivation is likely to play a role in sustaining public sector reforms in several other authoritarian countries in Asia. An overarching goal (EU membership) has also been relevant in several Central Eastern European countries (CEECs). It helped maintain focus despite changing governments. Post-accession in 2004, however, reform efforts flagged in several countries (Meyer-Sahling 2009 and 2011; Dimitrova 2010; Trauner 2009; Barysch 2006).17 17 Meyer-Sahling (2011) finds that only the three Baltic States continued civil service reforms. Slovakia, the Czech Republic, Poland and Slovenia experienced reform reversals and Hungary experienced a reform reorientation. He explains this diversity of outcomes with the patterns of government changes after accession and concludes that few 19 A general factor that is seen to facilitate a closer connection between reform attempts and implementation is the level of economic and of public sector development of a country at the outset. In an attempt to explain the success of Singapore‘s public sector reforms and the failure of Bangladesh‘s public sector reforms, Sarker (2006) argues that reforms are more likely to succeed if the level of economic development is higher, a formal market economy exists, the rule of law is strong, and the level of administrative infrastructure/capacity to implement reforms is advanced.18 Similarly Hesse (1997) offers a country‘s stage of development as a possible explanation for greater reform success. In addition, Hedger and de Renzio (2010) suggest that higher income levels are significantly associated with higher quality PFM systems.19 However, while this argument suggests that MICs are at an advantage with regard to public sector reform implementation, it does not mean that public sector reforms are necessarily implemented and sustained. They may still be undermined if top-down commitment and/or bottom up demand cannot be sustained. A third factor that has been proposed by parts of the literature, but that has been questioned by others, is citizen demand – facilitated by the relative openness of governments to citizens‘ scrutiny. Tendler (1997) emphasizes that public interest to sustain reforms depends on the availability and dissemination of information about service delivery and expenditure. Khemani (2007) tries to define what information may have an impact on politicians‘ incentives to sustain reforms and improve service delivery. She finds that to impact politicians‘ incentives, information needs to be on development outcomes that are targeted by public policies; it needs to be statistically representative at the level of the lowest electoral district (or similarly disaggregated politically relevant jurisdiction in a country), be made available in a credible and regular manner, allow comparison between electoral districts, and allow monitoring within a district. Only then, she argues, can it potentially lead to voter coordination in evaluating governments and promote competition between districts to improve performance. Given these significant demands for information to be a basis for tangible leverage, sustained public pressure appears to be relatively rare (see also Booth 2011). Further cross-country empirical analysis will be needed explore if and what forms of demand from citizens more broadly or from specific groups such as private sector entrepreneurs play any role in shaping the incentives for political decision makers not only to announce reforms, but also to follow through. In terms of strategy, approaches that are seen to contribute to sustaining the ability to implement reforms include: generating winners quickly (Haggard 1997); keeping a low profile to minimize opposition; and thinking carefully about the level of difficulty and breadth of the reform. The focus on ‗quick wins‘ (improvements in bureaucratic processes that many citizens utilize, but which may often be lengthy and cumbersome and/or involve pressure to pay petty bribes) has become attractive across a number of development issues. However, quick wins do not necessarily open the way for more substantial public sector reforms (as the Tanzanian and Ghanaian experience indicate) – while they may still be valuable in their own right. At least one case study suggests that keeping a low profile of reforms can be valuable in navigating political economy dynamics. In his analysis of Uruguay‘s 1995-6 civil service reforms, Panizza (2004, p. factors ―could have locked in the level of professionalization that had been reached at the time of accession‖ (p. 231). 18 The importance of sufficient managerial and technical capacity to implement reforms is also stressed by de Renzio (2011) and Diamond (2011) who focus on the specific case of PFM reforms. 19 This needs further systematic testing. 20 1) stresses reformers‘ strategy of “a reform „without losers,‟ which, instead of gathering support for adoption and implementation, sought to minimize opposition.� II.4 Summing Up: PE Factors Favoring and Hindering Public Sector Reforms and Their Frequency across Countries Overall, the review of the literature – while noting its limitations – has several implications. While potential reform triggers and certain incentives are found in many countries (including fiscal pressures, recent democratization, economic transformation, and global or regional integration), there are fewer countries in which specific factors facilitating or favoring actual implementation are present. In many developing countries, factors working against public sector reform are significant, most importantly disincentives for political decision-makers to reduce rent-seeking and privileges enjoyed by individuals and groups whose support they rely on, the potential political cost of pursuing reforms against bureaucratic resistance, and the fact that political time-horizons are shorter than the time needed for most public sector reforms to yield tangible benefits. Table 1 summarizes the factors revisited in section II. The table groups these factors by the stage of public sector reform, and by their negative or positive nature. Furthermore, the third column provides a rough reflection of whether the factors are likely to be present in many countries, some countries, or whether they are relatively rare (‗few‘). Table 1: Summary of triggers, disincentives, and favorable conditions for public sector reform initiation and implementation Grouping of +/- factors Political economy factor Frequency across countries Factors which can trigger Fiscal pressures or crises Some or induce initiation of PSR Recent democratization Some Global economic integration Many Donor pressures & involvement Some Disincentives for Time horizon: Disincentive for current government Many implementing PSR to incur cost if benefit of PSRs cannot be reaped during tenure Bureaucratic resistance increasing the political cost Many of implementation Rent-seeking related to maintaining privileged Many insiders and clientelism Limited demand for public sector reforms Many Proportional electoral rules Some Conditions strongly or Rapid and sustained growth Few possibly favoring PSR Post-crisis re-building efforts Few 21 adoption and Gov’t seeking legitimacy Many implementation Legacies of meritocratic PS Some Majoritarian electoral system Some Programmatic parties Some Significant demand for public sector reforms Few Proximate factors and Policy entrepreneurs Some strategies that may Quick win strategy - facilitate PSR Piecemeal reforms - implementation Source: authors based on the literature review. Announcing public sector reforms can frequently be politically opportune. In many countries, announcing that civil servants will be asked to work better (and that corruption will be rooted out) generally resonates positively with citizens, as well as with the business and/or the donor community – with the possible exception of countries in which past reform efforts still make many stakeholders distrustful of such efforts. Consequently, an overview of existing case studies and the wider literature suggests that public sector reforms are more frequently initiated than followed through. Reform commitment can wane or priorities shift after elections or the departure of key policy entrepreneurs, or actual reform intent may be limited from the outset. At the same time, in some countries, propitious conditions for substantial implementation of public sector reforms prevail. As reflected in Table 1, several features of favorable environments emerge from existing case study materials. Two factors are potentially powerful, but likely to be present in a limited group of countries, while other factors are present in a somewhat wider range of countries. The former two include countries with sustained rapid economic growth (which has been limited to a relatively small number of countries over the past decade), and the presence of political leaders who are seeking to re-build states after a period of conflict. Facilitating factors that are present in a wider range of countries include the presence of programmatic parties, countries with (earlier) legacies of meritocratic public services, and, as a more proximate cause, countries in which policy entrepreneurs with proposals for public sector reforms and access to political decision-makers come to play a role. This paper does not attempt to weigh the different factors or to test their prevalence or relative weight statistically, which would be extremely challenging to do given current data limitations. However, the review suggests that there is no single overriding factor. The multiplicity of incentives for and against public sector reforms can make it challenging to assess political commitment reliably. It implies monitoring government reform commitment over time while scanning numerous variables for their relevance, strength, and durability. This task seems easier if one or two are dominant and harder if a range of contradictory factors compete (e.g. fiscal pressures and clientelist structures). The next part of the paper turns to the review of the Bank‘s own engagement on public sector reforms. It seeks to explore if and how this engagement was affected by political economy challenges related to these reforms, and gain some understanding of how Bank teams approach potential challenges. 22 III. Findings from the World Bank Portfolio Review The World Bank is one of the largest financial and technical advice sources on public sector reforms in LICs and MICs.20 Hence, a review of its portfolio is highly pertinent better to understand political economy drivers, how these affect public sector reforms and how they are or can be addressed. This portfolio review has three parts. First, it maps how widespread and varied in scope public sector operations are across countries. A narrow spread would suggest selectivity with regard to countries with (more) propitious conditions. If instead engagement is wide, unfavorable conditions are likely in some countries. The second part explores whether the expected relationship between more or less difficult country contexts and success of public sector operations is apparent for the World Bank‘s support to public sector reforms for the decade 2000-2010. Based on this, the part then revisits the frequency and nature of political economy problems encountered. Third, the portfolio review revisits what effort task teams make to understand and mitigate political economy challenges, and whether stronger efforts are associated with better results. The portfolio review is based on a set of project documents produced at the design stage and ex- post, project ratings and existing cross-cutting assessments, as well as on selected interviews with task teams.21 The review includes development policy operations, which involves general lending (or grants) to the government once agreed upon policy triggers have been met, as well as investment lending (or grant) operations, i.e. operations in which the government borrows to cover costs associated with public sector reforms specifically. III.1 Public Sector Reform Efforts: Spread and Scope The Bank‘s recent public sector portfolio suggests borrowing for public sector reforms has been spread across a large share of (potential) client countries. Of a total of 128 potential client countries or territories with populations of over 500,000 in 2000, only 41 (32 percent) did not borrow for some form of public sector reforms between fiscal years 2000 and 2010 (see Table 2).22 They include countries where Bank operations are generally non-existent or limited (e.g. Islamic Republic of Iran) as well as those with significant other operations (Togo, Uzbekistan, Panama, or Egypt).23 20 In 2009, funding from the World Bank‘s International Development Association (IDA) alone accounted for 40% of all donor funding on PFM reforms, nearly 14% of all donor spending on PS policy and administration (surpassed only by EU institutions and the United States), and nearly 26% of all donor funding on decentralization according to the OECD aid statistics database. No detailed sector breakdown is available for the International Bank for Reconstruction and Development (IBRD). 21 Documents reviewed include Project Appraisal Documents (PADs), Implementation Status Reports (ISRs) and Implementation Completion Reports (ICRs). Ratings refer to outcome ratings confirmed by the Independent Evaluation Group (IEG) unless noted otherwise. These ratings are composites of ratings for sometimes quite divergent project components and do not formally take the level of risk of an operation into account – either in terms of country context or in terms of the change being sought. 22 Based on board approval dates. 23 The count only includes Public Sector Governance Board (PSGB) mapped development policy and investment operations for the decade of fiscal year (FY) 2000-2010. If operations mapped to the Bank‘s Financial Management 23 Table 2: Share of countries borrowing for public sector reforms (approved between FY00 and 10) Total number of Number of Number of Number of Number of countries counted countries with no countries with one countries with 2 countries with 4 PSG operation or more operations or more or more operations operations # 128 41 87 61 19 % 100 32 68 48 15 Note: PSGB mapped development policy and investment operations only. Source: Business Warehouse data on PSGB mapped policy and investment loans/grants (active and closed). Nearly half of all potential client countries (with populations greater than 500,000) had two or more Bank operations. A smaller group (19 countries) had 4 or more. It includes several large and/or federal countries (Argentina, Brazil, Mexico, and Russia), some countries with special conditions (Chad related to the Chad-Cameroon pipeline), some fragile states with significant Bank engagement (Afghanistan, Kosovo, Sierra Leone, Timor Leste), and various others (Ethiopia, Tanzania, and Uganda). Regionally, the spread is relatively even, with some patchiness in Sub-Saharan Africa and relatively strong engagement with MICs in Europe and Central Asia (ECA) and Latin America (LAC). The Middle East and North Africa (MENA) region stands out with relatively few public sector operations (only 7 out of 15 countries/territories with at least one lending operation).24 This pattern is consistent with the notion that incentives to initiate public sector reforms are significant across LICs and MICs. As such, the decision by governments to borrow for public sector reforms (rather than using potential grant-funded resources) could be seen to signal political commitment to such reforms – although this is not consistently the case, as will be discussed below. Furthermore, a closer look at which countries borrow for public sector reforms at least in part contradicts a widespread perception, as reflected in Schacter (2001), that public sector reform efforts are pre-dominantly driven by donor pressures in countries dependent on external aid. In fact, only 25 percent of the countries with at least one World Bank public sector operation in the past decade had aid dependency ratios of more than 10 percent, and a number of MICs with little dependency on aid have sought multiple operations.25 Thus, while there is a group of low income, aid dependent countries that seek World Bank support for public sector operations, the larger share of such operations has actually been focused on countries with modest to low levels of aid dependency. This suggests that other incentives to initiate public sector reforms among those noted in section two, such as globalization and regional integration and ‗contagion‘, play a role in a substantial number of countries. unit (OPCFM), Fee-Based Services (FBS) and IDF (Institutional Development Fund) grants were included, an even wider spread would result. 24 Several countries in the region have fee for service arrangements with the World Bank which are not counted here. 25 Based on the Bank‘s Development Data Platform (DDP) database results for ‗Net Official Development Assistance (ODA) received in % of GNI‘ for the 87 countries with at least one public sector operation over the past decade. 24 The broad spread of public sector operations across countries and the apparent willingness of many governments to incur loans for supporting such reforms are in line with the suggestions from the literature reflected in section II that pressures to initiate public sector reforms are widespread, and are not limited to aid-dependent contexts. However, if existing research is representative across countries, then the degree of political will to actually follow through with reforms is likely to be less frequent. This would suggest that as a consequence, challenges should be prominent in number of the 86 countries which have been undertaking public sector reforms with World Bank support over the past decade. In addition to a broad spread across countries, many of the World Bank‘s operations in s upport of PSRs are relatively broad in scope, and cover less as well as more politically sensitive reform areas. For example, introducing new reporting and accounting formats does not touch fundamental power interests or rent-seeking incentives directly. Hence the political sensitivity of such reform components is limited. At the same time, the incentives to really use these technical improvements for better management of public funds are frequently limited, and consequently hoped for wider improvements may not (fully) materialize. Furthermore, some public sector reforms potentially increase the degree of control or the ‗safety‘ of central executives – such as better debt management – and can be attractive to political leaders wishing to avoid future crises that may threaten their own position. In contrast, public sector changes addressing public political cleavages, group and power interests such as decentralization, tax reforms, or de- politicization of civil service management as well as reforms seeking to reduce rent-seeking opportunities (procurement reforms, strengthening internal and external audit, etc.) are politically sensitive in many contexts. At the same time, some PSRs are more politically salient than others. For example, in many countries there is an explicit demand from voters and local elites to pursue decentralization, while there may be little explicit demand for implementing an MTEF or internal audit committees.26 Consequently, for politically salient reforms, political leaders may face a balance of incentives and disincentives, while for difficult and low salience reforms, disincentives dominate. As noted in section two, the degree of challenge also depends on the stage of reforms: the adoption of reform plans and in some countries even of certain laws tends to be less challenging than the actual implementation of reforms or of legal provisions. Many World Bank public sector operations combine less challenging with more ambitious reform areas and stages from a political economy perspective. Many operations stretch across a range of reforms – addressing up to 15 or more distinct sub-issues of public sector reforms. Public sector reforms are supported by investment loans (ILs) or development policy loans (DPLs) – with the latter typically larger in size and moderate complexity. Among 28 operations reviewed in greater depth, the narrowest had 2 public sector reform elements and the broadest 17.27 Nine out of 28 covered 10 public sector reform aspects or more. The financial size of 26 Even if citizens in principle prefer a ‗clean‘ public sector that uses resources efficiently, open demands if any are more likely to be directed at the adoption of anti-corruption legislation or limiting the size of government. 27 Such as introducing/strengthening a MTEF, reducing payment arrears, strengthening external audit, capacity building for the MoF and/or for line ministries, reforming civil service hiring, performance management, wage decompression, and introducing asset declaration for politicians and civil servants. Note that this refers to distinct reform issues touched on in project document descriptions and monitored outcome indicators, and not to the way these are clustered into project components. 25 interventions includes smaller loans or grants covering several areas (such as a $2m investment grant covering 8 areas), large and focused ($100m investment loan addressing 2 areas), as well as medium-large and diverse ($49m for an investment loan addressing 14 areas). Some of these differences are related to the countries‘ size, but they also represent considerable diversity as to the financial ‗firepower‘ brought to bear relative to the number of issues covered and the degree of challenge presented by these. Overall, current practice of Bank support for public sector reforms is spread widely across countries and across issues. Most operations tend to be moderately or significantly complex and virtually all touch on some sensitive aspects which require substantial political commitment to succeed. These indications of limited selectivity imply that an important share of public sector operations would be expected to encounter political economy challenges in implementation. The following section explores the actual success rate of operations according to IEG outcome ratings, the patterns in the distribution of more and less successful operations across countries, and the type of challenges encountered. III.2 Frequency and Types of Political Economy Problems: Patterns across Countries and Operations To explore the frequency of political economy problems in public sector reform support provided by the World Bank, this section examines what share of recently closed World Bank operations was rated less well, whether poor performance has been related to political economy challenges, and if so to what type of political economy problems. Graph 1: Distribution of outcome ratings – all public sector operations, ILs and DPLs Outcome rating distribution in Outcome rating distribution Outcome rating distribution %, all public sector operations in % for DPLs only in % for ILs only 45.0 50.0 46.1 40.0 41.0 36.2 45.0 34.5 40.0 35.0 33.6 40.0 35.0 32.9 30.0 n = 58 30.0 35.0 30.0 25.0 25.0 n = 76 20.0 25.0 20.0 15.0 20.0 13.8 11.2 11.2 15.0 12.1 10.0 15.0 10.5 9.2 10.0 10.0 5.0 0.7 1.5 5.0 5.0 3.4 0.0 1.3 0.0 0.0 HS S MS MUS US HUS 0.0 0.0 HS S MS MUS US HUS HS S MS MUS US HUS †Closed PSGB mapped projects with approval dates between FY2000 and 08 for which IEG outcome ratings are available. Ratings are as follows: HS = Highly Satisfactory; S = Satisfactory; MS = Moderately Satisfactory; MUS = Moderately Unsatisfactory; US = Unsatisfactory; and HUS = Highly Unsatisfactory. 26 Across all 134 operations for which IEG outcome ratings are available, slightly over 40 percent are rated satisfactory or highly satisfactory; slightly over 30 percent are rated moderately satisfactory, and the remaining quarter is rated moderately unsatisfactory or worse. As reflected in Graphs 1a)-c), the rate of less satisfactory operations is somewhat higher for investment loans (29 percent) than for development policy loans (20 percent). A first observation is that given the broad spread and scope of operations noted in the previous section and widespread disincentives to pursue public sector reforms that emerges from the literature discussed in section II, the actual success rate of Bank public sector operations is rather high, with two-thirds of operations rated as moderately satisfactory or better. Three reasons appear to account for this fact. One is a higher success rate for Development Policy Loans – which tend to be focused on specific one-off actions, such as legal changes. Moreover, core elements of these actions have to be implemented as ‗prior actions‘ – i.e. prior to the operation being approved and becoming effective. However, a 71 percent success rate for investment lending operations is still high relative to the expectations resulting from the literature. A second reason is the fact that Bank operations are rated according to their specific development objectives, which may be achieved even when overall public sector transformation is not. Nearly 40 percent of all operations and over 45 percent of investment operations are rated as ‗moderately satisfactory‘ which implies the achievement of most of the development objective, but often weaknesses in individual components and often also in the wider impact of reforms. This may also mean that despite success in an operation as such, progress in the overall quality of public sectors frequently remains more limited. As noted by IEG‘s 2011 Annual Review, while a rather high share of public sector operations is rated as successful, overall progress on public sector reform was only observed in 47 percent of completed and reviewed country programs (IEG 2011: 8). A third factor is the fact that a good share of public sector operations has been focused on more propitious environments – while at the same time this has been combined with a better than expected performance in more challenging environments. To a degree, the expected pattern of project performance - more likely to be good in better and worse in worse environments - broadly holds also for Bank operations, but the relationship is less strong than might be expected. This is reflected in graphs 2 and 3 (covering all public sector operations and IL operations specifically).28 Operations that are rated as moderately unsatisfactory or worse are more likely to have been implemented in a more difficult environment than operations that were implemented in more propitious contexts; and there have been no unsatisfactory operations in a propitious environment. As would be expected, the association between the environment and operational outcome ratings is somewhat stronger for ILs and weaker for DPLs. 28 Note that four of the countries appear in more than one category, i.e. they had multiple operations with different outcome ratings. The chart reflects the governance indicators for ‗government effectiveness‘ from the World Governance Indicators (WGIs) compiled by the World Bank Institute (WBI) for 2000 and for 2010. While methodological shortcomings of the WGIs have been widely discussed, they offer a useful rough gauge of country context. 27 Graph 2: Prior government effectiveness and operational performance of public sector governance operations Prior government effectiveness and project outcome ratings HS 2 S 1 Project outcome ratings MS 0 0 10 20 30 40 50 60 70 80 90 100 -1 MUS US -2 n=134 -3 HUS Worldwide Governance Indicators for Government Effectiveness (2000) Note: project ratings range from: HUS = highly unsatisfactory, to HS = highly satisfactory. WGIs rate countries on a scale from 0 = no government effectiveness to 100 = very high government effectiveness. Colored arrows indicate governance percentiles (0-10, 10-25, 25-50, 50-75, 75-90 and 90-100) as defined by the index. Coverage: 134 public sector mapped development policy and investment lending operations implemented in 67 countries. Several countries had more than one operation, sometimes rated differently. Each operation is counted separately. Graph 3: Prior government effectiveness and operational performance of public sector governance operations -- investment loans only Prior Government Effectiveness and Project Outcome Ratings for Investment Operations HS2 S 1 Project outcome ratings MS 0 0 10 20 30 40 50 60 70 80 90 100 MUS -1 US-2 HUS -3 Worldwide Governance Indicators -- Government Effectiveness (in 2000) Coverage: 58 public sector mapped investment lending operations implemented in 43 countries. Several countries have had more than one operation, sometimes rated differently; each operation is counted separately. 28 However, there are two important elements of divergence from the broad pattern of an association between the country environment and the likelihood of operational success: (i) well rated operations are found across the entire range of countries, and (ii) in particular, there are a number of well rated operations in environments typically classified as very difficult (top middle and top left-hand corner of Graphs 2 and 3). Well rated operations in challenging environments include a number of operations in post-conflict countries such as Afghanistan, Kosovo and Sierra Leone, or Haiti.29 This observation is consistent with the observation noted in section two that some post-conflict environments entail significant commitment by government to re-build states. It may also be due to significant external incentives and involvement in some post conflict environments (see Fritz and Fialho Lopes 2012). Some of the well rated operations have supported more limited reforms in favorable environments. Examples are the Brazil Fiscal and Financial Management Technical Assistance Loan (TAL) or tax administration reforms in Bulgaria.30 Revenue administration reforms are sensitive from a political economy perspective but appear to have been manageable in the case of Bulgaria, a MIC with relatively strong government commitment (linked to the country‘s EU accession). Remarkably, the operation in Bulgaria very explicitly addressed the risk of internal stakeholder buy-in through careful change management as noted in the Implementation Completion Report (ICR), focusing on early gains to build confidence in the reforms and on involvement of stakeholders in the reform process.31 Even in successful operations, political economy challenges appear and come to bear on the relative success of project components. For example, the Brazil operation supported among other components improved budget integration (capital and recurrent) and the implementation of the Fiscal Responsibility Law (and a related Law on Fiscal Crimes) requiring an enforcement of borrowing limits across all sub-national governments. While the latter is challenging from a political economy perspective, the government was very strongly committed after experiencing successive macro-fiscal crises and the necessary legislation had been passed. In contrast, for the government enhancing the quality of public expenditures was less urgent and salient, and the benefits less likely to materialize within a short-term time horizon. This component achieved little traction and legislative reform in this area remained blocked. Furthermore, for 28 operations, project documents and Implementation Completion Reports were reviewed in greater depth to explore further whether and how frequently political economy challenges become a major obstacle. This in-depth exploration focused especially on less 29 For these countries, ex-post assessments stress that progress remains uncertain and subject to reversals. For Haiti, the assessment explicitly assesses operational performance as the achievement prior to the earthquake of January 2010. The natural disaster and its impact on administrative structures reversed many of the gains made. 30 The Brazilian operation engaged in four sub-areas, and the Bulgarian in three; compared to a maximum of 17 distinct reform areas among all reviewed operations. 31 The ICR for the operation attributes government commitment to the EU accession process and pre- and post- accession prodding. It notes that even so there was a negative publicity campaign against the revenue authority towards the end of the project period which raised concerns about sustainability. This is consistent with considering tax administration reforms as sensitive from a PE perspective, as they are likely to touch on economic (and power- related) vested interests. 29 successful projects, as for these the need to understand what could have been done better is most relevant (Table 3). Table 3: In-depth review sample Selection of ICRs reviewed – frequency of political economy factors assessed as key constraints Total DPLs ILs Rated moderately 10 -- 10 satisfactory or better Rated moderately 18 10 8 unsatisfactory or worse Operations rated MUS or worse for which pol econ 14 (78%) 6 (60%) 8 (100%) factors are significant For operations rated moderately unsatisfactory or worse, political economy challenges are described as key obstacles with very high frequency in Implementation Completion Reports. For the 18 less successful operations reviewed, more than 75 percent of ICRs providing ex-post assessments of these operations point to political economy factors as key obstacles. This includes all of the poorly performing Investment Lending operations. 32 As to the specific political economy factors inhibiting project success, the ex-post reports highlight:  underestimation of reform resistance (whether open resistance e.g. from trade unions, or more covert resistance within the bureaucracy or from vested interests);  overestimation of governments‘ reform commitment – including its ability to take on vested interests; and  failure to engage with stakeholders outside the acting government. This left several operations moribund when governments changed. While Implementation Completion Reports and IEG assessments confirm that political economy challenges emerge and are a major source of project under-performance, they offer rather less clarity with regards to what has made successful operations feasible even in challenging environments. Positive political economy drivers or successful strategies for managing political economy challenges receive less attention. Issues of potential opportunities in post-conflict environments as well as the question how in some instances teams seem to have been able to successfully manage political economy disincentives to reforms do not receive much explicit attention. 32 Of each type of operation, slightly over half of the operations were reviewed and compared to ten operations that performed moderately satisfactory or better (representing one third of the well-rated group). This was judged to be an adequate way to check for differences between well and less well performing loans. The inclusion of additional ILs or DPLs performing well could be considered; however, the additional insights to be gained are likely to be limited (while there would be some gain in robustness). The main other critique raised was that projects were overly complex in the areas they were seeking to address or with regard to implementing arrangements. 30 Furthermore, Implementation Completion Reports and IEG assessments suggest that in many ‗moderately satisfactory‘ operations in which several elements do not perform well or impact and sustainability remain uncertain, efforts to assess and manage political economy context and drivers have been insufficient. These ex-post assessments imply – but do not explicitly discuss – that adjustments in design and better management would be possible and might be important to strengthen the impact, sustainability, and development effectiveness even of moderately satisfactory or satisfactory rated operations. For example, the Implementation Completion Report for an operation in Zambia rated as ‗moderately satisfactory‘ notes that the original design was overly ambitious given limited political will and an upcoming leadership transition. For another operation in Southern Africa (a $41.2m IL) rated as ‗moderately satisfactory‘ 33, interviews with the current task team working on a follow up operation suggest that many envisaged reforms were carefully planned but never fully implemented, and that the political will to follow through was overestimated. A number of Implementation Completion Reports for moderately satisfactory or better rated operations point to significant risk for the development impact and the sustainability of achievements, and several reports cite continuing political economy challenges. For example, an operation might succeed in strengthening the capacity and professionalization of a country‘s Supreme Audit Institution (SAI), but whether these gains will persist and lead to better accountability de facto depend on continued support by political stakeholders (political parties, presidents), and possibly also on citizen demand for such changes (as the case with regard to an Economic and Financial Management Project in Central America).34 However, as seen above, overall evidence on the role and importance of citizen demand in sustaining reform efforts so far is slim – in part as it is hard to come by. A brief exploration of the correlation between project outcome ratings for public sector operations and of overall changes in government effectiveness over the decade 2000 to 2010, shown in Graph 4 also suggests that the association is rather weak. While there are many instances of the expected association (well rated operations and improving country contexts), there are also many contradictory cases (well rated operations, but deteriorating contexts or more poorly rated operations despite overall improving country contexts).35 This is consistent with the concern that even when public sector operations can be implemented successfully, this does not necessarily translate into broader positive changes of overall public sector performance. 33 IEG outcome rating (http://wbln1023.worldbank.org/oed/oeddoclib.nsf/DocUNIDViewForJavaSearch/8525682E0068603785257474004 54F51?opendocument) 34 FY01-07, $19m IL, IEG outcome rating MS. 35 Incidentally, the best rated operation in the sample, a DPL in a MENA country, the case of the greatest deterioration in government effectiveness between 2000 and 2010 (upper left-hand point in graph 4). 31 Graph 4: Outcome ratings of public sector operations and change in overall government effectiveness 2000 to 2010 Change in government effectiveness 2000 to 2010 and public sector operation outcome ratings 2 HS S 1 Project outcome ratings 0 MS -35 -25 -15 -5 5 15 25 35 -1 MUS -2 US -3 HUS Change in government effectiveness (WGI data) 2000 to 2010 Note: project ratings are as follows: HUS = highly unsatisfactory, US = unsatisfactory, MUS = moderately unsatisfactory, MS = moderately satisfactory, S = satisfactory, HS = highly satisfactory. Changes in overall government effectiveness were calculated based on WGI data for 2010 and for 2000; and are provided as percentile changes. The following section explores how teams capture and mitigate political economy challenges and assesses the impact the level of attention paid to political economy factors has on project performance. 32 III.3 Attention to Political Economy Drivers: Evolving Practices and Potential Pay-offs Given the considerable importance that political economy factors play in enabling or hindering progress of public sector reforms, an important question is how such factors are or can be taken into account in designing and implementing operations. This section addresses current and emerging practices and this regard, and reviews tentative evidence on whether increased or more systematic attention pays off in terms of better support and better results achieved. III.3.1 Understanding and acknowledging political economy dynamics and risks What efforts are teams currently making to assess and address political economy risks and variables to calibrate and manage World Bank support? And is there any (tentative) evidence that more attention to and more analysis of political economy factors can pay off in terms of operational performance? To investigate these questions, this section draws on three sources: Implementation Completion Reports of closed operations and semi-structured interviews with 13 team members and task team leaders of eight ongoing public sector operations;36 Furthermore, it revisits 17 public sector operations initiated in fiscal year (FY) 2008 that had been included in the 2009 Quality Assurance Group (QAG) review of attention to governance and anticorruption (GAC). The section compares the level of attention to governance and political economy at the project approval stage to project performance as of 2011. World Bank teams feel that they generally understand political economy factors and anticipate related operational challenges. That said, for a majority of the teams interviewed, efforts to understand political economy variables were limited to internal team discussion about the operation, but were not necessarily made explicit. Unsurprisingly, operations vary as to the frankness and the accuracy of the ex-ante assessment of political economy-related risks, and assessments ex-post of whether these were fair. Furthermore, the way in which risks are assessed (and suggested to be mitigated) is influenced by bureaucratic incentives. Most importantly, once planning for an operation has been started, there is a strong incentive to achieve its approval – which may in turn incentivize teams to downplay risks that cannot be readily mitigated. There can also be various pressures to aim for relatively ambitious designs.37 As in the literature, the risk of fluctuating political reform commitment featured prominently in Bank documentation and interviews with task teams. 38 For several closed operations – both satisfactory and unsatisfactory— government commitment was explicitly assessed as high ex ante, but turned out to be lower de facto – with the subsequent Implementation Completion Report arguing that the risk of uncertain or fluctuating commitment could have been recognized at entry. Several interviewees explained that the level of government commitment can easily be 36 Ongoing operations were selected for easier access to teams and institutional memory. Interviews were conducted in person and by phone. 37 Both of these issues are to some degree addressed by recent internal reforms: the introduction of an Operational Risk Assessment Framework (ORAF) and the emphasis on seeking feasible reforms, set out in the 2012 PSMA. 38 Apart from political economy, other key challenging factors that several teams referred to include civil service incentives/pay and capacity constraints, as well as the role and incentives of consultants in the implementation stage. 33 overestimated given it typically evolves as public sector reforms move from adoption to implementation stage or as the result of government changes. Evolving political economy dynamics repeatedly mentioned in the interviews included electoral cycles, veto players, challenges related to cross-government coordination and executive-legislative relations.39 This is consistent with the finding set out in section two that windows of opportunity are limited in the case of many public sector reform efforts – and the issue of likely changes is sometimes, but not consistently, addressed at the operational design stage. Implementation Completion Reports tend to criticize an overoptimistic assessment of commitment and opportunities and failure to anticipate veto points – especially legal reforms being blocked in parliament. Some operations argue that they do not expect problems even if changes in government are looming (or do not address this issue). Such challenges subsequently became manifest during implementation.40 While the sample size is too small to allow broad generalization, teams seem to be more realistic and frank in propitious environments (where a case can be made that risks are manageable) as well as in recognized difficult environments (where substantial risks are obvious), but less so in ‗intermediate‘ environments (including reviewed operations in Sub-Saharan Africa, Central America, and South Eastern Europe). Furthermore, in difficult environments the focus and scope of operations can still remain unaligned to the full extent of risks, despite teams‘ awareness of significant challenges (e.g. operations implemented in a post-conflict country in South Eastern Europe, a low income Latin American country; and a country in the Indian Ocean). Furthermore, explicit risk mitigation efforts prove to be more challenging than anticipated. Several ex-post reviews raise the point that components meant to target the demand side and/or accountability institutions remained underfunded or otherwise limited in implementation. Only one project was explicitly lauded ex-post for strong understanding of context and effective engagement with government counterpart and non-government stakeholders (an operation in East Asia). The latter was also praised for having a deliberately limited scope, learning from earlier operations.41 In several cases, continuity of task team members generated greater awareness of reform dynamics and possible pitfalls. Motivated by previous experience with variable reform commitment, some teams reported more explicit efforts to capture political economy issues and to take these into account in project design. In one case, the team commissioned an explicit political economy analysis of public sector reforms during the implementation phase of the most 39 One approved operation remained ineffective blocked in the country‘s parliament due to a combination of difficult institutional rules and inability of the executive to persuade opposition MPs to support approval of the loan. This eventually came ‗unstuck‘ after a hiatus of 15 months. 40 The PAD for a PSCAP in a Southern African country rates the risk related to an (already expected) change in government as ‗negligible‘ arguing: ―The program sets out to implement some of the be as committed as the current one to the program more difficult and strategic, win-lose changes in (elections will take place before the end of Phase I) the first phase, leaving more of the win-win changes for later phases. Will and capacity to manage change will be embedded in the leadership of the CS during the first phase. The focus on capacity building is a politically neutral approach to public service reform; any government is likely to welcome a public service able to perform well.‖ In fact the change in government brought considerable difficulties for the operation and its development impact. 41 ‗The project was focused on a small set of priority interventions for which government agencies‘ commitment was high …‘. As noted in section III (i) this operation combined a narrow scope with a considerable scale of $100m. 34 recent World Bank support to public sector reforms against the backdrop of fluctuating government commitment. However, long-standing involvement of team members can also have drawbacks. In one case, the World Bank had a significant and long track-record of supporting public sector reforms. Because these had been seen and reported as successful, emerging disappointment about results and heightened concern about governance was challenging to acknowledge for some time. This became easier as the team‘s composition changed. In sum, project documents and interviews reflect that teams make some efforts to understand political economy challenges and likely dynamics over time. However, to date, the majority of these efforts are limited and implicit in nature and do not involve explicit analysis. Interviews and the review of ex-post assessments of operations also reveal incentives to under-state political economy risks in ex ante project documents; as well as some challenges with the implementation of mitigation measures aimed at increasing the demand from citizens and NGOs for public sector reforms. In several cases, experience gave Task Team Leaders a sense of potential challenges, as well as how the government‘s commitment was evolving over time. The final sub-section of the portfolio review drills deeper into the question of whether there are indicators that greater attention to political economy variables pays off in terms of operational performance. III.3.2 Potential pay-offs from more explicit attention to political economy Several teams reported explicit adjustment in operational design and/or implementation based on their discussions about likely political economy challenges. Two teams that had invested in political economy analysis suggested the following implications for the project design: seeking a deliberate balance across project components, agreeing on rules of the game and being clear and open about these, making efforts to reach out to the parliamentary opposition42, agreeing with the government to undertake regular surveys of citizens‘ opinion on satisfaction with service delivery, and narrowing the scope of the project to what was deemed feasible based on lessons learned from earlier projects and the current analysis of reform context. Both operations that involved explicit political economy analysis ex ante are currently rated as ‗satisfactory‘. While not hard evidence, it supports the idea that taking political economy factors into account can be associated with better performance. This is also consistent with evidence from a sample of 17 public sector operations that were assessed in 2009 with regards to their relative attention to governance and political economy factors, and that were revisited with regards to their current performance for this paper. In 2009, the Bank‘s internal Quality Assurance Group (QAG) undertook an initial review of how key aspects of governance and anticorruption are integrated into lending operations.43 Three dimensions were covered: attention to Governance and Political Economy (GPE); Fiduciary Aspects (GFA); and the demand side of governance (DSG). Operations were rated on a six point scale ranging from (1) very responsive to (6) highly unresponsive. The GPE dimension was assessed by examining the adequateness of analysis of stakeholder interests and their underlying incentives; of formal institutions and informal rules; of the project‘s vulnerability to rent- 42 One team made an effort to understand and manage PE dynamics, but new political forces emerged unexpectedly. In another country, in contrast, the potential alternative party was readily identifiable. 43 In the context of the Bank‘s Governance and Anti-corruption Strategy and Implementation Plan launched in 2007. 35 seeking, elite capture, and clientelism; and of the design to respond to the governance and political economy constraints. Among an overall sample of 180 operations drawn from 344 newly approved operations in fiscal year 2008, the report included 17 public sector governance (PSG) operations (5 DPLs and 12 ILs). These were found to be somewhat more frequently responsive to GPE dimensions than the sampled Bank operations overall (see Table 4). Table 4: Operations included in the QAG GAC review and responsiveness on the GPE dimension Responsive to GPE factors PSG total PSG PSG all reviewed operations (N=17) DPLs ILs (N=180) (N=6) (N=11) # % # % # # Very responsive (1 &2) 32 18 6 35 4 2 Somewhat responsive (3) 50 27 4 24 1 3 Limited effort (4) 56 32 6 35 1 5 Unresponsive (5 & 6) 42 23 1 6 0 1 This paper compares the operations‘ level of GPE responsiveness at entry with their most recent status or outcome rating to explore any relation between the two. Table 5 summarizes the relationship between the QAG panel GPE ratings and status or outcome ratings. It shows that operations seen as ‗somewhat responsive‘ or better to GPE issues outperformed those with limited effort or being unresponsive to GPE issues. Table 5: Summary – QAG attention to GPE and project performance ratings Implementation or outcome Implementation or outcome rating MS or better rating MU or worse Very or somewhat responsive to GPE total PSG 9 1 of which DPLs ILs 5 4 0 1 Limited effort or unresponsive to GPE total PSG 4 3 of which DPLs ILs 1 3 0 3 The robustness of this finding is limited given the small sample size, sub-sectors of public sector reforms covered, as well as the variance in the status of the ratings (some operations are ongoing, others have been closed).44 Within these limitations, the data indicates that ignoring GPE issues is associated with a higher risk of poor project performance and that when greater attention is 44 More continuous monitoring is probably valuable – including additional years and operations – to generate a more robust sample over time, and one that potentially allows analysis also for specific lending instruments and sub- sectors. 36 paid to GPE factors, operational performance tends to be better.45 Furthermore, the break-down into DPLs and ILs indicates that the potential downside of neglecting political economy drivers has been higher for investment operations than for development policy operations. This is consistent with expectations, given the typically longer implementation periods, and the greater challenge in achieving the expected outcomes defined ex ante that are inherent in investment operations. Considering specific cases within the sample suggests that attention to GPE factors helped to support better results in several less propitious environments such as a country in West Africa or a South East Asian country, but not in all – e.g. not in another West African country. In a more propitious environment such as a Latin American MIC, in turn, an operation was able to achieve good results despite paying little attention to GPE dynamics. A continuation of the QAG GAC assessment and further monitoring of results would be highly desirable with a view to following over time and for a growing number of operations whether attention to GPE and other governance dimensions ‗pays off‘. Overall, the evidence from reviewing project documents and ex-post assessments, interviews with task teams and revisiting the QAG sample of public sector operations suggests that investing in more explicit attention to political economy and developing more political economy responsive approaches to public sector reform support can pay off, especially in environments where reform commitment is uncertain. Importantly, the interviews indicate that a better understanding of political economy factors does not automatically translate into a more ‗political economy savvy‘ approach. Task teams can face incentives that make them reluctant to limit the scope or hold back on an operation, while some task teams are pushed to develop overly ambitious designs. Thus, looking forward, it would be helpful for the public sector practice to track not only whether more explicit analytic work as such is undertaken and pays off, but also whether resulting insights are in fact translated into operational design (and possible obstacles for doing so). IV. Making Public Sector Reforms Work: Preliminary Conclusions This paper has investigated the importance of political economy challenges and dynamics in an effort to understand why progress on public sector reforms has been difficult in many countries, and also why in some cases public sector reforms progress despite typical obstacles. It analyzed how political economy factors affect public sector reforms by looking at the literature and evidence from a decade of World Bank projects supporting public sector reforms. Importantly, while the focus of this paper is primarily on the political economy dynamics surrounding public sector reforms and less on issues with the content of these reforms that are discussed elsewhere (World Bank 2012, Andrews 2009 among others), the paper and its findings are fully aligned with the view that flexibility and innovation in the content is essential to enable 45 No indication was found that responsiveness to GPE is targeted at countries where successful reforms are less likely. 37 greater feasibility as well as sustainability and impact of efforts to improve public sector performance. IV.1 Summary of Key Findings: Public Sector Reform Trajectories The literature underscores the notion that we should expect public sector reforms to be challenging; while indicating that many countries will nonetheless be motivated to initiate public sector reforms rather than ignore them entirely. It points to a variety of structural drivers for public sector reforms such as international or regional integration, but also fiscal crises that lead to a search for greater efficiency while at times reducing the power of vested interests, and recent democratization which can generate promises of more service oriented, less corrupt, and more accountable public administrations. Aid dependency is also assumed to play a role (a factor that is relevant for about a quarter of World Bank public sector operations). Politicians typically face disincentives to implement public sector reforms effectively given that this frequently requires taking on a variety of vested interests, while the benefits from public sector reforms tend to be long-term, uncertain, and often not very visible to voters, and factors consistent with broader theorizing about ‗limited access orders‘. The evidence both from the literature and from the portfolio review suggests that ‗competing‘ political economy drivers (some in favor, others against public sector reforms) frequently coexist. However, there are also countries and periods (‗windows of opportunity‘) when it is possible to achieve reform progress. In some instances, political economy factors align in a favorable way including structural conditions such as an experience of recent fiscal crisis combined with a growth opportunity creating space for reforms, and/or interested political leadership, sufficient incentives of key stakeholders, and a capable policy entrepreneur. In other cases, seeking recognition as a state (such as Kosovo) or accession to the European Union have been powerful anchors. Political and security imperatives to build sufficiently capable states are historically an important political motive to pursue public sector reforms that can also be observed in some recent cases (for example Rwanda). Table 6 summarizes the type of factors which make reforms more or less likely to be initiated and successful. Based on the discussion of different cases in the literature (section II), the table sets out three stylized trajectories: a trajectory in which reform efforts are not likely to emerge, a trajectory of reform initiation that is then not (or only partially) followed through, and a trajectory of reform success across initiation and implementation. Typically, trajectory one countries are likely to overlap with the group of countries not seeking Bank (or other donor) engagement on public sector reforms; while trajectory two and three countries account for the large number of countries that have engaged on public sector reforms and have sought Bank support over the past decade. 38 Table 6: Political economy factors across three stylized trajectories Reform less likely to Reforms likely to be Reform more likely to be emerge initiated but not initiated and implemented successfully implemented successfully or only partially Domestic/External  weak  present and significant  present and strong Political/Economic e.g. some ongoing e.g. opportunities for reform pressures and regional integration or international trade, EU incentives diffusion of reform accession prospect ideas, aid dependency Incentives/Constraint  high degree of  significant clientelism  limited degree of s embedded in the clientelism  short-term rent seeking clientelism political and  short-term rent dominant  belief that collective institutional and seeking dominant  significant veto points long-term gains administrative order  high number and and players who outweigh short-term intensity of potential oppose reforms gains for a small veto points  limited public and/or group  lack of public or private sector demand  low number and private reform for reform intensity of potential pressure veto points  government pre-  high degree of public occupied by survival or private reform and/or outright pressure predation Political dynamics  government does not  government initiates  government is under within relevant feel pressure to reform reforms, but does not increasing pressure for government(s)  a new government have the political change; or comes into office and capital to pursue these  a new government is uninterested in PSR against the opposition with less stake in the  no reform champion or of vested interest existing system takes natural counterpart for  government primarily office; or PSR exists pursues PSRs for  a reform champion public relations with emerges with the domestic stakeholders resources and political and/or donors skill to push ahead with PSR Reform less likely to Reform more likely be sustained to be sustained Government  weak, fluctuating,  some stakeholders in  sustained, strong, commitment (linked reforms not linked to government are likely reforms linked to to implementation broader agenda to be committed, but broader long-term arrangements, they cannot carry the goal coordination) reform through Public  weak  potentially some  strong, sustained Interest/Pressure interest but not effective Capacity (linked to  weak(er) capacity  can vary, usually  higher pre-existing state of development) weaker capacity Reform  no reform strategy  ineffective  effective (quick wins, Strategy/leadership emerges minimizing opposition, building 39 support coalitions, etc.) Using portfolio ratings reviewed in section III as a proxy, a very rough grouping of client countries against these trajectories shows that among countries borrowing for public sector reforms, somewhat over one third have one or multiple operations that are rated as satisfactory or highly satisfactory, and a further group of similar size have moderately successful operations, which situates them between the second and third trajectory. Finally, somewhat over a quarter of countries clearly fall into trajectory two, with operational success being rated as moderately unsuccessful or worse. Reviewing operational performance against a crude measure of country context furthermore indicates that operations that there is a share of public sector operations that is assessed as successful in environments that are typically judged as highly challenging. Several of these cases have combined very low levels of government effectiveness at the outset combined with significant government commitment – in several cases in post-conflict settings and entities seeking recognition as states. Thus, the strength of incentives for reforms and of political commitment rather than general country conditions such as income levels or the ex ante quality of the public administration appears to play an important role for whether a country falls more into trajectory two or trajectory three in terms of its public sector reform potential at a given moment in time. Some teams have begun to pay greater attention to political economy variables, including initial piloting efforts of systematic analysis. There are indications that such efforts pay off in terms of project performance tipping the balance towards better performance in intermediate environments. Given small numbers and the fact that these efforts are recent, it is as yet unclear whether this can also help to boost the sustainability and wider impact of reform efforts. The importance of good strategic approaches to public sector reforms and the need to utilize windows of opportunities emphasized in the existing case study literature suggests that the potential progress with and impact of public sector reforms is not pre-determined by the country context alone, but can in fact be influenced by how savvy supporters of reform utilize opportunities. IV.2 Implications for External Support to Public Sector Reforms and the Knowledge Agenda There are several implications for a research agenda as well as for an agency such as the World Bank which supports public sector reforms across a range of country environments. Continuous investment in knowledge will be important to provide practitioners with the best evidencing about what is likely to work and good strategies for stretching the boundaries of public sector reforms in varying contexts. To develop this knowledge calls for a clearer set of hypotheses, more empirical data collection, and more empirical testing, both quantitatively and qualitatively; particularly, on whether and how countries have advanced on public sector reforms ‗despite the odds‘ of typical political economy constraints, and when and how these reform efforts have translated into a substantially better public sector performance. Qualitatively, a more systematic, ‗medium n‘ exploration of success episodes of similar reforms in different country 40 contexts as well as expected and unexpected failures of comparable reform efforts in similar countries with close attention to political economy aspects in terms of interest groups as well as the role of individual stakeholders would be helpful. Such comparative process tracing could also shed light on the relative weights of political incentives versus bureaucratic resistance to reforms, as well as on how sufficient political commitment comes about to follow through with reforms. While closing these gaps in evidencing will take time, the analysis presented allows at least some implications for external support to public sector reforms by the World Bank and other donors. (i) First, it is important to identify country contexts more clearly as propitious, mixed, or unlikely to allow reforms, and to differentiate expectations and approaches to reforms more strongly between these. As the portfolio analysis has indicated, some environments typically considered as very challenging may in fact offer significant windows of opportunity for reforms, albeit with higher risks; while some intermediate environments can be most challenging. Given the World Bank‘s mission and reach, it should continue to engage in countries that offer less propitious environments for public sector reforms – but it should do so with a very clear view of the constraints and limited opportunities, and approaches that match such contexts. Project design should include a clear indication of the context, the likely feasibility of the changes sought, and especially their likely real impact and sustainability. (ii) Second, in contexts with some but limited potential for progress on public sector reforms, political economy analysis can play an important role in identifying which reform options maximize a ‗good fit‘ and a ‗feasible stretch‘. The aim is to avoid a wasting of resources and efforts on reforms that ultimately are not implemented and/or sustained – of which there is a high risk and likelihood in such contexts – while identifying what can be done to concretely move public sector performance forward. (iii) Third, more attention should be given to the nature of ‗windows of opportunity‘, and reforms should be tailored to these. Windows of opportunity may open for a short time and once they close again due to new elections, the removal of a policy entrepreneur, etc. opportunities for reform recede. This underlines that such windows should not be ‗frittered away‘ with overly lengthy preparation of (large-scale) reforms, but should be used to prepare and implement limited but tangible change. It may also be useful to have staff in a country continuously to be able to respond to such opportunities promptly when they arise. (iv) Fourth, reform support needs to strike a challenging balance between targeting feasible progress and having bite in terms of real change, which frequently means limiting discretion. Support to reform planning or preparation stages as well as predominantly technical reforms or de jure changes that do not significantly threaten the discretion and rents of elites are more likely to be successful since the balance between political gain versus pain is typically more favorable. The benefit of such an approach, especially in difficult environments, is the potential opportunity for incremental gains and of building engagement; while direct attempts to reduce discretion may either be rejected or thwarted. At the same time, an exclusive focus on reform preparation stages or on 41 technical issues risks leaving ‗real‘ problems unaddressed, leading to the observed disconnect between successful operations but a lack of real improvements in public sector performance in a share of countries. A political economy perspective should be used to determine how far and how the boundaries of real improvements can best be stretched. (v) Fifth, thinking about political incentives to implement reforms and strategizing with stakeholders – in government, as well as among civil society and the private sector – how to strengthen these incentives should become more explicitly integrated into public sector reform support. The recent emphasis on better ‗(local) fit‘ of reforms (see World Bank 2012) meshes well with such an approach; while a good fit on its own may still be insufficient to actually enable reforms that go against stakeholder incentives. Governments can be supported better to advertise their incremental gains in public sector performance, and national level politicians may like to make sure that money is better spent at local levels to increase voter satisfaction with government performance. At the same time, opportunities for civil society monitoring are growing in many countries. All of these offer opportunities to strengthen political economy reform facilitators. . (vi) Sixth, while a majority of World Bank public sector operations are in countries with limited or no dependency on aid, aid dependent countries account for a significant share of those contexts in which changing public sectors is particularly challenging. As Bräutigam and Knack (2004) have shown, aid dependency can contribute to undermining rather than strengthening the quality of public sectors. Developing approaches that are feasible and effective in these countries typically marked by clientelism, concern for political survival and limited demand for reforms is particularly important. Furthermore, there is a need to extend public sector reform engagement into countries – falling into column 1 in table 5 and often not seeking support – that face weak incentives for even starting public sector reforms, but that urgently require such reforms. Such implications are inevitably challenging to take on board within existing bureaucratic procedures and incentives at the Bank and other aid agencies. However, early efforts to include a more explicit understanding of political economy factors by some teams and the endorsement of such efforts by the Bank‘s new Public Sector Management Approach indicate that some evolution in such directions is well possible. A reasonably well functioning public sector matters for service delivery, economic competitiveness as well as environmental sustainability in LICs and MICs and the combined reviews presented in this paper emphasize that political economy factors can be important inhibitors of moving towards substantially better public sector performance. 42 References Adei, S. and Boachie-Danquah, Y. (2003) ‗The Civil Service Performance Improvement Programme (CSPIP) in Ghana: Lessons of Experience‘ African Journal of Public Administration and Management Vol. 14, No. 1 & 2, p. 10-23. Andrews, M. (2008) ‗PFM in Africa: Where we are, how did we get here, where should we go? – Lessons from recent PEFA data and World Bank PFM Performance Reports‘ Research Study for Brookings Institution and World Bank. Andrews, M. (2009) Isomorphism and the Limits to African Public Financial Management Reform. Faculty Research Working Paper Series, Cambridge: Harvard Kennedy School. Andrews, M. (2011) Does Change Space Influence Public Sector Management Project Success? World Bank, unpublished report. Antwi, K. B. Analoui, F. and Nana-Agyekum, D. (2008) ‗Public Sector Reform in Sub-Saharan Africa: what can be learnt from the civil service performance improvement programme in Ghana?‘ Public Administration and Development Vol. 28, No. 4, October 2008, pp. 253–264. Antwi, K.B. and Analoui, F. (2008a) ‗Challenges in building the capacity of human resource development in decentralized local governments: evidence from Ghana‘ Journal of Management Research News, Vol. 31 No. 7, pp. 504-17. Antwi, K. B. and Analoui, F. (2009) ‗Challenges of successful reform: An international perspective‘ Journal of Management Development Vol. 28, No. 6, 12 June 2009, pp. 489-494. Antwi, K. B. and Analoui, F. (2010) ‗Challenges of Making Donor-Driven Public Sector Reform in Sub-Saharan Africa Sustainable: Some Experiences from Ghana‘ International Journal of Public Administration Vol. 33, No. 12/13, pp. 635-646. Armijo, Biersteker and Lowenthal, A. (1994) ‗The Problem of Simultaneous Transitions‘ Journal of Democracy Vol. 5 No. 4, pp. 161-175. Baimyrzaeva, Mahabat (2010) Shortcomings of institutional reform in public sector governance: The case of Kyrgyzstan, Ph.D. Thesis, University of Southern California. Bangura, Y. and Larvi, G.A. eds. (2006) Public Sector Reform in Developing Countries - Capacity Challenges to Improve Services (Basingstoke: Palgrave Macmillan). Barysch, K. (2006) ‗ Enlargement two years on: Economic success or political failure?‘ Briefing Paper, April 2006, Available at: http://docs.google.com/viewer?a=v&q=cache:lxpUXXi6CLAJ:www.cer.org.uk/pdf/essa y_enlargement_two_%2520barysch.pdf+Eastern+Europe+reform+efforts+post+EU+ac cession&hl=en&gl=us&pid=bl&srcid=ADGEESjoqaIvbHkbJlQt5lji5ZXR7vaBOW9N2 43 Sy4O3- wq5YIUXlOPnt7dxXNqHuoeMjYpKH95AMxteHvMQ6dHFQHJipgjYh3fskHy0N9vt 2pO_Q7GXreqPpDuGLxHZgDeey9XuhDMwda&sig=AHIEtbSx0epi3MqUKTATDM 3fLzmwseEoyA Bekke, H. A. G. M.; Perry, J.L.; and Toonen, T. A. (1996) Civil Service Systems in a Comparative Perspective (Bloomington: Indiana University Press) Bertucci, G. and Jemiai, Y. (2000) ‗Public Sector Reform Revisited in the Context of Globalization‘ Division for Public Economics and Public Administration, Department of Economics and Social Affairs, United Nations, Paper prepared for the Seminar on Public Administration Reform in Asia, Beijing, 15 July 2000. Blau, P.M. (1955) The dynamics of bureaucracy (Chicago, IL: University of Chicago Press). Booth, D. (2011) ‗Governance for Development in Africa: Building on what works‘ Africa, Power, and Politics Policy Brief No. 1, April 2011. Bräutigam, D. A. and Knack, S. (2004) ‗Foreign Aid, Institutions, and Governance in Sub- Saharan Africa‘ in Economic Development and Cultural Change Vol. 52 No. 2, January 2004 (Chicago: University of Chicago Press), pp. 255-286. Bresser-Pereira, L. C. (1999) ‗From Bureaucratic to Managerial Public Administration in Brazil‘ in Bresser Pereira, L. C. and Spink, P. (1999) Reforming the State – Managerial Public Administration in Latin America (Boulder: Lynne Rienner Publishers). Bresser-Pereira, L.C. (2003) ‗The 1995 Public Management Reform in Brazil: Reflections of a Reformer‘ in Schneider, Ben Ross and Heredia, Blanca (2003) Reinventing Leviathan: the politics of administrative reform in developing countries (Miami: North-South Center Press, University of Miami). Brinkerhoff, D.W. (2000) ‗Assessing political will for anti-corruption efforts: an analytic framework‘ Public Administration and Development Vol. 20, No. 3, pp. 239-252. Brunsson, N. (1989) The organization of hypocrisy: Talk, decisions and actions in organizations (Chichester: John Wiley). Campos, J. E. and Pradhan, S. (Eds.) (2007). The many faces of corruption. Tracking vulnerabilities at the sector level. The World Bank, Washington, D.C. Cheung, A.B.L. (2003) ‗Public Sector Reform in Singapore: reinventing government in a global age‘, in Cheung, A.B.L. and Scott, I. eds. (2003) Governance and PSR in Asia (London: Routledge), pp. 117-137. Cheung, A.B.L. (2005) ‗The Politics of Administrative Reforms in Asia: Paradigms and Legacies, Paths and Diversities‘ Governance Vol. 18, Issue 2, April 2005, pp. 257–282. 44 Crozier, M. (1964) The bureaucratic phenomenon (Chicago, IL: University of Chicago Press). Cruz, C. and Keefer, P. (2010) ‗Programmatic Political Parties and Public Sector Reform‘, APSA Annual Meeting Paper, September 2010. De Renzio, P. (2011) ‗Buying Better Governance: The Political Economy of Budget Reforms in Aid-Dependent Countries„, GEG Working Paper 2011/65 (September 2011).Devarajan, S., Khemani, S. and Walton, M. (2011) ‗Civil Society, Public Action and Accountability in Africa‘ Policy Research Working Paper 5733, July 2011 (Washington DC: The World Bank). Diamond, Jack (2011) Political Economy of PFM reforms. Mimeo. Dimitrova, A. L. (2002) ‗Enlargement, Institution-Building, and the EU‘s administrative capacity requirement‘ West European Politics, Vol. 25, No. 4, pp. 171-190. Dimitrova, A. L. (2010) ‗The new member states of the EU in the Aftermath of Enlargement: Do new European Rules remain empty shells?‘ Journal of European Public Policy, Vol. 17, No. 1, pp. 137-148. Echebarría, K. and Cortazar, J. C. (2007) ‗Public Administration and Public Employment Reform in Latin America‘, in Lora, E. eds. (2007) The State of State Reform in Latin America (Washington DC: IDB). Eisenstadt, S. (1973) Traditional Patrimonialism and Modern Neopatrimonialism (Beverly Hills: Sage). El-Said, H. (2000) Jordan: The political economy of public sector growth and reform in a rent- seeking economy, 1973-1998 (Amman, Jordan; Paris: CERMOC). Fatile, J.O. and Adejuwon, K.D. (2010) ‗Public Sector Reform in Africa: Issues, Lessons and Future Directions‘ Journal of Sustainable Development in Africa (Volume 12, No.8, 2010) Fukuyama, F. (2004) State-Building: Governance and World Order in the 21st Century (Ithaka, NY: Cornell University Press). Fukuyama, F. (2011) The Origins of Political Orders (New York: Farrar, Straus and Giroux). Gaetani, F. (2003) ‗Public Management Policy Change in Brazil: 1995-1998‘ International Public Management Journal No 6, Vol. 3, pp. 327-41. Garcia-Zamor, J-C. and Khator, R. (Eds.) (1994) Public Administration in the Global Village (Westport, CT: Praeger). 45 Geddes (1994) Politician‘s Dilemma – Building State Capacity in Latin America (Berkeley: University of California Press). Geddes, B. (1991) ‗A Game Theoretic Model of Reform in Latin American Democracies‘ The American Political Science Review Vol. 85, No. 2 (Jun., 1991), pp. 371-392. Gehlbach, S. and Keefer, P. (2011) ‗Private Investment and the Institutionalization of Collective Action in Autocracies: Ruling Parties and Legislatures‘ (Last revised on August 1, 2011). Available at SSRN: http://ssrn.com/abstract=1663057. Giambiagi, F. and Ronci, M. (2004) ‗Fiscal Policy and Debt Sustainability: Cardoso‘s Brazil, 1995-2002‘ IMF Working Paper WP/04/156.Grindle, M. ed. (1997) Getting Good Government (Cambridge: Harvard University Press). Grindle, M. S. (2004) Despite the Odds: Contentious Politics and Education Reform (Princeton, NJ: Princeton University Press). Haggard and Kaufmann (1992) The politics of economic adjustment (Princeton, NJ: Princeton University Press). Haggard (1997) ‗Reform of the State in Latin America‘ in Burki, S. J. and Perry, G. (1997) Development in Latin America and the Caribbean (Washington DC: World Bank). Hannan, M.T. and Freeman, J. (1984) ‗Structural inertia and organizational change‘ American Sociological Review Vol. 49, pp. 149–164. Harun (2007) ‗Obstacles to Public Sector Accounting Reforms in Indonesia‘ Bulletin of Indonesian Economic Studies Vol. 43 No. 3, pp. 365-376. Haruna, P. H. (2003) ‗Reforming Ghana's Public Service: Issues and Experiences in Comparative Perspective‘ Public Administration Review Volume 63, Issue 3, May 2003, pp. 343–354. Hedger, E. and Kizilbash Agha, Z. (2007) ‗Reforming public financial management when the politics aren‘t right: A proposal‘, ODI Opinion 89. London: ODI. Hedger, E. and de Renzio, P. (2010) ‗What do Public Financial Management assessments tell us about PFM reforms?‘ ODI Background Note, July 2010. Heredia, B. and Schneider, B.R. (1998) ‗The Political Economy of Administrative Reform: Building State Capacity in Developing Countries‘ March 1998. Available at http:// lasa.international.pitt.edu/LASA98/Heredia-Schneider.pdf. Hesse, J. J. (1997) ‗Rebuilding the State: Public Sector Reform in Central and Eastern Europe‘ in Lane, J. E. (1997) Public Sector Reform – Rationale, Trends and Problems (London: Sage). 46 Hood, C. (1996) ‗Exploring variations in public management reform of the 1980s‘, in Bekke, H. A. G. M.; Perry, J.L.; and Toonen, T. A. (1996) Civil Service Systems in a Comparative Perspective (Bloomington: Indiana University Press). Hutchful, E. (2002) Ghana‘s Adjustment Experience – the Paradox of Reform (UNRISD). IEG (2004), ‗An Evaluation of World Bank Assistance to Transition Economies‘ Report No. 29761 (World Bank: Washington DC). IEG (2008) Public Sector Reform: What works and Why? An IEG Evaluation of World Bank Support (Washington, DC: World Bank). IEG (2011) IEG Annual Report 2011: Results and Performance of the World Bank Group (Washington, DC: World Bank). Ingraham, P. (1996) ‗The Reform Agenda for National Civil Service Systems: External Stress and Internal Strains‘ in eds. Bekke, H., J. Perry and T. Toonen. Inter-American Development Bank (2006) The Politics of Policies – Economic and Social Progress in Latin America, 2006 Report (David Rockefeller Center for Latin American Studies: Harvard University). Issa, F. H. H. (2010) ‗Implementing Change by Continuous Learning: The Public Service Reform Programs in Tanzania‘ International Journal of Public Administration Vol. 33, pp. 474–486. Jacobs, C. (1998) 'Institutional Strengthening and Technical Co-operation: Developing a Best Practice Model' Journal of International Development Vol. 10, No. 3, pp. 397—406. Jacobsen, D. I. (2005) ‗Sand in the machinery? Comparing bureaucrats‘ and politicians‘ attitudes toward Public Sector Reform‘ European Journal of Political Research Vol. 44, pp. 767–799. Keefer, P. (2011) Collective action, political parties and pro-development public policy. World Bank Policy Research Working Paper no. 5676. Khemani, S. (2007) ‗Can information campaigns political obstacles to serving the poor?‘ in Devajaran, S. and Widlund, I. (2007) The politics of Service Delivery in Democracies. Better access for the poor, Expert Group on Development Issues (EGDI) (Stockholm: Swedish Ministry of Foreign Affairs). Kingdon, J. W. (1995) Agendas, Alternatives and Public Choices (New York: Harper Collins College Publishers). Kiragu, K. (2002) ‗Improving Service Delivery through Public Service Reform: Lessons of Experience from Selected Sub-Saharan African Countries‘ Discussion Paper presented 47 at the Second Meeting of the DAC Network on Good Governance and Capacity Development, Paris: OECD Headquarters, 14-15 February 2002. Kiragu, K. and Mukandala, R. (2003) ‗Public Sector Pay Reform – Tactics Sequencing and Politics in Developing Countries: Lessons from Sub-Saharan Africa‘, Draft Report. Kitschelt, H. (2011) Clientelistic Linkage Strategies. A Descriptive Exploration. Paper prepared for the Workshop on Democratic Accountability Strategies. Duke University. May 18- 19, 2011. Kitschelt, H. Wilkinson, S. (2007) Patrons, Clients, and Policies: Patterns of Democratic Accountability and Political Competition. Cambridge: Cambridge University Press. Kulibaba, N. and Rielly, C. (1994) ‗Improving the Effectiveness of Agricultural Policy Reform‘ In Africa: A Synthesis of Lessons Learned. Collaborative Research Report No. 360. February 1994. US Agency for International Development, Agricultural Policy Analysis Project, Phase 11: Washington. DC. Leftwich, A. (2000) States of Development: On the Primacy of Politics in Development (Cambridge: Polity). Leftwich, A. (2011) ‗Politics, Leadership and Coalitions in Development‘ Background Papers, Research and Policy Workshop, Frankfurt, Germany 10-11 March 2011. March, J.G. and Olsen, J.P. (1995) Democratic governance (New York: Free Press). Merton, R. (1940) ‗Bureaucratic structure and personality‘ Social Forces Vol. 17, pp. 560–568. Meyer-Sahling, J- H. (2009) ‗Sustainability of Civil Service Reforms in Central and Eastern Europe Five Years After EU Accession‘ SIGMA (2009) No. 1, April 2009. Meyer-Sahling, J-H. (2011) ‗The Durability of EU Civil Service Policy in Central and Eastern Europe after Accession‘ Governance, Vol. 24, Issue 2, pp. 231-260. Moon, M-J. and Ingraham, P. (1998) Shaping Administrative Reform and Governance: An Examination of the Political Nexus Triads in Three Asian Countries Governance Volume 11, Issue 1, pages 77–100, January 1998. Morgan, P., Baser, H., and Morin, D. (2010) Developing Capacity 2010for Managing Public Service Reform: The Tanzanian Experience 2000-2008 Public Administration and Development Vol. 30, p. 27–37 Musoni, P. (2003) ‗Reconstructing Governance and Public Administration Institutions for Effective, Conflict-Sensitive Rule of Law‘ United National Ad Hoc Expert Group Meeting, Yaoundé, Cameroon, July 2003. 48 Nelson, J. (1994) ‗How market reforms and democratic consolidation affect each other‘ in Nelson, J. Intricate Links (Washington DC: Overseas Development Council). Nelson, J. M. (2007) ‗Democratic Politics and Pro-Poor Social Services: Unpacking the Concept of Reform‘ in Devajaran, S. and Widlund, I. (2007) The politics of Service Delivery in Democracies. Better access for the poor, Expert Group on Development Issues (EGDI) (Stockholm: Swedish Ministry of Foreign Affairs). North, D., Wallis, J. J., and Weingast, B. (2009) Violence and Social Order: A Conceptual Framework for interpreting recorded human history (Cambridge: Cambridge University Press). North, D., Wallis, J.J., Webb, S., and Weingast, B. (2010) ‗Limited Access Orders: An Introduction to the Conceptual Framework‘. Stanford University http://politicalscience.stanford.edu/faculty/weingast/LAO_CUP_2_Intro2100614.pdf O‘Neil, T. (2007) ‗Neopatrimonialism and public sector performance and reform‘ Good Governance, Aid Modalities and Poverty Reduction: Linkages to the Millennium Development Goals and Implications for Irish Aid, Research project (RP-05-GG) of the Advisory Board for Irish Aid, Background Note 1, September 2007. Painter, M. (2003) Public administration reform in Vietnam: problems and prospects Public Administration and Development Vol. 23, Issue 3, August 2003, pp. 259–271. Painter, M. (2004) The Politics of Administrative Reform in East and Southeast Asia: From Gridlock to Continuous Self-Improvement? Governance Volume 17, Issue 3, July 2004, pp. 361–386. Panizza, F. (2004) ‗A Reform Without Losers: The Symbolic Economy of Civil Service Reform in Uruguay, 1995–96‘ Latin American Politics & Society, Vol. 46, No. 3, Fall 2004, pp. 1-28. Panizza, F. and Philip, G. (2005) ‗Second Generation Reform in Latin America: Reforming the Public Sector in Uruguay and Mexico‘ Journal of Latin American Studies, Vol. 37, No. 4, pp. 667-691. Pastor, M. and Wise, C. (1999) ‗The Politics of Second-Generation Reform‟ Journal of Democracy Vol. 10, No. 3, July 1999, pp. 34-48. Peters, G. and D. Savoie (1995) Governance in a Changing Environment (Buffalo: McGill- Queen‘s University Press). Polidano C., Hulme D. (1999) ‗Public management reform in developing countries: issues and outcomes‘. Public Management Vol. 1, No. 1, pp. 121–132. 49 Polidano, C. (2001) ‗Why civil service reforms fail‘ Public management review Vol. 3, No. 3, pp. 345-361. Pollitt, C. and Bouckaert, G. (2000). Public management reform: A comparative analysis (Oxford: Oxford University Press). Posner, D. (2005) Institutions and Ethnic Politics in Africa (NY: Cambridge Univ. Press). Pritchett, L. and de Weijer, F. (2010) ‗Fragile States: Stuck in a Capability Trap?‘ A Background Paper prepared for the World Development Report 2011, 5 November 2010. Quality Assurance Group (22 April 2009) Annual Report on Portfolio Performance – Fiscal Year 2008. Sarker, A. E. (2006) ‗New Public Management in Developing Countries – An analysis of success and failure with particular reference to Singapore and Bangladesh‘ International Journal of Public Sector Management, Vol. 19, No. 2, pp. 180-203. Schacter, M. (1995) 'Recent Experience with Institutional Development: Lending in the Western Africa Department' in Petter Langseth, Sandite Nogxina, Daan Prinsloo and Roger Sullivan (eds) Civil Service Reform in Anglophone Africa. Report of workshop proceedings. Pretoria: Economic Development Institute, Overseas Development Administration, and Government of South Africa. Schacter M. (2000) ‗Public Sector Reform in Developing Countries Issues, Lessons and Future Directions‘ Policy Branch (Canadian International Development Agency: Ottawa). Schacter, M. (2001) ‗The Heart of the Matter: Donors, Development Assistance, and PSR‘, Policy Brief, No. 10 February 2001 (Institute on Governance: Ottawa). Schick, A. (1996) The spirit of reform: Managing the New Zealand state sector in a time of change (Wellington: State Services Commission and the Treasury). Schneider, Ben Ross and Heredia, Blanca (2003) Reinventing Leviathan: the politics of administrative reform in developing countries (North-South Center Press). Schwarz, H. (1994) ‗Public choice theory and public choices: Bureaucrats and state reorganization in Australia, Denmark, New Zealand and Sweden in the 1980s‘ Administration and Society Vol. 26, pp. 48–77. Shin, R. and Ha, Y-S. (1999) ‗Political Economy of Policy Reform in Korea: Review and Analysis‘ Review of Policy Research Vol. 16, Issue 2, June 1999, pp. 65–98. Sozen, S. and Shaw, I. (2002) ‗The International Applicability of new public management: lessons from Turkey‘ International Journal of Public Sector Management Vol. 15. I, pp. Issue 6, pp.475 – 486. 50 Stevens, M. and Teggemann, S. (2004) ‗Comparative Experience with Public Service Reform in Ghana, Tanzania, and Zambia‘ in Levy, B. and Kpundeh, S. (2004) Building State Capacity in Africa – New Approaches, Emerging Lessons (WBI Development Series). Suleiman, E.N. and Waterbury (1990) The Political Economy of Public Sector Reform and Privatization (Boulder: Westview Press). Tendler, J. (1997) Good Government in the Tropics (Baltimore: The Johns Hopkins University Press). Therkildsen, O. (2000) ‗Public sector reform in a poor, aid dependent country, Tanzania‘ Public Administration and Development, Vol. 20, pp. 61.71. Therkildsen, O. (2001) ‗Efficiency, Accountability and Implementation – Public Sector Reform in East and Southern Africa‘ Democracy, Governance and Human Rights Programme Paper No. 3, February 2001, UN Research Institute for Social Development. Thomson, A. (2010) An Introduction to African Politics (NY: Routledge). Tomassi, M. (2003) ‗Crisis, Political Institutions, and Policy Reform – The good, the bad, and the ugly‘ Leitner Program Working Paper 2003, Vol. 1. Toonen, T. A. J. (2001) ‗The comparative dimension of administrative reform: creating open villages and redesigning the politics of administration‘ in Peters, G. and Pierre, J. (2001) Politicians, Bureaucrats and Administrative Reform (London: Routledge). Trauner, F. (2009) ‗Post-accession compliance with EU law in Bulgaria and Romania: a comparative Perspective‘ in Schimmelfennig, F. and Trauner, F. (eds): Post-accession compliance in the EU‘s new member states, European Integration online Papers (EIoP), Special Issue 2, Vol. 13, Art. 21, http://eiop.or.at/eiop/texte/2009-021a.htm. Van de Walle, N. (2003). "Presidentialism and Clientelism in Africa's Emerging Party Systems." Journal of Modern African Studies Vol. 41, No. 2, pp. 297-321. World Bank (1996) Bangladesh: Government that Works – Reforming the Public Sector , University Press, Dhaka. World Bank (1997) World Development Report 1997: The State in a Changing World (New York: Oxford University Press). World Bank (1999) ‗Rethinking Civil Service Reform‘ PREM Notes (Public Sector), No. 31 (October). World Bank (2000) ‗From Patronage to Professional State: Bolivia Institutional and Governance Review‘ Report Number 20115 – BO. 51 World Bank (2000b) Reforming Public Institutions and Strengthening Governance. A World Bank strategy. Washington DC. World Bank (2009) Governance and Anticorruption in Lending Operations: A Benchmarking and Learning Review. Washington DC. Available at: http://siteresources.worldbank.org/EXTLAWJUSTINST/Resources/GAC_review_2009. pdf. World Bank (2010) Issues Paper 1.A ‗What is the public sector?‘ Draft: Oct 5, 2010. World Bank (2012) Public Sector Management Approach. Washington DC: World Bank. 52