New insights from existing evidence to Domestic inform decisions, address knowledge Revenue gaps, and enhance operational learning Mobilization February 2023 click on individual headings for immediate access to the related section Domestic revenue mobilization (DRM) has become an increasingly important part of international and country-level policy agendas, yet promoting and sustaining reforms in this area has been challenging. In the operations reviewed, political economy constraints were identified and analyzed at the time of approval and, for the most part, reasonable mitigation measures were identified ex ante; however, tax reforms were still derailed. Prior actions in development policy operations often benefited from being paired with investment projects or technical support from development partners. Although coordinating with development partners was often discussed in project documents, in just a few cases was there extensive coordination, with mixed results. Results frameworks to track progress on domestic revenue mobilization were often not well articulated, sometimes missing baseline or targets, being over- ly optimistic, or lacking clear results chains. This Evaluation Insight Note draws on a detailed review of six Independent Evaluation Group evaluations of World Bank interventions supporting domestic revenue mobilization. Drawing from the Independent Evaluation Group’s rich knowledge repository, Evaluation Insight Notes respond to the need for more rapid and focused evaluative evidence. These notes systematically analyze data from a range of evaluations, validations, and other studies to generate insights in a timely manner around important strategic and operational issues. What Are the Lessons from the World Bank’s Support for Domestic Revenue Mobilization? Domestic revenue mobilization (DRM) has become an increasingly important part of international and country-level policy agendas. Since the 2015 World Conference on Financing for Development in Addis Ababa, Ethiopia, DRM has risen in importance in the international policy agenda—forming part of 2 of the 17 Sustainable Develop- ment Goals—and has figured prominently in successive International Development Association Replenishments. In the years leading up to the COVID-19 pandemic, high fiscal deficits and already high and rising debt levels made enhancing DRM a priority for developing economies. This importance has increased further with the drop in tax revenues that occurred at the onset of the pandemic. This Evaluation Insight Note gathers insights from World Bank projects and operations supporting DRM through a detailed review of six Independent Evaluation Group (IEG) Project Performance Assessment Reports (PPARs).1 It synthesizes key issues that af- fected each project’s ability to sustain results across five countries (Croatia, Guatemala, Liberia, Pakistan, and Panama) and one state (Rio de Janeiro in Brazil). Interventions covered include four development policy operations (DPOs), one technical assistance loan, and one specific investment loan (now called investment project financing). Three of the DPOs were programmatic series consisting either of two (Guatemala, Pakistan) or three (Panama) operations, and one was a stand-alone operation (Rio de Janeiro). The interventions evaluated were approved between June 2007 and May 2014 and closed between January 2014 and June 2017. Most of the operations sought to address issues beyond DRM. In these cases, only DRM-relevant aspects were assessed for this note. What Are the Results from the Six Project Performance Assessment Reports? The operations reviewed had disappointing DRM-related results at the time of closing (table 1), and these were further eroded by policy reversals over time. While DRM- 1  Project Performance Assessment Reports (PPARs) are undertaken several years after a project or operation closes to assess, among other things, whether or not results were sustained. 2 related objectives in the operations were relevant to country needs, design was weak, with two-thirds rated modest for the quality of design. Achievement of DRM-related objectives was rated modest in all but one case. Only the operation in Pakistan was rated as having substantial achievement of DRM-related objectives, and even in this case, there were subsequent, substantial policy reversals. For the 20 DRM-related results indicators in the six interventions, 11 (or 55 percent) were fully achieved at program closing (figure 1).2 However, of these, only 5 had been sustained when the PPAR was prepared. All PPARs identified room for improvement in the World Bank’s efforts to strengthen DRM, mainly in relation to design of operations but also in relation to implementation.3 World Bank projects faced challenges in political economy constraints, matching the right intervention to the problem, properly leveraging donor support, and preparing a results framework. As such, this Evaluation Insight Note catalogues issues that may be informative for other DRM operations. Table 1. Overview of Relevance, Achievement, and Bank Performance Ratings Achievement of Relevance of Relevance of DRM Objectives Country Objectives Design (PPAR) Bank Performance Guatemala Substantial Modest Modest Moderately unsatisfactory Pakistan High Modest Substantial Moderately satisfactory Panama High High Modest Moderately satisfactory Rio State, Brazil High Modest Modest Moderately unsatisfactory Croatia Substantial Modest Modest Moderately unsatisfactory Liberia High n.a. Modest Moderately unsatisfactory Sources: World Bank 2019, 2020a, 2020b, 2020c, 2021, 2022. Note: Budget support operations (development policy operations) are shaded purple. DRM = domestic revenue mobilization; n.a. = not applicable; PPAR = Project Performance Assessment Review. 2  These results are roughly similar to the broader achievement of targets for results indicators related to domestic revenue mobilization (DRM) in development policy operations (DPOs) across the World Bank for fiscal years (FY)16–19, where 69 percent were fully or mostly achieved, 5 percent partially achieved, and 26 percent not achieved. 3  This is comparable to ratings for World Bank projects at closing instead of several years later for a PPAR in this Evaluation Insight Note. For FY16–19, Bank performance was rated satisfactory for two-fifths of all DRM-related investment projects and one-fifth of DPOs with at least one DRM-related prior action; a similar share (three-quarters) of both investment projects and DPOs were rated moderately satisfactory or higher. 3  chievement and Sustainability of Indicators Related to Domestic Revenue Figure 1. A Mobilization Results 12 5 of 11 not sustained 10 at the time of PPAR 8 6 4 2 0 Achieved Partially achieved Not achieved Source: Independent Evaluation Group. Note: Figure compares data at closing and several years later, when the PPAR was prepared. DRM = domestic revenue mobi- lization. PPAR = Project Performance Assessment Review. What Are the Four Main Insights from This Synthesis? In the operations reviewed, political economy constraints were usually identified and analyzed at the time of approval and, for the most part, reasonable mitigation measures were identified ex ante; however, tax reforms were still derailed. The most frequent challenge identified in project documents to the achievement of development outcomes was political economy constraints, which regularly derailed tax reforms either early on or after several years. Although these risks were regularly iden- tified and analyzed in project documents, mitigation measures proved insufficient to 4 overcome political resistance. While the World Bank judged the potential reform payoff worth the risk, it raises the question of why the World Bank would have believed that a particular instance was different. Can World Bank operations be expected to have suf- ficient leverage if government support for reforms is weak, vested interests are strong, or underlying conditions change? Are there other commitment mechanisms that could have supported sustainable reforms in light of vested interests? In the most successful of the operations assessed, in Pakistan, after some early suc- cess, tax exemptions substantially larger than at the baseline were reintroduced, while implementation of several other tax reforms stalled in the face of opposition from in- terest groups (World Bank 2020c). In Guatemala and Panama, opposition also blocked the implementation of tax reform (World Bank 2019, 2022). Prior actions and mitigation efforts were insufficient to meaningfully move along the results chain toward greater DRM. Underlying conditions eventually changed in Pakistan and Panama, where the initial impetus for reform was rooted in crisis but was not sustained after pressures abated (World Bank 2020c). In Croatia, strong pressures to urgently prepare a project supporting the country’s accession to the European Union were supportive in getting the project started (World Bank 2020a); however, concerns about readiness were raised before approval, and the rapid preparation undermined project implementation. The cases of Croatia, Pakistan, and Panama raise the question of whether crisis or urgent responses are an appropriate entry point for tax reform or if they increase the likelihood of policy reversal. Conversely, in Rio State, vested interests were more mundane: the development policy financing did not overcome insufficient incentives to improve per- formance (World Bank 2021). In Guatemala, political economy challenges were cited as the main reason for under- performance (World Bank 2019). The World Bank recognized substantial risks with respect to the new government’s ability to overcome well-known and entrenched opposition from vested interests. Opposition to tax reform came from the private sec- tor and manifested itself in delays in parliamentary approval of tax reform legislation and a failure to implement measures that had been legislated or decreed. Approved legislation was immediately challenged in the constitutional court. The World Bank had hoped that an extensive consultation process with support from experts across the political spectrum would overcome strong opposition.4 But the opposition’s ability 4  The project document discussed an extensive consultation process on policy reforms, built on efforts started over a decade earlier. The incoming administration had begun an intensive consensus-building process on reforms even before it took office, which enabled them to pass a tax reform package within the first two months of taking office. The reform package drew heavily on a proposal developed in 2008 by a group of 40 economists from across the political spectrum. Aware of the potential for legal challenges to any significant reform, the authorities drew on internal and independent expertise on constitutional tax law. 5 to effectively block tax reform once again proved too strong. Although the World Bank could have encouraged the government to expand outreach to stakeholders, it is not clear how much more this would have achieved, given what had already taken place. Additional technical assistance was unlikely to overcome the constraints identified. Underlying conditions eventually changed in Pakistan and Panama, where the initial impetus for reform was rooted in crisis but was not sustained after pressures abated. In Pakistan, critical tax reforms stalled because of opposition from vested interests (World Bank 2020c). The first development policy credit (DPC) was approved in May 2014 and the second in June 2015, but revenue improvements were not sustained beyond 2017, when an internal political crisis resulted in the disqualification of the prime minister by a supreme court decision and a major weakening of government. This undermined the government’s capacity to deliver on its reform commitments. These events were exacerbated by the end of the arrangement supported by the International Monetary Fund (IMF) under the Extended Fund Facility (EFF), which had served as an effective disciplining instrument. After initial declines starting in 2014, tax exemptions were reintroduced in advance of the 2018 elections, increasing them from 1.6 percent to 2.5 percent of gross domestic product (GDP) between 2017 and 2019. As a result, tax collection deteriorated from 12.9 percent to 11.6 percent of GDP. Stalled reform also led to incomplete implementation of general sales tax reform, the preservation of tax exemptions for the agriculture sector, and ad hoc statutory regulatory orders. The DPC was the World Bank’s first policy-based loan to Pakistan in more than a decade. Several earlier attempts by the World Bank to support taxation-related policy reforms through development policy financing had failed, undermined by a lack of political will, strong vested interests, and gaps in the government’s institutional capacity (World Bank 2020c). Following previous IEG guidance, the DPC program was split into two parallel programmatic series to help mitigate political economy 6 risks and implementation constraints.5 In addition, the program document included a strong analysis of political economy risks and proposed broadly appropriate mitigation measures. The weakest point of the mitigation strategy related to communication: although the program document suggested undertaking “systematic work with media and other stakeholders to track and inform public opinion on the benefits of reform” (World Bank 2020c, 40), the PPAR found no evidence that such support was provided in a systematic manner. Despite the high risks, the World Bank saw the benefits of this operation—helping the government catalyze and consolidate much-needed reforms during the reform window opened up by the political transition and a stable government—as outweighing the costs. In Panama, a crisis launched interest in reform. In 2010, the Organisation for Economic Co-operation and Development placed the country on a list of countries with inad- equate tax transparency, which was sanctioned by the World Bank, and the Panama Papers were published.6 International perceptions of inadequate international tax transparency threatened to undermine the country’s standing as a financial and busi- ness center and a destination for foreign investment (World Bank 2022). As a result, the government prioritized addressing this status. The World Bank used the opportunity to push domestic tax reform, including a broadening of the domestic revenue base (World Bank 2022). But elimination of tax exemptions proved more difficult than envisaged in Panama (World Bank 2022). There were strong political headwinds to domestic tax reform out- side the executive branch of government; the private sector did not see the rationale for increasing domestic tax revenue in light of large nontax revenues, including from the Panama Canal, and low fiscal deficits and public debt. In addition, the supreme court ruled unconstitutional the operation-supported reorganization of the Directorate of Government Revenues into an autonomous agency, which would have strength- ened the agency’s authority and capacity to increase tax revenue. Risks associated with opposition within society were to be mitigated through consultation with civil society. However, opposition to reform came from private sector interests and legal groups that successfully challenged the constitutionality of the reform. 5  The program document referred to recommendations from a 2005 PPAR’s review of four earlier structural adjustment loans to Pakistan, which advised that “complex or politically difficult sector reforms are best supported through dedicated sector operations. A multisector operation can play only a secondary or facilitating role when dealing with . . . deep-rooted reluctance to reform (as in power)” (World Bank 2014, vii). 6  The Panama Papers provided considerable negative publicity in 2016 when the International Consortium of Investigative Journalists and 100 international media partners began publishing massive documentary evidence of foreign taxpayers’ presence in Panama, some of which was for tax evasion and potentially illicit activities. 7 In Rio de Janeiro, World Bank optimism for DRM reform was grounded in what staff considered a genuine desire on the part of the state governor to improve Rio de Janeiro’s DRM performance (World Bank 2021). Evidence of this was seen in the appointment of a highly qualified professional to lead the tax administration agency. At the same time, there were indications of the lack of political will to put in place a working tax audit program with international practices. Moreover, auditors and in- spectors of the tax administration service had little to no incentive to improve perfor- mance because most were already at the top of their pay scale, which undermined efforts to implement a results-based management system to improve revenues at the state Secretariat of Finance. In Croatia, the DRM project was incentivized by Croatia’s bid for European Union (EU) membership (World Bank 2020a). It was prepared expeditiously in response to an ur- gent government request as part of their negotiations for EU accession and was intend- ed to align its operations with EU practices and build capacity for implementing the EU acquis on taxation. However, expedited preparation proved to be the project’s downfall: it allowed little time for stakeholder consultation and project preparation. At project approval, the Croatia Tax Administration (CTA) had only endorsed a high-level descrip- tion of the strategic modernization framework. Concerns about lack of readiness were raised at the project’s internal review meeting, but the decision was made to proceed with the project on the expedited schedule. This allowed little time for stakeholder consultation, and several critical preparatory elements were pushed to the implemen- tation stage as effectiveness conditions and dated covenants, including adoption of the strategic plan for CTA modernization and new legislation for civil service salaries. Early stages of the project implementation focused on civil works for the new headquarters complex and on information and communication technology procurement, but lack of agreement on the location of the new headquarters led to a three-year delay and the eventual dropping of the civil works component (World Bank 2020a). Two project restructurings were not enough to overcome weak design and preparation at entry. Restructurings recentered on capacity building but did not anticipate the extent of the CTA’s weak institutional capacity, which resulted in protracted delays in procurement and in completing tasks before project closing. Given the momentum for EU accession, the project could have better secured the government’s commitment to tax administration reforms up front. Moreover, frequent changes in government also led the management of the tax administration to change frequently (with the directors-general often lacking a background in tax administration). This contributed to a lack of steady ownership and momentum for the institutional reform process. 8 Prior actions in DPOs often benefited from being paired with investment projects or technical support from development partners. Using the right interventions for the circumstances matters, and the reform actions of DPOs often benefit from being paired with investment projects or technical support. The Rio State DPO was not sufficient for reform success and was not well suited to address weaknesses in tax administration, where pairing with an investment project may have helped. Pakistan had a combination of interventions: a longer-term budget support series paired with adaptable, comprehensive technical assistance. However, in Croatia, despite concerns about project readiness, an investment project would have benefited from including policy reforms. The Rio de Janeiro operation was one of several stand-alone operations in the state that touched on tax administration and was the only one that made a serious effort to influ- ence the quality of tax administration (World Bank 2021). Two other operations (the Rio State Development Policy Loan I and the Rio Municipality Development Policy Loan) also touched on tax administration, but the issues they addressed were not a coherent and mutually reinforcing set of tax reforms. Taken together, these budget support operations did not constitute a coherent and consistent effort to tackle problems in tax administra- tion. The stand-alone operation as designed was ill-suited to reform tax administration in any significant way. Although the World Bank had good knowledge of the areas the loan covered, the operation did not have a clear theory of change or a realistic plan for how to advance and preserve the reforms, given a challenging political and institutional envi- ronment. Moreover, tax administration office staff lacked information technology capacity and were not able to carry out the changes in hardware and software required to achieve program objectives. There was inadequate complementary technical assistance to solve problems that required deep and sustained change over several years. An investment project loan anchored on a longer-term institutional partnership with state implementing agencies might have been a better tool for supporting the underlying reforms. In Guatemala, the program document identified the risk that the effectiveness of the tax reforms depended on the ability of the tax administration to apply and enforce the new provisions (World Bank 2012b). It identified donor support aimed at strengthening tax administration as mitigating measures. Despite this, the PPAR did not find evidence of targeted technical assistance, and political economy issues derailed the reforms (World Bank 2019). 9 Although the World Bank had good knowledge of the areas the loan covered, the operation did not have a clear theory of change or a realistic plan for how to advance and preserve the reforms, given a challenging political and institutional environment. In Pakistan, the programmatic budget support series had two operations, as part of a coordinated international effort with the IMF, the Department for International Development (DFID), the United States Agency for International Development, and the Asian Development Bank. The program envisioned a gradual intensification of reform and benefited from a longer-term strategy that provided for several interrelated DPCs, including a follow-up operation (World Bank 2020c). It was also accompanied by a long-term, adaptable technical assistance package provided by DFID. Available technical assistance was comprehensive, covering all policy areas, and delivered through multiple channels, including ongoing World Bank investment projects, trust funds executed by the government of Pakistan, and support from other development partners, including the IMF, DFID, and the United States Agency for International Development. DFID provided £300 million in budget support and £40 million in technical assistance, with grants for budget support linked to overall progress on the IMF- and World Bank–supported programs, including achieving specific targets in tax collection. In addition, the World Bank and DFID established the Multi-Donor Trust Fund for Accelerating Growth and Reforms, funded by DFID, to provide technical assistance for implementing the reforms supported under the World Bank’s DPCs. The trust fund financed technical assistance in several areas of focus under the Fiscally Sustainable and Inclusive Growth series, which was very responsive to the needs of both the implementing agencies and the World Bank team. This offered the project team additional opportunities to address emerging bottlenecks during implementation and monitoring. In Croatia, reforms to foster the administrative and financial autonomy of the CTA, or tax policy reforms to broaden the tax base, could have enhanced the effectiveness of tax administration (World Bank 2020a). Instead, much of the project’s early efforts were focused on civil works and information technology infrastructure at the cost of institu- tional reform, which slowed modernization efforts. 10 Although coordinating with development partners was often discussed in project documents, extensive coordination occurred in only a few cases, with mixed results. Interventions involved a range of engagement with development partners, particularly with the IMF on policy reform dialogue. While coordination did support results in some cases (for example, Pakistan), it showed little impact in others (for example, Croatia and Liberia), suggesting that partners coalescing in relation to an agenda may not on its own deliver results. In Pakistan, the World Bank’s development policy financing prior actions were more effective when they were coordinated with an IMF program and with technical assistance from DFID (World Bank 2020c). In Guatemala, Panama, and Rio State, no IMF program was in place to support policy reform (World Bank 2019, 2021, 2022). While there were parallel Inter-American Development Bank (IADB) policy reform and investment operations in Guatemala and Panama, they did not support the same reforms or capacity constraints in tax administration. In Croatia, the World Bank coordi- nated with the IMF on analytical and diagnostic work (World Bank 2020a). In Liberia, the World Bank brought in multiple donors into a multidonor trust fund, which increased synergies and reduced the burden on a fragile government with limited capacity (World Bank 2020b). However, while the World Bank may have increased synergies, weak project management and supervision undermined results. In Pakistan, close coordination and alignment with the IMF’s arrangement under the EFF provided considerable synergies, especially in tax policy and administration and in trade policy and financial sector reform (World Bank 2020c). The EFF was implement- ed in 2014–16, in parallel with the Fiscally Sustainable and Inclusive Growth DPO, and provided additional leverage through demanding structural benchmarks, indicative targets and quarterly monitoring. After the end of the EFF arrangement in 2016, reform momentum greatly diminished, and the World Bank on its own did not have sufficient leverage in the most sensitive policy areas (taxation, trade, and privatization). In Guatemala, no IMF-supported program was in place during this period, although the World Bank consulted the IMF during preparation and policy dialogue (World Bank 2019). The program was prepared in close coordination with the IADB, which had a parallel budget support operation. The respective teams coordinated in the design of the policy matrix to avoid duplicating efforts and to ease the reporting and monitoring burden of the authorities. The IADB’s approved operation aimed at promoting fiscal 11 sustainability and reducing the debt level and fiscal deficit. While overlap was avoided, there were few synergies to overcome political economy challenges. In Croatia, the World Bank collaborated with the IMF on analytical and diagnostic work, including preparation of the CTA’s modernization strategy and an update to and a joint diagnostic of CTA reform priorities and implementation (World Bank 2020a). The IMF took the lead in tax policy advice and provided support to tax administration reform through a resident adviser in Slovenia. The World Bank and the IMF worked together on the establishment of the Large Taxpayer Office: the World Bank took the lead in providing advice on criteria for selecting large taxpayers and improving the Large Taxpayer Office’s infrastructure, while the IMF focused on modernizing the processes and governance of the Large Taxpayer Office. The World Bank program also complemented assistance by other donors, especially the EU. The EU provided technical assistance for enhancing capacity, mainly for the administration of the value-added tax and the value-added tax Information Exchange System. The World Bank project aimed to improve the overall capacity of the CTA to implement the EU acquis on taxation, to support the government’s fiscal consolidation effort, and to improve taxpayer services. In Pakistan, the World Bank’s development policy financing prior actions were more effective when they were coordinated with an IMF program and with technical assistance from DFID. In Liberia, the World Bank sought to support the government in implementing the comprehensive public financial management reform strategy by aligning development partner resources into a World Bank—managed multidonor trust fund. However, more frequent dialogue with other donors could have provided opportunities to discuss spe- cific challenges. Although donors participated in missions and working group meetings, several donors indicated an expectation of more intensive dialogue and reporting from the World Bank on project implementation. In Panama, while there was supportive collaboration with the IADB and IMF, it was not sufficiently catalytic to lead to a successful program (World Bank 2022). Discussion of donor coordination included reliance on joint analytical work supporting the 12 development policy loan series (such as the 2006 Country Financial Accountability and Procurement Assessment Report and the 2013 Public Expenditure and Financial Accountability [PEFA] assessment) and was complemented by IADB loans focusing on fiscal regulation. The IMF relied on the World Bank for inputs on tax reform. The IMF focused on anti–money laundering issues, while the World Bank led the dialogue on international tax transparency. Cooperation and coordination with the OECD and the Global Forum was also noted. Results frameworks to track progress on DRM were often not well articulated, sometimes missing baselines or targets, being overly optimistic, or lacking clear results chains. Results indicators to measure the impact of tax reforms sometimes had limited links to the specific reforms or prior actions. In many cases, more appropriate options were available. The quality of indicators related to tax were often an issue, limiting the ability to monitor the impact of the operation or project. Some targets for results indicators were overly optimistic, such as in Panama and Rio State (World Bank 2021, 2022). Other results indicators had little or no link to the specific components of the operation, such as in Liberia and Pakistan, or in Croatia, where there were also no baselines or targets (World Bank 2020a, 2020b, 2020c). Improvement in the articulation of indicators was possible using instruments such as the Tax Administration Diagnostic Assessment Tool, although this would require funding for the assessment. While this is a straightforward issue for an investment project, incremental funding would have been needed for a DPO. In Croatia, the project paid little attention to the design of a meaningful results frame- work, which could have guided implementation (World Bank 2020a). Some indicators had only weak connections to project objectives, with no baselines or targets at the outset. Baselines and targets were only provided during the second restructuring, and the surveys of taxpayers and tax officers—on which three of the four outcome indictors were based—were discontinued before project closing, suggesting limited ownership of the project’s monitoring framework. In Liberia, revenue mobilization and administration was one of three subcomponents under one of the project’s objectives (to improve fiscal policy management; World Bank 2020b). The results indicator was PEFA indicator PI-12 (multiyear perspective in fiscal 13 planning, expenditure policy, and budgeting), which was not closely linked to the sub- component on tax mobilization and administration. In Pakistan, the results framework lacked meaningful indicators, with seven policy areas and 19 prior actions, but only six results indicators (World Bank 2020c). Only two indicators were DRM focused, one of which was only loosely associated with the prior actions. Some targets for results indicators were overly optimistic, such as in Panama and Rio State… Other results indicators had little or no link to the specific components of the operation, such as in Liberia and Pakistan, or in Croatia, where there were also no baselines or targets. In Rio State, the operation’s documentation included little information about the pro- gram’s performance, and information on revenue collection was missing (World Bank 2021). Moreover, objectives were unlikely to be achieved during the program’s 17-month duration, but instead would require sustained longer-term support. Tax collection over the 17 months between loan approval and closing was too short to be an adequate indicator of tax administration quality because it did not control for the economic cycle or issues of tax composition and equity. The results-based system consisted of 20 indi- cators to help guide the state Secretariat of Finance’s policies for human resources and the assignment of financial and technological resources. However, the indicators did not have target values, and their purpose was not specified. The state Undersecretariat of Revenue was responsible for producing the quarterly outcome indicators and mak- ing them available, but the evaluation team was not able to verify their production. In Panama, a target under the first pillar was increasing tax revenues by 2 percent- age points of GDP (World Bank 2022). This was unrealistic given the complexity of the reforms, which sought to broaden the base, eliminate exemptions, and increase tax rates in the face of strong political headwinds both inside and outside government. 14 Even more so, it would have meant an almost 20 percent increase in revenue-to-GDP ratio over a three-year period, a tall order under any circumstances. This optimism was one of the key shortfalls identified in project design and also reflected the project’s optimism in overcoming strong political headwinds to domestic tax reform, including a private sector that did not see the rationale for increasing domestic tax revenues in view of large nontax revenues. Methodology This Evaluation Insight Note draws on a detailed review of six IEG evaluations of World Bank interventions supporting DRM. The interventions were selected from among closed projects and operations for which PPARs had been done, based on having significant DRM content, using different financing instruments, and covering a diverse group of countries.7 Performance ratings (assigned independently by IEG) were not taken into account in the selection process. The projects and operations were implemented in five countries (Croatia, Guatemala, Liberia, Pakistan, and Panama) and one state (Rio de Janeiro in Brazil) covering a range of income levels (one low- income country, one lower-middle-income country, three upper-middle-income countries, and one high-income country). Four of the interventions were financed by the International Bank for Reconstruction and Development, one by the International Development Association, and one with a blend of concessional and nonconcessional financing. In terms of analytic approach, the team reviewed the PPARs, seeking specific factors (or their absence) that were credited with influencing the sustainability of reforms by the country. It subsequently grouped together these individual observations across cases into themes. 7  Tax administration accounted for less than 25 percent of the thematic context of the selected operations (using the World Bank’s theme cod- ing). Projects and operations with significant DRM content were defined as those with more than 6 percent tax content based on independently assigned theme codes (domestic revenue administration [theme code 412]). The objectives of the operations, pillars or components, and prior actions implemented (in the case of DPOs) were reviewed to ensure there was enough material on taxation to be evaluated. 15 References World Bank. 2007. “Croatia—Revenue Administration Modernization Project.” Project Appraisal Document 39219-HR, World Bank, Washington, DC. World Bank. 2012a. “Brazil—State of Rio De Janeiro Fiscal Efficiency for Quality of Public Service Delivery Development Policy Loan.” Program Document 62793-BR, World Bank, Washington, DC. World Bank. 2012b. “Guatemala —Fiscal Space for Greater Opportunities First Programmatic Development Policy Loan.” Program Document 70208-GT, World Bank, Washington, DC. World Bank. 2014. ”Pakistan—First Fiscally Sustainable and Inclusive Growth Development Policy Credit.” Program Document 86373-PK, World Bank, Washington, DC. World Bank. 2017. Tax Revenue Mobilization: Lessons from World Bank Group Support for Tax Reform. Learning Product. Independent Evaluation Group. Washington, DC: World Bank. World Bank. 2019. “Guatemala—Enhanced Fiscal and Financial Management for Greater Opportunities DPL Series.” Project Performance Assessment Report 141817, Independent Evaluation Group, World Bank, Washington, DC. World Bank. 2020a. “Croatia—Revenue Administration Modernization Project.” Project Performance Assessment Report 149797, Independent Evaluation Group, World Bank, Washington, DC. World Bank. 2020b. “Liberia—Integrated Public Financial Management Reform Project.” Project Performance Assessment Report 153799, Independent Evaluation Group, World Bank, Washington, DC World Bank. 2020c. “Pakistan—First and Second Programmatic Fiscally Sustainable and Inclusive Growth Development Policy Credit.” Project Performance Assessment Report 149804, Independent Evaluation Group, World Bank, Washington, DC World Bank. 2021. “Brazil—Rio State Fiscal Efficiency for Quality of Public Service Delivery Development Pol- icy Loan.” Project Performance Assessment Report 158234, Independent Evaluation Group, World Bank, Washington, DC. World Bank. 2022. “Panama—First, Second, and Third Programmatic Fiscal Management and Efficiency of Expenditures Development Policy Loan.” Project Performance Assessment Report 168515, Independent Evaluation Group, World Bank, Washington, DC. 16 This note was produced by Patrick Hettinger (senior economist, IEG) drawing on earlier work by Zeljko Bogetic (Lead Economist, IEG). The work was overseen by Jeffrey Allen Chelsky (manager, IEG) under the guidance of Oscar Calvo-Gonzalez, director, IEG). Corresponding author: Patrick Hettinger phettinger@worldbank.org People who provided useful comments or inputs include Stephen Porter, Estelle Raimondo, and Rashmi Shankar. 17