Debt Sustainability Assessments & Their Role in the Global Financial Architecture ECONOMY AND FINANCE How to Ensure Debt Sustainability Accelerates Sustainable Development Matthew Martin June 2024

Summary One complaint about the Bretton Woods Institutions’(BWI) debt sustainability frameworks that is strongly raised by International Monetary Fund(IMF) member states, academic and independent analysts, and civil society, is that they are not fully compatible with the National Sustainable Development Plans countries have adopted to reach Agenda 2030, or with the National Defined Contributions/National Adaptation Plans adopted to confront the climate crisis. Drawing on steps taken by the BWIs to adapt to the Sustainable Development Goals(SDGs), as well as independent methodologies developed by other analysts, this paper proposes ways in which debt sustainability analysis(DSA) methodologies could be more fully adapted to the SDGs. It begins by making the case for why such adaptation is urgently needed, and then deals with three issues: i) adapting DSAs to overall SDG spending needs; ii) adapting DSAs to urgent environmental crises(climate change adaptation, nature and biodiversity collapse, and natural disasters); and iii) adapting DSAs to urgent social crises(rising extreme inequality and poverty, and global pandemic health events). The paper pays particular attention to feasibility and ease of implementation of the proposed changes, so that they do not overburden BWI staff or government officials working on the analysis and are immediately practicable and actionable. They can therefore relatively easily be fed into the forthcoming review of the Low-Income Countries Debt Sustainability Framework(LIC-DSF), and of the Staff Guidance Note and tools for the Sovereign Risk and Debt Sustainability Framework(SRDSF) for Market Access Countries(MACs). The paper ends by urging that the reviews should emphasise the high positive multipliers SDG spending can have on growth and explore how they can be funded by enhancing non-debt and lower-cost financing. It also urges the BWIs to keep a much closer eye on total public(external and domestic) debt-service burdens to avoid crowding out key spending for the SDGs and to accelerate the provision of rapid liquidity relief on debt where needed.

Table of Contents Summary ............................................................................................................................ 2 1. Introduction: Why should DSAs be adapted to the SDGs?....................................... 1 2. Adapting Debt Sustainability Analysis to the SDGs................................................. 2 2.1 Definition and Background/History of Past Efforts................................................. 2 2.2 The Way Forward: Integrating the SDGs into DSAs................................................. 4 3. Adapting Debt Sustainability Analysis to Combat the Climate Crisis........................ 5 3.1 Definition and Background/History of Past Efforts................................................. 5 3.1.1 Adapting DSAs to Climate Analysis: the SRDSF.............................................. 5 3.1.2 Much More Limited Adaptation: the LIC-DSF................................................ 7 3.2 The Way Forward: Integrating SDGs 13-15 into DSAs............................................. 7 4. Adapting DSAs to Combat the Extreme Inequality Crisis........................................10 4.1 Definition and Background/History of Past Efforts............................................... 10 4.2 The Way Forward: Integrating SDG10 into DSAs................................................... 10 5. Overall Conclusions and Recommendations..........................................................13 Table 1: Summary of current situation and reform proposals....................................14 References...............................................................................................................16 Interviews...............................................................................................................18 Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture

1. Introduction: Why should DSAs be adapted to the SDGs? In 2015, all the world’s governments agreed on Agenda 2030 and the Sustainable Development Goals(SDGs). These have since become the basis for all national development plans and goals, especially in countries of the global South, covering the period 2015-2030. Both of the Bretton Woods Institutions have integrated the SDGs into their work programmes. The World Bank already has two SDGs(1 – end poverty, and 10 – reduce inequality) as its core goals for its own operations, and is in the process of adding a climate goal and strengthening its inequality goal(World Bank 2024); and the IMF has framed much of its recent work around adapting to and being compatible with the SDGs, notably on confronting the climate crisis, promoting gender equality and reducing inequality, because of the strong negative impact such factors have on its mandate of promoting growth and financial stability(IMF 2023). Yet until 2020 for low-income countries(LICs) and 2022 for“market access countries”(MACs), debt sustainability analysis by the BWIs remained stuck in a world which took virtually no account of Agenda 2030. While the SDGs implied a doubling or trebling of government spending in many countries(Sachs and SchmidtTraub 2014), the BWIs did not go beyond calculating some extra spending needs in a few countries and developing a toolkit for doing similar work in other countries. They failed to finish the work by indicating at country level how such spending needs could be financed without compromising debt sustainability and by helping countries mobilise funding on the basis of such SDG scenarios. Faced with massive additional spending needs, especially as it became clear the world was failing to mobilise the financing for them, they reverted back to“incrementalism”(small increases in spending) in country fiscal frameworks, and limited DSAs to analysing the risks of default arising from such frameworks and/or changes in macroeconomic prospects/financing costs, rather than as identifying how to finance sustainable development. This was an entirely“unsustainable” position. According to the Intergovernmental Panel on Climate Change (IPCC) forecasts, countries like Tuvalu or Chad have little sustainable future unless they make plans NOW to combat the climate crisis, with plans that are sustainably funded. Many countries are potentially vulnerable to the climate crisis undermining their growth prospects and economic stability. In the same way, IMF and World Bank research has shown many times how extreme poverty, income inequality and gender inequality are undermining growth in many countries(on inequality, see IMF 2014 and 2017; on poverty see World Bank 2018). These negative impacts are not limited to individual countries: to the degree that life becomes less tenable and extreme poverty and inequality more widespread in many countries of the global South, there will also be higher levels of cross-country migration and insecurity across the world. The failure to ensure that SDG spending was adequately funded and therefore increasing – or that debt service was being kept at reasonable levels where countries were trying to spend more on the SDGs – meant that by 2023 Global South countries were spending much more on debt service than on key SDG sectors. According to the Debt Service Watch 1 database, debt service in 2023 is almost exactly equal to total SDG core social spending (on education, health and social protection) across 139 countries borrowing from the World Bank. In Africa and LICs, it exceeds social spending by almost 50 per cent. Looking at individual sectors, debt service is on average 2.5 times education spending, 3.7 times spending on health and 11 times social protection expenditure. The relationship with climate spending is equally startling(partly because delivery of climate finance via government budgets has been very low): on average across 42 countries for which data are available, debt service is 12 times climate adaptation spending in 2023, rising to 13 times in 2024. 1 Debt Service Watch is a database compiled by Development Finance International and launched in late 2023, which tracks debt service and spending on the core social and environmental SDGs, across all countries which borrow from the World Bank. It differs from other debt service data in that it covers both external and domestic debt service and is compiled in real time as soon as budget documents and debt management reports are released by developing countries, so that current data are for 2023. Development Finance International (2023a) presents its overall debt and social sector findings, and the summary database; and Development Finance International(2023b) presents its findings on debt service and climate adaptation spending. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 1

Put more positively, as discussed in detail below, investments to combat the dual crises of climate and inequality for a genuinely just green transition – and to prevent or recover rapidly from future catastrophic events such as natural disasters or health pandemics- provide the best prospects of highest returns and a path to dramatically accelerated growth in most countries, much better than expected outcomes from traditional infrastructure spending. To the degree that these effects are not being demonstrated in debt sustainability analysis, they are being ignored and governments are not being encouraged to mobilise funding to support transformative spending, nor to see how investments in these areas could show a path to greater debt sustainability and borrowing capacity. The BWIs have recognised this in recent years and begun to adapt their DSAs to these needs, but they still have a long way to go. BWI staff interviewed for this study have recognised that much more could and should be done – on which they are in agreement with member government officials, independent analysts and CSOs. In addition, all stakeholders agree that i) such analysis must take account of limited personnel and budgets in the BWIs(and even more limited personnel in member countries); and ii) it must produce clear and transparent findings for all stakeholders, for which governments and the BWIs can be held accountable. The remainder of this paper is structured as follows: • Section 2 looks at the broad issue of adapting DSAs to overall SDG spending and financing needs; • Section 3 examines what more can be done to adapt DSAs to take account of urgent environmental crises confronting the planet(focusing on the climate crisis- SDG 13 – but also emphasising biodiversity and the marine environment – SDGs 14 and 15); • Section 4 examines how to adapt DSAs to take account of urgent social crises, as exemplified by the extreme inequality crisis(SDG10), which is perpetuating extreme poverty(SDG1) and undermining attainment of all the other social and environmental SDGs; • Section 5 concludes by drawing together the analysis and prioritising recommendations. The original outline of this paper envisaged a separate section to deal with how DSAs could incorporate the impact of environmental and social"shocks” such as climate-related natural disasters or pandemics on debt sustainability. However, these events should no longer be considered shocks. In the countries most strongly affected by natural disasters, such events happen at least every two years, and with increasing regularity and frequency across almost all affected countries(see Section 3); and the latest forecasts for pandemics indicate that there is a 14-23 per cent chance of another pandemic happening between now and 2030(see CGD 2021 and Marani et al 2022). Therefore, this paper suggests that these“shocks” should be considered as forecastable events and included in the baseline scenarios for debt sustainability analysis. Alternatively, they could be included as“stress tests” on the same basis as other likely events, including commodity shocks, changes in financial market conditions, etc. The types of events and how they can be simulated are discussed in each of the environmental and social sections below. 2. Adapting Debt Sustainability Analysis to the SDGs 2.1 Definition and Background/History of Past Efforts If debt sustainability analysis is to be truly compatible with the SDGs and Agenda 2030, it should involve i) working out at the level of each country how much the SDGs would cost to attain and integrating this fully into government forecasts of financing needs between now and 2030; and ii) working out what financing“terms” governments could afford, to fund these needs while keeping debt sustainable. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 2

Neither of these steps is impossible. Development Finance International(then known as Debt Relief International) helped more than 30 Heavily Indebted Poor Countries(HIPCs) to analyse precisely these issues for themselves during 2005-2012 through work under the HIPC Debt Strategy and Analysis Capacity-Building Programme(DSA CBP), funded by seven donor governments and increasingly by the beneficiary governments themselves. This was achieved by encouraging them to conduct their own DSAs, using as a“high case scenario” the spending needs and financing sources for the then Millennium Development Goals, which in most countries were falling way behind schedule and required accelerated implementation and more concessional financing. In several countries such as Burkina Faso, Rwanda and Tanzania, use of these scenarios at donor Consultative Group and Round Table meetings helped to mobilise much more concessional financing and to accelerate progress on many of the core MDGs. However, in many more countries which had less support from development partners, the high case scenarios were not presented to donors because they were seen as “unrealistic”, i.e. too expensive. The most ambitious such exercise was undertaken in Bolivia in 2007, funded by GIZ and implemented in partnership with the Centro de Estudios Monetarios Latino-Americanos. It identified how to finance the new government’s ambitious development programme, the Plan Nacional de Desarollo(National Development Plan, NDP): Bolivia Digna, Soberana, Productiva y Democrática para Vivir Bien, 2006 – 2011(see Gaceta Oficial 2007). This plan was even broader than the SDGs, including wellbeing, community, cultural and religious issues. The workshop identified massive funding needs for the plan, concluded that the prospects for concessional financing were limited, and recommended as the main feasible financing source taxing the hydrocarbons sector heavily (which government did), and feasible amounts of new external and domestic loans. Methodologically, after initial investment in developing a methodology and forecasting template through global and regional consultations, for each country a two-week workshop building on existing national development plans and Poverty Reduction Strategy Papers, and involving around 30 national staff, was able to cost the NDP goals, and work out a high case scenario to fund them while keeping debts sustainable. The programme was decentralised to regional partner organisations in 2012, after which time its methodology was largely replaced by the LIC-DSF and MAC-DSF. Efforts have continued to link debt sustainability analysis and the SDGs. The most comprehensive effort to assess country spending needs for the SDGs was undertaken by the UN Sustainable Development Solutions Network (Sachs and Schmidt-Traub 2014). During 2014-2018 they first produced a paper bringing together sectoral assessments of spending needs to estimate total global SDG spending needs, then sponsored a process of bringing together key global experts, including from the IMF and World Bank, covering all the SDGs to refine methodologies for assessing spending needs into tools countries could use. However, this methodology was not taken up by the whole UN system due to lack of funding and was partly included in subsequent IMF work. The spending needs for many SDGs are currently included in some countries’ National Development Plans (NDPs). In some cases, countries have developed the costings themselves; in others(e.g., Benin), the United Nations Development Programme(UNDP) played a key role(assisted by other UN agencies) in making such plans SDG-compatible and helping countries to cost the resulting spending needs. However, such detailed costing efforts were limited to a relatively small number of pilot countries, partly because it was not obvious where the financing would come from, and partly because UNDP lacked the funds to provide such support in all countries. 2 UN Trade and Development(UNCTAD) has also made efforts to model SDG costs across a much broader range of countries but using estimated shares of GDP which have been derived by other UN sectoral specialist agencies, and mainly in order to look at the global impact of funding the SDGs on worldwide debt sustainability and financing needs(for example, see UNCTAD 2019). 2 The author was involved in discussions of spending needs estimates in Benin, as part of a DFA/INFF mission for UNDP, in which donors were largely dismissive of the idea that the country could find sufficient funds to fund all the SDGs. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 3

In addition, the IMF has made efforts to cost some of the“core” SDGs for a limited number of countries, and to develop a methodology to replicate this exercise across all countries. This has involved doing two rounds of work on detailed country costings for some of the SDGs(education, health, electricity, water and sanitation, and rural roads). The first was released in January 2019, looking at SDG costings for countries with different income levels, and with detailed case studies of Benin, Guatemala, Indonesia, Rwanda and Vietnam to present different country types(IMF 2019). The second major multi-country study, released in April 2021 and updated to take account of the negative impact of the COVID pandemic on SDG progress and prospects, concluding that even more financing is needed. Case studies were also completed for Cambodia, Nigeria, Pakistan and Rwanda(IMF 2021b). However, such costings by the different organisations have not generally been included in the key documents used by governments to push donors to mobilise more funding, or to discuss with their own citizens why more tax revenue would be needed to reach the goals. Efforts were made in some early UNDP-sponsored DFAs/INFFs (Benin, Cameroon) to include such costings and identify funding sources, and by the government of Rwanda to use IMF costings to guide donor meetings, but these have not been replicated more recently because the shortage of funding for all SDGs has dominated discussions. The IMF has also for more than a decade been incorporating second-round effects of investment on accelerating growth into its forecasts of specific economies and policy advice, through separate case studies and model analysis conducted in Article IV 3 documents, Selected Issues papers and other studies. 4 Most of the analytical work conducted with this model has been on the growth impact of large infrastructure project investments but, according to the authors, there is no reason why it cannot also be used to simulate the impact of capital investment, including investment in human capital. What has been missing in all of these analyses is a formal incorporation of their results into the LIC-DSF or the various DSFs for MACs. 2.2 The Way Forward: Integrating the SDGs into DSAs There is therefore a sound basis on which to integrate the SDGs into DSAs. Four further steps are needed: 1. Make the costing methodologies used by the IMF and other UN agencies for each SDG consistent, so that governments can use results with full confidence that they will be acceptable to all agencies; 2. Include all of the SDGs. Costing methodologies now exist for all the SDGs, and it is essential to broaden coverage beyond the five IMF sectors. 3. Incorporate fully the effects of SDG spending on growth and debt sustainability, including spending on human and environmental capital as well as on physical infrastructure(for more, see Sections 3 and 4). 4. Integrate these costs with financing prospects in the LIC-DSF and SRDSF forecasts as an accelerated “SDG needs and impact” scenario. There does not seem to be any lack of willingness among the institutions interviewed for this study(or among independent organisations such as SDSN and DFI) to do more of this work, and in a cooperative joint manner. However, all say funding is lacking for work on methodology to be completed and done comprehensively and routinely in all countries. 3 Article IV of the IMF Articles of Agreement includes reporting obligations and regular inspection missions. 4 A good example of this type of model is the Debt, Investment and Growth(DIG) model developed in IMF(2012), which the IMF is still recommending as the simplest way to model the impact of investments on accelerating growth. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 4

However, we need to take account of the current context. Even before COVID, the SDGs were way off track in most countries. After COVID, increasing numbers of international leaders and experts have concluded that the SDGs are a pipe dream. Even the UN Secretary General’s SDG Stimulus Plan envisages finding only another US$500 billion a year in financing. Assuming that this plan succeeds(which many sceptics doubt unless there is a fundamental change in global taxation, debt relief, or ODA efforts), this would mean that we are still US$1.9 trillion short of annual global spending needs to reach the SDGs. It is also not evident that any new money would go to LICs and LMICs, who need it most and whose share of total ODA has fallen since 2018. As a result, most countries are going to have to make some very hard choices about which of the SDGs to prioritise. In practice, most countries have already been making such choices through key political pledges by their leaders or parties in election campaigns – for example Benin and Sierra Leone focusing on water, Gabon and Sierra Leone on health. Such choices should not be made by the BWIs or the broader development partner community, but instead by participatory development of SDG acceleration/stimulus plans, so the priorities chosen may vary by country(though in similar exercises around Poverty Reduction Strategy Papers, PRSPs, there was a remarkable degree of citizen consensus in almost all LICs/LMICs on top priorities: education, health, nutrition/food and water). On the other hand, it is possible for the international community to provide advice on which sectors might produce the greatest investment multipliers and impacts on growth, as well as advances in sustainable and human development. 5 Here the evidence is clear: investments in a just green transition(i.e. in reaching both the social and the environmental SDGs) produce far higher multipliers than investments in traditional infrastructure, energy, or land/sea use- both directly and indirectly, by reducing inequality and climate damage(see, for example, IMF 2021a and 2017). These greater multiplier effects would enhance countries’ debt-carrying capacity and debt sustainability. On the other hand, failure to deal with the climate and inequality crises would dramatically undermine growth and security prospects, reducing debt-carrying capacity and sustainability. It is for these reasons that the rest of this paper focuses on integrating into DSAs the top priority spending on the environmental and social SDGs. 3. Adapting Debt Sustainability Analysis to Combat the Climate Crisis 3.1 Definition and Background/History of Past Efforts This is where the IMF and World Bank have made the most advances in adapting methodology, notably in the 2020 Sovereign Risk and Debt Sustainability Framework(SRDSF) for Market-Access Countries. 3.1.1 Adapting DSAs to Climate Analysis: the SRDSF The SRDSF is the framework which has made the most systematic and comprehensive adaptations(for more details see IMF 2022b), through a climate change module with two“sub-modules”. The first of these models the impact of adaptation investments, which build resistance to the effects of climate change, the second covers climate change mitigation, which involves efforts to reduce greenhouse gas emissions to limit increases in temperatures. 5 It is vital to underline that“investments” mean both recurrent and capital spending. There persists in many international agencies a preference for capital spending and an urge to reduce recurrent spending, even when it is obvious from many sectoral studies that in social and environmental sectors, recurrent spending(especially on staff wages, training and maintenance) is as vital as capital spending. See Development Finance International(2016). Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 5

Each sub-module allows projections over a 30-year horizon under two scenarios: an“extended standardised baseline” scenario based on the default costs in the template, and a customised scenario, where users can adjust the costs to country-specific characteristics. The customised scenario also allows users to adjust the financing terms of the climate-related investments, providing scope for example to show the difference between financing with non-concessional or concessional debt. It also allows users to adjust the long-term GDP growth path, which in principle would provide space to incorporate the results of the positive growth impact of any just green transition spending. Each is also based on a clear costings methodology applicable across all countries, reflecting the“currently best set of estimates” by the IMF for individual country costs of adaptation(ranging between 0.3 per cent and 2.4 per cent of GDP for different regions and types of countries – see IMF 2022a) and mitigation(between 1 per cent and 4 per cent of GDP). 6 To facilitate the task of integrating climate adaptation into the DSF, the adaptation module is“pre-populated” with these estimates, making it easy to use while giving the user plenty of flexibility to change the default assumptions. There is also a clear definition of which countries to analyse: • For countries which request financing from the Resilience and Sustainability Facility(RSF) of the IMF, both submodules have to be analysed. • Use of one or both of the modules is also compulsory in pre-defined groups of countries in which the fiscal costs and risks of adaptation or mitigation are expected to be significant. ⁻ The adaptation submodule is compulsory in(1) the set of countries for which the natural disasters stress test is triggered 7 and(2) the top 25 per cent of countries at highest risk from climate change, as judged by an IMF-calculated Adaptation Ranking Index. 8 ⁻ The mitigation submodule is compulsory for all countries with an ambitious zero net carbon emission target(targeting zero net carbon emissions before 2050), as well as for the 25 largest CO2 emitters per unit of output, who have yet to set a target. • Use of the adaptation module is also compulsory in debt-restructuring cases, to provide guidance to teams who need to formulate realistic debt restructuring envelopes – though there is no evidence that as a result more emphasis is placed on restructuring, helping to raise climate spend. For the remaining IMF member states, use of either module remains optional, based on the views of the authorities and the IMF mission as to the likelihood of high climate risk. However, according to IMF staff interviewed for this study, no country for which the module is not mandated has chosen it as an option. Furthermore, based on the IMF’s own estimates, the fiscal costs of climate adaptation could be sizeable for many developing countries. If costs are high, this can be supplemented with a second step, the building of a customised country-specific scenario which looks at how these costs can be funded while maintaining debt sustainability. Here, missions are encouraged to look for other sources as estimates of costs. These include studies by the IMF, World Bank, or regional development banks(RDBs); Nationally Determined Contribution 6 These are based on costs for European countries with ambitious mitigation targets as reported to the European Commission. 7 The SRDSF also contains a“natural disaster stress test” – which, interestingly, is triggered for all small state MACs(unlike the LIC-DSF test), as well as MACs with evidence of frequent or severe disasters. This is not discussed here in detail because it is not seen as an adaptation to climate change, but rather an adaptation to natural disasters which could have multiple causes; and because the climate change modules are much better adaptations. For details of the LIC-DSF stress test, see section 3.1.2 below. 8 This in turn combines information on(i) propensity to natural disasters, from EM-DAT;(ii) climate-related adaptation cost estimates, from IMF 2022a; and(iii) climate-related adaptation risk, measured by a Composite Index calculated with data from the Notre Dame University ND-GAIN Index, the IMF-INFORM index and the United Nations Institute for Environment and Human Security’s World Risk Index(WRI). Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 6

(NDC) reports to the United Nations Framework Convention on Climate Change(UNFCCC); or alternative cost proxies based on other countries with similar rankings in the IMF INFORM risk index(see IMF 2024a). This can also include the impact of spending to promote resilience on growth(and on protection and recovery from natural disasters) using the IMF’s Debt-Investment-Growth and Natural Disasters(DIGNAD) model, as has been done for Bangladesh and Rwanda in Resilience and Sustainability Trust(RST) Board papers. 9 However, the cost levels presented by the IMF in its Staff Guidance Note are much higher than the standardised scenario, ranging up to 3.8 per cent of GDP for adaptation(50 per cent higher than in the standard scenario) and 14 per cent for mitigation(four times as high as standard), thereby raising the question as to whether the standard scenario is really useful as an indicator of potential costs and risks of climate change for debt sustainability. 3.1.2 Much More Limited Adaptation: the LIC-DSF In contrast to the major reforms made in the integration of climate into the SRDSF, there has been only very limited progress with the LIC-DSF. Adaptations have been much more limited in scope as follows: • The impact of climate change is limited to physical risk, such as climate-induced natural disasters, and omits the impact of adaptation, mitigation and resilience spending of the types included in the SRDSF. • Climate is relegated to a“stress test”, rather than being treated as an extension or modification of the baseline scenario. • It is used only for small states vulnerable to natural disasters as defined by the IMF(20 countries) 10 , as well as other LICs that have met a frequency criterion(two disasters every three years) and economic loss criterion(above 5 per cent of GDP per year), based on the EM-DAT database during 1950–2015. The use of such historical data on frequency and severity should also be examined, as all more recent data and forecasts indicate growing frequency and severity, so the analysis might better be conducted for all countries where natural disasters are forecast to be frequent and severe. • In spite of the fact that country eligibility requires countries to be hit regularly by shocks, the stress test natural disaster shock occurs only in year two of the projection and is not repeated over the longer term. • For this paper, the author has checked the remaining LIC-DSFs for small states 11 for any inclusion of a climate stress test and found none – which stands in marked contrast with repeated and extensive analysis of climate adaptation and resilience spending plans(for example for Cape Verde and Timor Leste) and warnings even in the texts of the DSAs of major unquantified downside climate risks to debt sustainability. The lack of analysis of climate impact in the DSA therefore stands out like a sore thumb. 3.2 The Way Forward: Integrating SDGs 13-15 into DSAs Currently, there are seven main criticisms of the way in which climate has so far been integrated into DSAs: Country Coverage: The number of countries covered by the analysis is potentially far too low. In terms of the adaptation submodule, it is only the 25 per cent of countries which are considered most at risk(plus any applicants for the RSF) which have to be analyzed, whereas costs for adaptation are likely to be substantial for at least the top 50 per cent of countries(for example, 68 countries are members of the V20 group and around 80 are considered“climate vulnerable”). In terms of the mitigation submodule, the threshold set for using the module- countries which set a target of net zero by 2050- is no longer ambitious as it has been adopted by 93 9 As well as IMF working papers for the Maldives, St Lucia and Vanuatu; Selected Issues papers for Solomon Islands, Timor-Leste and Uganda; and Climate Macroeconomic Assessment Program pilots for Madagascar and Samoa. For more details see IMF 2024b. 10 Based on the countries defined as extreme or high vulnerability in Annex 1 of IMF(2016) 11 For Bhutan, Cape Verde, Guyana, the Marshall Islands and Timor Leste. Djibouti’s DSA is unpublished but is understood based on interviews also not to include any climate stress test. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 7

per cent of countries according to the UNFCCC. It would therefore make sense for the extended baseline scenario modules on adaptation and mitigation to be used for all countries to test whether climate spending constitutes a key risk for their debt sustainability and/or could accelerate growth enough to increase debtcarrying capacity. Underestimating Climate Spending Needs: Estimates of climate spending needed are far too low. As discussed above, country-specific costs for these aspects are 50 per cent higher for adaptation and 300 per cent higher for mitigation than those in the standard scenarios. The Staff Guidance Note acknowledges that its cost estimates for individual countries, which would generally be used for the customised scenario, include only two types of adaptation investments: strengthening physical assets and investing in coastal protection, i.e. protection against floods, storms and sea level rise. The mitigation costs presented for non-EU countries in the Staff Guidance Note also exclude investment in buildings, on the grounds that it is harder to distinguish which are normal building maintenance and which are climate mitigation investments. The Staff Guidance Note also fails to capture investments needed to protect against other important climate risks, including droughts and heat waves. Major investments for adaptation, mitigation and resilience are omitted from SRDSF baseline and customised scenarios. In addition, to keep the baseline scenario agreed with the IMF for the first five years“clean”, the DSA adds the extra climate spending in only from year t+6 onwards, and therefore, in spite of the urgency of the climate crisis, does not provide any simulation of a potential additional scaling up of spending on the climate crisis before 2030. The Staff Guidance Note on the SRDSF acknowledges that there will be major modeling and data advances in the calculation of country spending needs for adaptation and mitigation, and that therefore the Fund overall, and individual country missions, should keep assumptions under review. IMF staff interviewed and participants at the seminar in April 2024 acknowledged that there is therefore a need to review the assumptions made in the Guidance Note published in August 2022. Ignoring the Positive Impact of Just Green Transition Spending: It is important to realise that all the reforms present additional spending on confronting the climate crisis negatively as“additional costs and risks” for financing needs and debt sustainability. In other words, they can only as currently constructed have a negative impact on the assessment of sustainability, by adding an additional high risk from high climate spending and related borrowing. 12 This could then be offset by assuming more concessional terms for financing such spending, in the customised scenario. There is no mention of major positive multiplier effects of spending on a just green transition on growth, jobs, or debt-carrying capacity, in spite of work by the Fund showing higher economic multipliers from such spending(IMF 2021a). As discussed above, the IMF has simulated the impact of resilience spending on growth and on reducing the impact of natural disasters using the DIGNAD model, but it has not used DIGNAD to project the impact of broader climate adaptation or mitigation investments on growth. The IMF and World Bank have also both been working extensively on improving the efficiency of climate-related investments through climate spending-specific Public Investment Management Assessments and Public Expenditure Review analyses(see IMF 2024c and World Bank 2022) and broader efforts to improve“Green Public Financial Management”, which they suggest could improve investment outcomes by up to 20 per cent, but have not integrated these results into climate adaptation or mitigation spending scenarios. Failing to Combine the Multiple Impacts and Spending Needs of Climate Change: Climate change analysts distinguish the different spending needs resulting from climate change as being for resilience(to protect against 12 The Fund argues that it is implicitly taking into account the higher growth effects of climate spending, by assuming no deviation from the baseline growth path even though climate change will have negative effects on growth. However, a much more transparent solution would be to show the negative effects of climate change on growth in the baseline scenario, and the positive effects of anti-climate change investments in the extended baseline. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 8

disasters), loss and damage(to rebuild the country after disasters), adaptation and mitigation. As currently constructed, the adaptation of the LIC-DSF to climate covers only the loss and damage impact on spending and other macroeconomic variables. The adaptation of the SRDSF covers adaptation and mitigation separately(and for different countries) in different long-term submodules, and loss and damage in a separate disaster stress test. Nowhere is there a combined overall climate impact scenario(for more on this, see Maldonado and Gallagher 2022). Ignoring Other Environmental Goals: Other key environmental goals should be included in the DSFs in the same way as climate. Finance for Biodiversity(F4B) and the School of Oriental and African Studies(SOAS)(see Kraemer and Volz 2022) have recently analysed the scale of natural capital and biodiversity-related risks(SDGs 14 and 15) and, together with major global CSOs such as AVAAZ(2023), have argued they should be included in the debt sustainability analysis in the same way as been done with climate in the SRDSF. This is based on the facts that nature-collapse-related risks are increasingly being taken very seriously by others who are assessing risks to economic sustainability and that work by the World Bank(2021) and credit rating agencies has provided the tools for assessing the scale of these risks and their macroeconomic impact, and for integrating this relatively easily into DSFs. SOAS has conducted such an exercise using the SRDSF tool and has shown in case studies of Bangladesh and Viet Nam that the impact of a nature collapse is greater than even the IMF’s combined worst case macroeconomic stress tests included in DSFs, and that there would also be very severe negative effects on debt sustainability in Indonesia and Nigeria. As one example, a nature collapse shock would raise Bangladesh’s debt to GDP ratio by three times as much as did the COVID-19 pandemic. However, this work has not included any assessments of the public spending needed to prevent nature collapse, and the positive or negative impacts this could have on debt sustainability. There is little doubt that following the December 2022 Biodiversity Summit COP 15 and the 2023 Climate Change COP 28 in which it was agreed that climate and biodiversity are inextricably linked, and that countries should therefore include their nature spending needs in the NDCs for the UNFCCC, pressure to include this extra spending in the DSFs will grow in future years. Failure to Link to Other IMF Targets and Processes: It should be noted that even if climate and nature goals are included in the DSA projections they are a long way from becoming top priority spending as classified by other IMF tools. It will also be vital to protect the spending analysed in this scenario from any broader budget cuts, by including it in the“indicative spending floors” which are fixed in IMF programmes but currently limited to a subset of social spending(for a recent analysis of these floors, see Oxfam 2023). As a result, their title should be changed to“indicative social and environmental spending floors.” Equally, the IMF could consider linking its own lending and other facilities(such as the Catastrophe Containment and Relief Trust, CCRT) to a more vulnerability- rather than income-led classification of countries, which would allow countries highly vulnerable to climate change(and other environmental or natural disaster) shocks to access CCRT debt-service cancellation and concessional lending provided by the RST, regardless of their income level. Alternatively, the IMF could provide other debt-relief modalities such as disaster debt service suspension clauses to countries of all income levels(see also Gallagher et al 2023). Not Giving Sufficient Emphasis to Climate Scenario Results in Summary Presentations of DSA Results: Looking across the range of DSAs for which climate modules have been used, there are many where the(usually very significant) implications of climate are not mentioned in the summary presentation of the DSA results and of how the decision on risk rating has been taken. This seems partly to reflect the separation of the impacts of adaptation and mitigation, again making the case for why the results of the two modules should be combined in presentation. One example is the 2023 DSA for Trinidad and Tobago(IMF 2023b), in which the combined adaptation and mitigation modules lead to an increase of 30 percentage points in the debt/GDP ratio, but all the emphasis in the summary is placed on the potential impact of failure to reform the pension system for an ageing population. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 9

4. Adapting DSAs to Combat the Extreme Inequality Crisis 4.1 Definition and Background/History of Past Efforts As discussed, many experts see the extreme inequality crisis 13 in many countries in the world as having major negative effects on economic growth, political stability, insecurity and crime, progress on education and health indicators, and many other key aspects of development – which are just as significant as those being produced by the climate crisis. In addition, analysts of the climate crisis including the IPCC have emphasised that the climate crisis cannot be successfully tackled without also sharply reducing inequality and poverty through a “just” green transition. Nevertheless, in contrast to efforts on climate, this is an area in which the IMF and World Bank have made no adaptations to their methodology, in either the SRDSF or the LIC-DSF. Indeed, the only mention of the“social sector SDGs” in the DSFs is the inclusion of a module in the SRDSF, tracking the potential negative demographic risks for debt sustainability(in countries with a growing proportion of elderly citizens) of growing costs for social security, pensions and public healthcare(see IMF 2022b). 4.2 The Way Forward: Integrating SDG10 into DSAs In spite of the lack of progress so far, it is easy to see how spending which would have a key impact on reducing extreme inequality and poverty could be factored into debt sustainability analysis. As with climate, this could be done through a module in the SRDSF(and preferably a similar module introduced as part of the review of the LIC-DSF, as discussed in 3.2.2. above for climate). This module could track the increases in the key types of spending needed to reduce inequality and their impact on debt sustainability. These types of spending have been shown in multiple studies to be education, health and social protection; additional sectors with strong impacts are water and sanitation, public housing, public transport, access to electricity and small-scale infrastructure such as rural roads and markets. It is suggested that the core sub-module of this module would include the sectors where the type of spending has the most significant impact on inequality across all countries: this could apply only to education, health and social protection; or be expanded to cover the other sectors the IMF included in its previous analysis of SDG costs – electricity, water and sanitation, and rural roads. The additional sectors(especially public housing and public transport) could be put in a second sub-module, mainly because reliable estimates of their costs across multiple countries are available only for OECD and some MAC countries, and used only when the spending levels in these sectors reach a trigger threshold as a percentage of total government expenditure. This would resemble the use of the adaptation and mitigation sub-modules for climate currently contained in the SRDSF, which are applied to different countries based on particular triggers. As with the SRDSF climate module, such a module could provide projections over a 30-year horizon under two scenarios: an“extended standardised baseline” scenario based on default costs in a template, and a customised scenario, where users can adjust the costs to country-specific characteristics. The customised scenario would also allow users to adjust the financing terms of the inequality-related investments, providing scope to show the difference between financing with concessional debt, tax, or grants. 13 Some interviewees have indicated that the suggestions in this section might be more palatable if they were phrased as adapting DSAs so as to ensure the end of extreme poverty(SDG1), given that the key sectors in which spending would be needed would be broadly similar. However, the World Bank has stated clearly that it will be impossible to end extreme poverty without dramatic reductions in inequality, so the focus is on SDG10. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 10

It would also allow users to adjust the long-term GDP growth path – thereby in principle providing space to incorporate the results of the positive growth impact of any anti-inequality spending. The IMF’s own work(IMF 2017) has shown the very substantial impact reducing inequality could have on accelerating growth by up to 5 per cent a year in the countries with highest inequality: it would accelerate per capita real GDP growth by between 0.15 per cent and 0.4 per cent(the highest numbers applying to the most unequal countries) for every percentage point by which the Gini coefficient is reduced. Therefore, it will be vital that this accelerated growth rate be included in the module, to provide a realistic assessment of the positive as well as negative impact of anti-inequality spending on debt sustainability. As mentioned in Section 2, there exist clear costings methodologies for each sector, most of which(for education, health, electricity, water and sanitation, and rural roads) have been agreed by the IMF and applied across eight countries. These reflect the“currently best estimates”(to use the phrase applied in the SRDSF climate module) by the IMF for individual country costs for spending on these sectors which will be sufficient to reach the SDGs for different regions or types of countries. It would also be highly desirable to add in costs for social protection, given that widely accepted methodologies exist, courtesy of the International Labour Organization(ILO)(ILO 2024), and taking into account the crucial role social protection played in protecting overall progress on all SDGs during the COVID-19 pandemic. To facilitate the task of integrating anti-inequality spending into the DSF, the module could again be“pre-populated” with estimates of total likely spending as a proportion of GDP for different country income level groups, making it easy to use while giving the user plenty of flexibility to change the default assumptions. Several experts interviewed for this paper have suggested that, given the close links between the climate and inequality crises, and the need to tackle both urgently in virtually all countries, all countries borrowing PRGF or RSF should be analysed using both the climate and inequality modules. However, if this is initially too ambitious, it would be relatively easy to define which countries should be analysed using the module: • For countries which request financing from the Poverty Reduction and Growth Facility(PRGF) of the IMF, the first submodule would have to be analysed. • Use of one or both of the submodules would also be compulsory in predefined groups of countries in which the fiscal costs and risks of reducing inequality are expected to be significant. ⁻ The first submodule could be compulsory in, for example, the top 25 per cent of countries with the highest inequality levels, as measured by their Gini coefficients after current tax and transfer measures. A simple threshold for such an analysis could be set at a Gini of 0.4, which would cover around 57 countries, and would match the levels considered to be“high inequality” by the UN and the World Bank. 14 More complex methods of setting such a threshold, using other inequality and poverty indicators, could also be devised. ⁻ The submodule could also be used in all countries which have less high inequality(for example a Gini of between 0.35 and 0.4) but where the country’s government has set itself a clear goal to reduce inequality and growing inequality has been identified as a problem by the World Bank and/or IMF(this would cover countries such as Ethiopia, Kenya, Mongolia, Senegal, Sierra Leone and Viet Nam). ⁻ The second submodule could be used in countries where the country’s government has set itself a clear goal to reduce inequality, and where data on costs of spending on the broader sectors are available(which from DFI’s experience are generally OECD Member countries and some middle-income countries in Asia and Latin America). 14 For the definition of these levels and the countries which would be covered by them, see Martin and Kripke 2023, used as a submission into the review of the UN progress on SDG10. These are also the levels which were suggested as representing“high inequality” which could be judged as“macro-critical” by Fund staff interviewed as part of the process of compiling the paper. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 11

• Use of the main submodule could also be compulsory in debt-restructuring cases, to provide guidance to teams who need to formulate realistic debt restructuring envelopes, on the same grounds as the use of the climate adaptation submodule in the current SRDSF. For the remaining IMF member states, use of either module could remain optional, based on the views of the authorities and the IMF mission as to the likelihood of a high inequality risk. On the other hand, if the process of including estimates of spending needed to reach the core SDGs is relatively straightforward, it would make sense for the extended baseline scenario module on inequality to be used for all countries, in order to test whether anti-inequality spending constitutes a key risk for their debt sustainability and/or could accelerate growth enough to increase debt carrying capacity. 15 As with the climate module, based on the IMF’s own estimates in its existing studies, the fiscal costs of reducing inequality by reaching universal education, health care and social protection could be sizeable for many developing countries. If costs are high when the core submodule is run, then this would be supplemented with a second step, with the building of a customised country-specific scenario which looks at how these costs can be funded while maintaining debt sustainability. As with the adaptation submodule of the SRDSF, missions would then be encouraged to look for other sources as estimates of costs: virtually every country has available costings for education, health and social protection as part of its national and sectoral development plans(which would need to be updated for the delays caused by the pandemic and the recent widespread high inflation), or the IMF could in cooperation with the country authorities conduct the same calculations using globally-agreed costing methodologies as it did for the country SDG case studies since 2019. This more customised scenario could also include the impact on growth of spending to promote resilience against health pandemics, as well as extra costs of health prevention/treatment and social protection during pandemics), though this would require a further adaptation of the IMF’s DIGNAD model, in consultation with global pandemic experts. In addition, the customised scenario would also allow analysis of an alternative financing path which would allow debts to stay sustainable while funding the key anti-inequality and anti-poverty spending needs. This once again is a relatively straightforward task, involving making additional assumptions about increasing especially non-debt creating sources of financing such as tax revenues(which should be provide higher funding in a country with higher per capita GDP as a result of acting against inequality), ODA and other concessional flows, lower external and domestic borrowing costs, and debt relief where needed. These were the types of scenarios DFI helped countries to simulate under the HIPC CBP, and it has recently repeated this exercise for 40 Sub-Saharan countries in a post-COVID context for UNAIDS. These proposals would be implemented in both the SRDSF and the LIC-DSF. In addition, both could include separate ways of dealing with“shocks”. As with the current SRDSF and LIC-DSF“natural disaster stress tests”, the IMF and World Bank could use a“pandemic stress test”, judging the impact of a renewed pandemic on economic prospects. The case for including such a stress test should not need to be argued, given the massive actual negative impact COVID-19 had on economic growth, budget revenue, exports and additional borrowing by countries during 2020-2021, resulting in a major deterioration of debt sustainability assessments among MACs and LICs – and the assessment by the Independent Panel for Pandemic Preparedness and Response that we should take the pandemic threat as seriously as climate change. 16 Given that the COVID-19 pandemic had a major impact on every country, it would be sensible to include this stress test for all countries rather than limiting it to a subgroup based on where the impact on GDP/ budget 15 See section 3.2.1) above for similar argumentation around the climate module. 16 For more details on this panel and its conclusions, see https://theindependentpanel.org/ Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 12

revenue/exports was greatest. Based on the latest expert assessments of the potential frequency of pandemics, the stress test could involve simulating a shock comparable to COVID-19 occurring once in each decade. DFI has recently been including the GDP and revenue impact of COVID-19 as a“stress test” in analysis of social sector(especially HIV response) financing prospects conducted for UNAIDS and covering Sub-Saharan African countries, and has found this relatively straightforward(Hurley and Martin 2024). As emphasised in section 3.2, it would be preferable for all of the inequality and pandemic spending costs and financing reactions(including what are now seen as predictable regular pandemic shocks) to be included in one scenario, rather than separating them, but it may be that for technical or messaging reasons they need to be kept separate. 5. Overall Conclusions and Recommendations This paper has examined how the Debt Sustainability Analyses conducted by the IMF and World Bank could be adapted to take more account of the Sustainable Development Goals and Agenda 2030. It has looked first at the case for making such adaptations, made even more urgent by the need to help IMF and World Bank member states to prioritise national spending(and ensure that its financing does not compromise debt sustainability) in the post-COVID context of polycrises and limited global concessional funds. In particular, the paper emphasises the vulnerability of all countries across the world to two urgent crises: the climate emergency(and related risk of nature collapse and natural disasters); and extreme inequality and poverty, which have been worsened by COVID-19 and are undermining growth and stability. The paper next looked at existing efforts to include the SDGs in debt sustainability analysis. It found that this had been done successfully by many countries during the period of the Millennium Development Goals, in countries such as Bolivia, Burkina Faso, Rwanda and Tanzania, and that there already exist many building blocks which would make such inclusion possible(notably agreed estimated global costings and methodologies for country-specific costings for all of the SDGs). Nevertheless, it underlined that in the current“polycrisis” period, with only seven years to go to reach the SDGs, integrating the spending needed to reach all of the SDGs into the DSFs would simply produce a conclusion that virtually all countries would have unsustainable debt levels. As a result, it will be essential for countries to prioritise which SDGs they wish to include. The paper identifies the two greatest threats to debt sustainability(but also the greatest potential opportunities for growth and greater debt-carrying capacity if we combat them successfully) as being the dual crises of climate and nature emergency, and extreme inequality and poverty. It therefore recommends that the international community should focus on prioritising the adaptation of the DSFs to these issues, while ensuring that countries have flexibility to prioritise particular sectors within these areas. The paper next examined progress so far in adapting the SRDSF to forecasting long-term scenarios related to climate change adaptation and mitigation, as well as including natural disaster“stress tests” in its methodology to examine the impact of natural disasters in both the SRDSF and the LIC-DSF. Overall, these are major steps forward to including SDG13 in the DSA methodology, which provide a clear framework for how similar work could be done on other SDGs. However, the paper raises criticisms and suggests ways to improve this work in terms of broadening country coverage; calculating spending needs more accurately and(given the urgency of climate crisis action) including them in forecasts from year one of the projection; including the potential positive impact of just green transition spending; combining the multiple impacts of climate into one scenario; including the other environmental goals to prevent nature collapse; linking up the implications of climate-adapted DSFs to other IMF processes such as indicative spending floors and country lending eligibility; and giving more weight to climate module results in overall DSA risk assessments. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 13

Finally, the paper examined the lack of progress on adapting DSFs to the key types of spending which will confront the extreme inequality and poverty crisis. Currently the only mention of these types of spending is in an SRDSF module examining negative risks of ageing populations for higher health and social protection spending. However, the paper finds that it would be very easy to replicate what has been done in regard to integrating climate spending, for key anti-inequality spending. In particular, it finds that it would be easy to define the types of spending to include in an additional module or sub-modules, using an extended standard baseline scenario, and to define the country groups in which this should be done. It also finds that accepted estimated costs exist for the key sectors to be included, with which a template could be“pre-populated”. It would also be easy where necessary(i.e. where costs are very high) to include customised country-specific scenarios analysing costs and potential sustainable financing sources in more detail(for which clear and simple methodologies exist); and it would be feasible- and essential – to include the major positive impacts on growth which reducing inequality would have. Given that studies suggest a global pandemic is likely once every decade, it also suggests that a pandemic“stress test” should be included in both SRDSF and LIC-DSF. Table 1: Summary of current situation and reform proposals Type of Adaptation Current Situation Reform Proposals Climate/ Environment SRDSF: Adaptation and • Replicate SRDSF Climate Module in LIC-DSF Mitigation Costs in“Extended Review Baseline” • Include Nature Collapse Risk and Prevention Spend SRDSF and LIC-DSF: Natural • Combine all Climate Impacts Disaster Shock in“Stress • Include Positive Impact of Spending on Tests” Growth • Customised Scenario with Country Costs and Sustainable Financing • Give More Weight to Climate Module Results in Overall Risk Inequality/ Poverty No Current Adaptation Both SRDSF and LIC-DSF: • Anti-Inequality/ Poverty Module –“Extended Baseline” • Trigger High Net Gini(e.g. 0.4) • Include Positive Impacts of Spending on Growth • Customised Scenario with Country Costs and Sustainable Finance • Pandemic Shock Stress Test • Preferably Combine All Inequality/ Pandemic Impacts However, it is also vital not to forget two broader suggestions which are made in the paper: • There is no reason why the twin crises of climate and inequality should be considered separately. Ideally, climate and inequality would be considered and tackled together in all relevant countries so as to promote a“just green transition” across all countries and maximise the positive multiplier effects of the combined spending. This would mean that DSF modules on climate and inequality would be used simultaneously and their combined effects on sustainability, financing needs and growth shown. 17 17 Some interviewees suggested that to ensure that this happens naturally, the RSF and the PRGF could be merged: however, the different beneficiary countries, funding sources and purposes of the facilities might prevent this. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 14

• In the same way, other tools used by the IMF and World Bank could be made more consistent with the debt sustainability assessments: this applies notably to indicative spending floors in IMF programmes, where the definition of protected priority spending still revolves around social spending and should be expanded to include climate adaptation or mitigation or any efforts related to nature. Of all the suggestions made in this paper, the most important are to include in the scenarios the immediate and longer-term positive impacts and multiplier effects of spending on climate, nature and anti-inequality sectors on growth; and the impacts of increasing sources of finance which create no or highly concessional additional debt (tax revenue, grants, concessional loans and debt relief). It will also be highly important that baseline scenarios are made even more realistic than they are currently, taking into account the major negative effects on growth of growing climate and pandemic disaster events, as well as the permanent effects of more gradual climate impacts such as desertification/drought, sea warming, and those of inequality, on undermining growth(rather than, as currently, not mentioning them). These aspects cannot be seen as things which are“beyond the scope of a DSA” and therefore left to separate initiatives; nor can it be assumed that they are included in budgets or macroeconomic frameworks. The levels of spending in those documents almost always fall way short of the SDGs, and of ambitions governments have expressed in national development plans or election campaigns, thereby enhancing the growing cynicism of citizens across the world about politicians and democracy. They are vital in order to ensure that a revised DSA framework does not simply become a source of evidence reinforcing the view that climate and anti-inequality spending on the scale needed to reach the SDGs will dramatically worsen debt sustainability and increase debt risks: this is a major risk if the only impact of reforms is to add in large additional spending costs, and is likely to help prevent countries from reaching any of the SDGs or overcoming climate, nature, or inequality crises. Instead, the clear message emerging from revised DSAs with climate and inequality modules should be similar to that proven successfully during 2010-2015 with the Millennium Development Goals, that low- or no-cost financing plus major multiplier effects can allow us to reach the top priority SDGs and confront the climate, nature and inequality crises without provoking a widespread debt crisis. Their aim should be to assist countries in arguing the case for mobilising more concessional financing or debt relief and accelerating their efforts to collect progressive tax revenue. However, there is no getting away from the likely initial impact of including large extra amounts of spending (before positive multipliers, extra tax revenue and concessional flows kick in), which will be to increase debt ratios. To ensure that the SDGs are not sacrificed to keep debt levels sustainable, it will be essential to keep country debt burdens as low as possible and avoid what has happened over the last decade – a rapid rise in debt-service ratios so that SDG and climate spending is massively crowded out by debt service. To make this possible, and to accelerate the provision of debt relief where needed, much more emphasis must be placed in interpreting debt sustainability on the liquidity burden of debt service – and with just as much emphasis on external as domestic debt service – than has been in previous iterations of the DSF. This could be done by making debt service/budget revenue the primary risk indicator in interpreting DSA results. An even better way to sharpen the focus on this issue would be to add to the DSA framework an indicator showing the ratio of debt service to climate and/or anti-inequality spending(for which data exist, as Debt Service Watch shows), to measure the risk that high service is crowding out key SDG spending. The case for(and feasibility of) adapting the SRDSF and LIC-DSF more fully to country SDG needs and spending priorities is clear. It can build on the reforms already made to the SRDSF, and the suggestions made in this paper. The forthcoming review of the LIC-DSF(and updating of implementation guidelines for the SRDSF) should take a strong lead in this area, fundamentally reforming the LIC-DSF and the SRDSF, and thereby enhancing the contribution of the IMF and World Bank to Agenda 2030 and attaining the SDGs. Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 15

References AVAAZ(2023),‘Out of Oblivion: Why Biodiversity Must Be at the Heart of All IMF Policies’, April, (https://avaazimages.avaaz.org/Avaaz%202023%20-%20IMF%20spring%20meetings%20-%20biodiversity.pdf) Center for Global Development(2021),‘What’s Next? Predicting the Scale and Frequency of Future Pandemics, CGD Talks, Washington, July 20(https://www.cgdev.org/event/whats-next-predicting-frequency-and-scalefuture-pandemics) Development Finance International(2023a),‘The Worst Ever Global Debt Crisis: New Data from Debt Service Watch’, October(https://www.development-finance.org/en/news/831-11-october-the-worst-debt-crisis-evershocking-new-debt-service-numbers) Development Finance International(2023b),‘The Worst Ever Global Debt Crisis: Putting Climate Adaptation Spending Out of Reach’ November,(https://www.development-finance.org/en/news/832-debt-service-watch) Development Finance International(2016),‘Ensuring Debt Sustainability In Developing Countries through Productive Expenditure’, report by DFI to UK Department for International Development, July, mimeo. Gaceta Oficial(2007),‘Plan Nacional de Desarollo: Bolivia Digna, Soberana, Productiva y Democrática para Vivir Bien, 2006 – 2011’, Decreto Supremo 29272, (https://www.conicyt.cl/documentos/art_eliana28oct2010/BoliviaPlan_Desarrollo_Nac_2016_2011_cap%20.pdf) Gallagher, K.P., Bhandary, R.R., Ray, R. and Ramos, L.(2023),‘Reforming Bretton Woods Institutions to Achieve Climate Change and Development Goals’, One Earth, Volume 6:10 (https://www.sciencedirect.com/science/article/pii/S2590332223004487#bib63) Hurley, G. and Martin, M.(2024, forthcoming),‘Financing the Ending of HIV as a Public Health Threat by 2030, reports on Eastern and Southern, and West and Central Africa, UNAIDS, Geneva. International Labour Organisation(2024),‘Financing Gap for Universal Social Protection’, Working Paper 113, April, https://www.ilo.org/publications/financing-gap-universal-social-protection-global-regional-and-national International Monetary Fund(2024a),‘Climate-Driven INFORM Risk’, database accessed on 15 May, (https://climatedata.imf.org/pages/adaptation#ad2) International Monetary Fund(2024b),‘IMF DIGNAD Toolkit’, accessed on 15 May, (https://climatedata.imf.org/pages/dignad) International Monetary Fund(2024c),‘Public Investment Management Assessment: Home Page’, accessed on 15 May,(https://infrastructuregovern.imf.org/content/PIMA/Home/PimaTool/What-is-PIMA.html) International Monetary Fund(2023a),‘The IMF and the Sustainable Development Goals’, (https://www.imf.org/en/About/Factsheets/Sheets/2023/IMF-Sustainable-development-goals-SDGs) International Monetary Fund(2023b),‘Trinidad and Tobago: 2023 Article IV Consultation-Press Release; and Staff Report’, May, (https://www.imf.org/en/Publications/CR/Issues/2023/05/04/Trinidad-and-Tobago-2023-Article-IV-ConsultationPress-Release-and-Staff-Report-533070) Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 16

International Monetary Fund(2023b),‘How to Assess Spending Needs of the Sustainable Development Goals: The Third Edition of the IMF SDG Costing Tool’, IMF How To Note 2023/005, December, (https://www.imf.org/en/Publications/imf-how-to-notes/Issues/2023/12/14/How-To-Assess-Spending-Needs-ofthe-Sustainable-Development-Goals-The-Third-Edition-of-the-541463) International Monetary Fund(2022a),‘Macro-Fiscal Implications of Adaptation to Climate Change’, Staff Climate Note 2022/002, March,(https://www.imf.org/en/Publications/staff-climate-notes/Issues/2022/03/16/MacroFiscal-Implications-of-Adaptation-to-Climate-Change-512769) International Monetary Fund(2022b),‘Staff Guidance Note on the Sovereign Risk and Debt Sustainability Framework for Market Access Countries’ August, (https://www.imf.org/en/Publications/Policy-Papers/Issues/2022/08/08/Staff-Guidance-Note-on-the-SovereignRisk-and-Debt-Sustainability-Framework-for-Market-521884) International Monetary Fund(2021a),‘Building Back Better: How Big Are Green Spending Multipliers?’ IMF Working Paper 21/87, March(https://www.imf.org/en/Publications/WP/Issues/2021/03/19/Building-Back-BetterHow-Big-Are-Green-Spending-Multipliers-50264) International Monetary Fund(2021b),‘A Post-Pandemic Assessment of the Sustainable Development Goals’, Staff Discussion Note 2021/003, April(https://www.imf.org/en/Publications/Staff-DiscussionNotes/Issues/2021/04/27/A-Post-Pandemic-Assessment-of-the-Sustainable-Development-Goals-460076) International Monetary Fund(2019),‘Fiscal Policy and Development: Human, Social, and Physical Investments for the SDGs’, Staff Discussion Note 2019/003 (https://www.imf.org/en/Publications/Staff-Discussion-Notes/Issues/2019/01/18/Fiscal-Policy-and-DevelopmentHuman-Social-and-Physical-Investments-for-the-SDGs-46444) International Monetary Fund(2017),’Inequality Overhang’, IMF Working Paper 2017/076, March (https://www.imf.org/en/Publications/WP/Issues/2017/03/28/Inequality-Overhang-44774) International Monetary Fund(2016),’Small States’ Resilience to Natural Disasters and Climate Change: Role for the IMF’, IMF Policy Paper, December,(https://www.imf.org/external/np/pp/eng/2016/110416.pdf) International Monetary Fund(2014),‘Redistribution, Inequality and Growth’, Staff Discussion Note 14/02, February,(https://www.imf.org/external/pubs/ft/sdn/2014/sdn1402.pdf) International Monetary Fund(2012),‘Public Investment, Growth, and Debt Sustainability: Putting Together the Pieces’ IMF Working Paper 12/144, June,(https://www.imf.org/external/pubs/ft/wp/2012/wp12144.pdf) Kraemer, M. and Volz, U.(2022),‘Integrating Nature into Debt Sustainability Analysis’, SOAS Centre for Sustainable Finance/Finance for Biodiversity Initiative, April,(eprints.soas.ac.uk/37715/1/NatureDSA.pdf) Maldonado, F. and Gallagher, K.P.(2022),‘Climate Change and IMF Debt Sustainability Analysis’, Task Force on Climate, Development and the IMF Working Paper, April, (https://www.bu.edu/gdp/files/2022/02/TF_WP_005_FIN.pdf) Marani, M., Katul, G.G., Pan, W.K. and Parolari, A.J.,(2022),‘Intensity and frequency of extreme novel epidemics’, Proceedings of the National Academy of Sciences, Volume 118: 35, (https://www.pnas.org/doi/10.1073/pnas.2105482118) Martin, M. and Kripke, G.(2023),‘A Call to Action to Save SDG10’, DFI/Oxfam/Pathfinders/UNAIDS Policy Briefing, June,(https://www.development-finance.org/files/SDG10_policy_brief_English.pdf) Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 17

Oxfam(2023),‘IMF Social Spending Floors: a Fig Leaf for Austerity?’, Briefing Paper, April,(https://policypractice.oxfam.org/resources/imf-social-spending-floors-a-fig-leaf-for-austerity-621495/) Sachs, J.D., and Schmidt-Traub, G.(2014)‘Financing for Sustainable Development: Implementing the SDGs through Effective Investment Strategies and Partnerships’. New York: Sustainable Development Solutions Network(SDSN),(https://www.semanticscholar.org/paper/Financing-Sustainable-Development%3AImplementing-the-Schmidt-Traub-Sachs/ca9af2d591bdc06d855abbdfa874a39964f8bdd2) United National Conference on Trade and Development(2019),‘Financing a Global Green New Deal: Trade and Development Report 2019’, chapter 4,(https://unctad.org/system/files/official-document/tdr2019ch4_en.pdf) World Bank(2024),‘The World Bank and The 2023 Agenda’,(https://www.worldbank.org/en/programs/sdgs2030-agenda) World Bank(2022),‘Climate Change Public Expenditure and Institutional Review Sourcebook’, (http://documents.worldbank.org/curated/en/700291662048371012/Climate-change-public-expenditure-andinstitutional-review-sourcebook) World Bank(2021),‘The Economic Case for Nature: A Global Earth-Economy Model to Assess Development Policy Pathways’, June,(https://openknowledge.worldbank.org/entities/publication/fcc11682-c752-51c4-a59f0ab5cd40dc6f) World Bank(2018),‘Growth, inequality and Poverty: a Robust Relationship?’, Policy Research Working Paper 8578,(https://documents1.worldbank.org/curated/en/248621536264773958/pdf/WPS8578.pdf) Interviews ⁻ African Network on Debt and Development(AFRODAD) ⁻ Boston University ⁻ Debt Justice UK ⁻ Erlassjahr ⁻ European Network on Debt and Development(EURODAD) ⁻ International Monetary Fund(IMF) ⁻ Latin American Network on Debt and Development(LATINDADD) ⁻ London University School of Oriental and African Studies ⁻ United Nations Department of Economic and Social Affairs(UN DESA) ⁻ United Nations Conference on Trade and Development(UNCTAD) ⁻ World Bank Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 18

About the author Matthew Martin is Director of Development Finance International(DFI, www.development-finance.org), a non-profit consultancy helping countries of the Global South to mobilise more top-quality financing (including debt relief) to spend on reducing inequality and poverty. Since 1994, DFI has helped more than 80 countries, as well as civil society organisations (CSOs), parliaments, trade unions, international organisations and donor governments – through capacity-building, analysis, research and advocacy support. Prior to establishing DFI, he worked at Oxford University, the World Bank and the Overseas Development Institute. Acknowledgments This paper draws on debt service and SDG spending data prepared by David Waddock, Maria Holloway, Jeannette Laouadi and Myriam Sallah, and DFI's work building DSA capacity in 32 countries funded by Austria, Canada, Denmark, Germany, Ireland, Sweden, Switzerland and the United Kingdom. The author is also grateful for valuable inputs and comments from institutions interviewed listed above, as well as from participants at an FES/IPD/JNUSA seminar in September 2023 and an IMF/World Bank seminar in April 2024. New York Office The office of Friedrich-Ebert-Stiftung(FES) in New York serves as the liaison for FES offices worldwide with the United Nations(UN) in New York and the international financial institutions(International Monetary Fund and World Bank) in Washington, D.C. The office addresses peace, justice and economic issues, working closely with academia, civil society, multilateral institutions and their Member State governments to convene multi-stakeholder debates. The formats of our work include international conferences, expert workshops and high-level meetings involving government representatives as well as published policy briefs and analytical studies. Our overarching mission is to bring a special focus to the perspectives of trade unions, women, and developing and emerging-market countries in an effort to forge consensus toward multilateral solutions to international and global challenges. Jubilee USA Network Jubilee USA Network is an interfaith alliance of 75 religious, labor, development, and climate organizations and more than 750 faith groups from the United States and around the globe. Jubilee USA addresses the structural causes of poverty, inequality and environmental challenges in our communities and countries. The network's efforts build an economy that serves, protects, and promotes the participation of the most vulnerable. Imprint Publisher: Friedrich-Ebert-Stiftung e. V.| Godesberger Allee 149 | 53175 Bonn| Germany Email: info@fes.de Register no.: VR2392 Bonn Register of Associations Bonn Local Court Issuing Department: Division for International Cooperation| Global and European Policy Responsibility for content: Dr. Michael Bröning| Executive Director| Friedrich-Ebert-Stiftung| New York Office| 747 Third Avenue, Suite 34D| New York, NY 10017| USA Phone+1-212-687-0208 https://ny.fes.de Contact/Order: Christiane.Heun@fes.de The views expressed in this publication are not necessarily those of the Friedrich-Ebert-Stiftung. Commercial use of media published by the FriedrichEbert-Stiftung(FES) is not permitted without the written consent of the FES. Publications by the Friedrich-Ebert-Stiftung may not be used for electioneering purposes. © 2024 www.fes.de/bibliothek/fes-publikationen Series: Debt Sustainability Assessments& Their Role in the International Financial Architecture 19