Summary

Assessing the sustainability of developing countries’ debt is an essential task of the International Monetary Fund (IMF) and the World Bank, but it is also a controversial one. The two institutions are currently reviewing the debt sustainability framework for low-income countries (abbreviated as LIC-DSF). This paper provides a primer on the framework’s basic elements, an overview of critiques from a variety of perspectives, and analysis of selected issues, including the endogeneity of analysis of sustainability, analysis of debtors’ institutional strength, analysis of domestic political conditions, role and exercise of staff judgment, and collaboration between the World Bank and the IMF. The paper recommends eight improvements, including streamlining the institutional indicators, better separating economic analysis from political judgment, adopting certain procedures for applying judgment, and selectively introducing climate considerations.

Introduction

International financial assistance to emerging-market and developing countries requires analysis of the extent to which borrowing countries’ sovereign debts are financially sustainable. In this kind of analysis, carried out by the International Monetary Fund (IMF) and the World Bank, debt sustainability is defined as the debt’s ability to be serviced by the government without policy reforms that destabilize the country.¹ The Debt Sustainability Analyses (DSAs) conducted by the IMF and the World Bank impact a broad range of stakeholders—the debtor country, its people and institutions, and official and private creditors—through their consequences for the design of lending programs and, when necessary, debt restructuring. Because they sometimes force actors to recognize financial losses and distribute them, DSAs and the institutional procedures and analytical frameworks by which they are generated can be highly controversial.

The IMF uses two frameworks for this analysis: the Sovereign Risk and Debt Sustainability Framework for Market Access Countries (SRDSF), which applies to advanced and emerging-market countries, and the Low-Income Country Debt Sustainability Framework (LIC-DSF). The IMF shares the latter with the World Bank, whose principal mandate is these countries’ economic and institutional development. The success of sovereign debt restructuring, lending programs, and other major financial decisions of the international community depends on the usability and reliability of these frameworks.

Successfully predicting financial crises and specifying a maximum level of debt or debt service that a government can support is analytically challenging, however. Some countries with relatively low levels of debt nonetheless default, while others with high debt can service it seemingly without difficulty in perpetuity. DSAs therefore raise false alarms (type I errors) and miss crises (type II errors), each in about one-third of cases. Knowing this, the IMF and the World Bank reserve room in their DSA procedures for what they call staff judgment, which allows for the introduction of financial, institutional, and political considerations that are not explicitly modeled in their frameworks. But use of judgment is often opaque, and the frameworks, as operationalized, have generated conclusions that are sometimes inconsistent across country cases.

Such conclusions have fueled concern that DSAs’ findings—that debts are sustainable or unsustainable—could be influenced by a desire to justify (or undermine the case for) emergency lending or debt relief. Even commentators who are sympathetic to the Bretton Woods institutions, including former officials of these organizations, observe that DSAs are sometimes based on economic assumptions that, while not arbitrary, are at least uncertain and challengeable. A particular point of concern is whether DSAs and, in the case of emerging-market countries, the IMF’s Exceptional Access Policy (EAP) have weakened over time in their ability to counteract pressure from members to lend without imposing a restructuring on creditors. The distribution of tens of billions of dollars hinges on the choice of assumptions that underpin such findings. 

Notwithstanding misgivings, member states, including those in the Group of Seven (G7) and the Group of 20 (G20), official and private creditors, and the borrowers themselves vest responsibility for developing DSAs for low-income countries in the IMF in collaboration with the World Bank. The reasons behind this grant of authority are several and explored later in this paper. Suffice it to say at the outset that improving outcomes of debt crises and debt restructuring for sovereign borrowers and their creditors alike requires strengthening the debt sustainability framework of the Bretton Woods twins.

This paper diagnoses the strengths and weaknesses of the LIC-DSF and offers recommendations for improving it. The paper is written as the IMF and the World Bank are conducting a review of the framework, which is expected to be completed and published in mid-2026.² It offers a primer on the basic elements of the LIC-DSF, an inventory and review of critiques of it (and related elements of the SRDSF) from different perspectives, analysis of selected aspects of institutional dynamics and political economy, and an eight-point agenda for reform.

This paper focuses largely, albeit not exclusively, on the qualitative and institutional aspects of the LIC-DSF and the DSAs it generates, drawing on instructive country cases. It does not delve into recommendations of a more technical nature, such as changes to the formal model and the thresholds defining sustainability, but instead highlights these in other papers.³ It also leaves aside broader questions about the design of development financing, balance-of-payments lending, and the institutional arrangements for sovereign debt restructuring.

The paper recommends that the current review of the LIC-DSF should: 1) eliminate threshold effects at the boundary between low- and middle-income country categories; 2) reconsider the analysis of institutional indicators; 3) explicitly and more cleanly separate economic analysis from political judgment; 4) experiment more creatively with political risk analysis; 5) adopt certain procedures for applying staff judgment; and (6) selectively introduce climate risk into DSAs, among other things. The paper emphasizes that political analysis is an inescapable feature of debt sustainability analysis and offers suggestions for conducting it without being politically partisan. Recommendations here that cannot be addressed in the current review should be taken up in subsequent reviews. Several of them apply as well to the upcoming review of the SRDSF.

The Centrality of the IMF and the World Bank

Discourse over the plight of highly indebted countries has raised questions about which institutions should develop DSAs and the division of labor among them.

Several organizations have developed alternative debt sustainability frameworks and conduct parallel assessments of debt sustainability. These include UN Trade and Development (UNCTAD), the Asian Development Bank, the African Development Bank, the Inter-American Development Bank, and the European Union.⁴ The European Commission, the European Stability Mechanism, and the European Central Bank collaborate in producing DSAs for post-program monitoring of the countries that had been ensnared in the euro crisis.⁵ China’s Ministry of Finance has developed a template for analyzing debt sustainability for countries in the Belt and Road Initiative.⁶ Martín Guzmán and Joseph E. Stiglitz (2024b, 15) suggest that Chinese authorities may be quietly conducting their own DSAs on countries undergoing restructuring. BlackRock, a private holder of distressed bonds, has reportedly done something similar, and nongovernmental organizations and research institutions have offered alternative frameworks (discussed below).

National governments have conducted such DSAs as well. A market-access country, Argentina, conducted its own in 2020, with technical assistance from the IMF as the prelude to its private debt restructuring of that year.⁷ Zambia conducted its own DSA in July 2023 during its recent debt restructuring.⁸ Development of alternatives could be potentially important in future contingencies.

Nonetheless, while alternative analyses might inform the position of individual creditors in debt restructuring negotiations and their arguments with IMF staff, the IMF and the World Bank continue to own debt sustainability analysis and the underlying analytical framework for low-income countries—with the IMF in the lead relative to the World Bank. Alternative DSAs are supplements rather than substitutes for the IMF–World Bank framework and remain peripheral to decisions on international financing, early warning, and debt restructuring. With the exception of arrangements in the euro area,⁹ there is no prospect of a consensus on entrusting any other institution with conducting country-specific DSAs to guide collective decisions on debt restructuring or financial assistance.

The focal position of the Bretton Woods institutions is underpinned, among other things, by their global reach, advantages of scope and scale, access to information, and relative autonomy of staff. Among the G7 and G20, creditor and debtor states alike generally agree that DSAs should be grounded in financial analysis and that the IMF and the World Bank are best equipped in terms of expertise and bureaucratic organization to provide it. Even though they too are creditors, the Fund and Bank are regarded as more objective than other actors involved in debt restructuring. If the Bretton Woods institutions harbor their own agendas, those arguably align with a widely endorsed mandate (such as financial stabilization and development). Expertise is an important, but not the only important, source of authority of the Fund and Bank.

Moreover, the Bretton Woods twins are legally chartered global international organizations. Notwithstanding misgivings about the distribution of voting shares within their governing arrangements, the decisions and policies of their governing and executive boards represent a common position of a diverse, universal membership that includes debtor and creditor states alike. These qualities are not shared by other institutions in the debt restructuring space, such as the G20, the Paris Club, or creditor representative groups. While the United Nations organizations share universal membership and formal charters, the Fund and Bank have the distinctive advantage of lending capacity. The combination of program financing with expertise under a broad political mandate uniquely empowers the Fund and Bank to carry out debt sustainability analysis.

The combination of program financing with expertise under a broad political mandate uniquely empowers the Fund and Bank to carry out debt sustainability analysis.

DSAs play a particular role in the political economy of emergency lending within the IMF. They serve as a bulwark against pressure from members to lend into an unsustainable situation, thereby indulging moral hazard and jeopardizing Fund resources and, ultimately, its preferred creditor status. Such pressure comes not only from official and private creditors but also the debtor, whose finance minister’s job could be on the line.¹⁰ Observers fear that, given tensions over the analyses that prevail among staff, management, and the board, the wiggle room provided by the discretionary elements of the DSA process are used to weaken findings of sustainability.¹¹

The Current Debt Sustainability Framework for Low-Income Countries

To impose some degree of consistency on analysis of debt sustainability, the IMF and the World Bank established the LIC-DSF. This framework guides analysis when, for example, staff designs lending programs and proposes them to the IMF and the World Bank Executive Boards. In the context of a sovereign debt restructuring, the analysis determines the amount of debt reduction that the creditors must concede to restore sustainability and thus unlock institutional financing. A primer on the LIC-DSF is provided in the IMF’s most recent review of the framework in 2018.¹²

At the core of the framework lies a set of probit regression models that estimate the probability of debt distress and default while controlling for debt indicators, the strength of institutions and policies, and growth. These models are used to identify the thresholds for four debt indicators, above which the risk of debt distress is considered unacceptably high. Of the four indicators, there are two stock indicators and two flow indicators. The stock indicators are the present value (PV) of public and publicly guaranteed (PPG) external debt to gross domestic product (GDP) and to exports. The flow indicators are the ratios of service on PPG external debt to exports and to fiscal revenue. In addition, the LIC-DSF considers the PV of total PPG-debt-to-GDP, that is, domestic and external debt combined. The thresholds are chosen to minimize both false alarms (type I errors) and missed crises (type II errors).

Using an indicator of institutional strength, which is referred to as the Composite Indicator,¹³ low-income countries are grouped into “weak,” “medium,” or “strong” categories that reflect their capacity to carry debt. The acceptable threshold for each of the four indicators varies according to the debt-carrying capacity: for a country whose debt-carrying capacity is rated as strong, the external-debt-service-to-exports threshold is 21 percent, while it is 15 percent for a country whose capacity is rated as medium and 10 percent for a weak-capacity country.¹⁴ With respect to the ratio of the present value of total public-debt-to-GDP, countries rated as strong are expected to sustain a ratio of 70 percent, moderate-rated countries 55 percent, and weak-rated 35 percent. At the end of the process, staff issues two separate findings, one on the risk of distress and the other with respect to sustainability.

The Executive Boards, management, and staff of both the IMF and the World Bank are understandably uncomfortable relying solely on the so-called mechanical signals generated by the framework template. Staff can temper these risk signals with their own judgment when factors that are missing from the model could be important to determining sustainability, as occurs in about 20 percent of low-income country cases.¹⁵ When doing so, staff are mandated to be explicit and specific about why and how judgment is rendered.

The succession of DSAs across many countries over the past two decades, however, has puzzled analysts and raised nagging questions about the administration of the frameworks. Although the most salient cases have involved emerging-market countries, such as Argentina, such questions have also been raised in low-income cases as well. Given the inherent difficulty of predicting distress, a certain number of false positives and negatives should be anticipated. But the issues raised by the case history are not only apparent in hindsight: findings of (un)sustainability were debated at the time in many cases, with some analysts, even inside the institutions, taking the contrary position to the finding in the published DSA.¹⁶ Such cases have generated debate outside and soul searching within the institutions.

Critiques and Debates

Commentary on and critique of the LIC-DSF is offered from several different perspectives, which are reviewed in this section. Because such critiques sometimes overlap with assessments of the SRDSF, this review considers some of contributions that apply to the SRDSF that also have implications for the LIC-DSF.

Experts and Insiders

During the post-pandemic restructurings, the LIC-DSF and the SRDSF came under intense scrutiny and often faced criticism. Brad Setser (2024) questions whether the DSAs for several countries were too optimistic and therefore let creditors off the hook too easily in restructuring negotiations. Setser particularly criticizes the DSA for Sri Lanka, which was conducted under the SRDSF, for its exclusive focus on total public debt at the expense of considering external debt and debt service separately. Setser reflects that the country possesses narrow tax revenue and export bases, which along with scarce foreign exchange reserves arguably constrain its effective debt-carrying capacity relative to countries such as Ghana, even though the latter is listed as low-income and takes the LIC-DSF. The restructuring target for Sri Lanka is a ratio of public-debt-to-GDP of 110 percent, while it is 55 percent for Ghana, of which external debt should be no more than 40 percent in 2028. As Setser sums up, “Across a host of metrics, Sri Lanka is judged able to support about twice as much debt as Ghana because it was assessed according to the market-access country debt sustainability model. . . . The IMF’s targets . . . leave Sri Lanka exposed to a high risk of a future debt crisis.”

Indermit Gill and Brian Pinto (2023 and 2024) observe that, two decades ago, there was a rationale for creating a debt analysis framework for low-income countries that was different from the framework for emerging-market countries with access to international financial markets and advanced countries that mostly borrow in their own currencies. But, by the mid-2010s, low-income countries could increasingly borrow domestically and abroad in foreign currency, making private interest rates the marginal cost of capital for sovereign low-income borrowers. Under current circumstances, Gill and Pinto (2023) argue, external and domestic debt are intertwined and derive from basic expenditure and revenue decisions together. Forecasting debt distress and sustainability therefore calls for unified analysis (as conducted for market-access countries) of total public debt and its dynamics over time as captured in the standard debt formula.¹⁷

If, using this formula, debt is projected to rise and remain above a certain, specified threshold, it can be deemed unsustainable—with all the consequences for preemptive corrective action, financial assistance, and debt restructuring that follow. For this calculation, the chosen interest rate should be an amalgamation of international and domestic interest rates, weighted according to the share of external and domestic liabilities in the sovereign’s debt structure. Recognizing that servicing external debt is subject to the foreign exchange constraint, however, Gill and Pinto (2023) would retain analysis of servicing capacity in light of reserve adequacy.

Setser (2025), on the other hand, would retain the specific consideration of external public debt in the LIC-DSF. For its part, the IMF seems to be moving in the opposite direction in the current review. He provides three reasons for retaining it: 1) payments on external debt drain the national economy in a way that those on domestic debt do not; 2) foreign exchange constraints remain binding in many cases (for instance, Sri Lanka and Zambia defaulted because they exhausted reserves); and 3) debt restructuring focuses primarily on, and the concept of comparability of treatment applies exclusively to, external debt.

Clemens G. Von Luckner, Sebastian Horn, Aart Kraay, and Rita Ramalho (2024), who are all affiliated with the World Bank, examine alternative methods for predicting debt distress for low-income countries. They test a very large number of alternative econometric models to identify the best predictors of distress.¹⁸ They find that sophistication in modeling does not buy predictive power. A very simple model—based on institutional strength, reserves as a share of imports, and especially external debt service as a share of exports—performs just as well as or better than models with a larger number of explanatory variables and about as well as the existing LIC-DSF. Because it is more parsimonious and thus more transparent, they contend, the simpler three-variable model should be preferred.

Von Luckner and coauthors confirm that external and domestic distress are almost perfectly correlated and agree in principle that total public debt will drive crises. But they find that adding domestic debt to external debt–based models provides little predictive value, although this finding could change as domestic debt rises. These authors caution that even the best performing models fail to predict distress in roughly one-third of cases in which it does in fact occur and raises a false alarm in a similar proportion of cases in which it does not. Decisions based on these models should incorporate additional, relevant information and not dispense with judgment.

Antoine Gaudin, Brendan Harnoys-Vannier, and Martin Kessler (2024) identify an upward bias in DSA forecasts of the total debt ratio of more than 10 percent of GDP on average for low-income countries at the five-year time horizon. Overestimation is due to slippage on fiscal reform, especially revenue raising; excessive optimism with respect to growth; and neglect of country-specific factors. Their study supports earlier findings of “optimism bias” but shows that it applies mainly to larger low-income countries.

Private Sector

Objections to the debt sustainability analyses and frameworks from the private sector are motivated by a variety of considerations, such as the desire to minimize loss, maintain equity among creditor classes, sustain the liquidity of their investments, and participate in new investment and transaction opportunities when the country concerned re-accesses markets after stabilization. The private sector is a broad category that covers a variety of views, but stakeholders’ main grievances tend to fall into five categories.¹⁹

First, whereas DSAs for low-income countries are generally biased upward, those specifically for countries that are undergoing debt restructuring are instead biased downward. Pijus Virketis (2024), for example, finds that the DSAs in IMF programs for countries undergoing debt restructuring are pessimistic by a wide margin, a consequence of forecast errors on both the local currency GDP and the exchange rate. Such pessimism increases the sacrifice that all creditors, private and official, must endure in a restructuring, though it expands breathing room for the debtor.

Second, when private analysts identify assumptions that they suspect are unrealistic (usually too pessimistic), IMF country teams sometimes resist taking this feedback on board. Private analysts were particularly unhappy with the refusal to revise the DSA forecast assumptions for the local currency GDP and the exchange rate in the case of Sri Lanka, which led to the adoption of macro-linked bonds to give holders the benefit of an upside scenario.²⁰ The case contrasts with Ghana, where the government resisted bondholders’ proposals to include a value recovery instrument.

Relatedly, private sector investors and analysts generally accept the need for a buffer between the debt reduction and the threshold of sustainability, so that once the restructuring is accomplished, the country is not positioned on the knife edge, where the risk of yet another default remains substantial. But, in their view, such a buffer should not be excessive, and the DSA should not be premised on unrealistic assumptions in order to generate it.

Third, the private sector and official bilateral creditors (OBCs) differ, apparently irreconcilably, on the discount rate to apply to present value calculations about debt. The low discount rate used by OBCs and international financial institutions (5 percent) undervalues the contribution that the private sector makes to debt restructuring, as their cost of funds and rate of return on alternative investments is much higher—in double digits. Finding a way to bridge this fundamental difference remains an analytical challenge for debt restructuring.²¹

Fourth, private sector analysts, including the financial advisers for the bondholders and the banks involved in restructuring, want transparency into the DSAs earlier and with greater granularity than is currently the case. They want to be able to weigh in before the DSA is set in stone, and they want the benefit of the DSA in early discussion with the debtor, including with respect to comparability of treatment vis-à-vis official creditors.²²

Nonetheless, fifth, the private sector generally accepts the need for a centrally generated DSA that establishes the envelope for restructuring the debt—which they understand is necessary to address the collective action problem among creditors. Notwithstanding criticisms, investors and advisers in the private sector are generally favorable toward the IMF’s central role in this regard, as the most expert and objective institution. Indeed, in a restructuring scenario, it does not seem likely that private investors would emerge better off from reassigning the DSA task to an alternative institution.

Non–Paris Club Creditors

Official creditors that are relatively new to debt sustainability analysis and restructuring have a relatively distinct set of criticisms of the frameworks. Foremost among the non-traditional creditors, China tends to take the view that, in debt restructuring cases, DSAs are too quick to pull the trigger on an unsustainable designation. Chinese officials often argue that countries are instead sustainable if a longer time horizon is adopted. In advancing these arguments, Chinese commentators draw from domestic experience, in which the country successfully grew out of debt overhangs during its decades of rapid development. They argue low-income countries (and middle-income countries) that adopt a similar model should be able to grow out of debt problems with sufficiently long bridge financing, occasional debt service relief, and maturity extensions, but without face-value reductions to principal.²³

Foremost among the non-traditional creditors, China tends to take the view that, in debt restructuring cases, DSAs are too quick to pull the trigger on an unsustainable designation.

During the post-pandemic period, these objections were voiced during arguments over the debt restructurings within and outside the Common Framework.²⁴ Chinese officials entered the process in a defensive posture, concerned that a disproportionate share of the financial losses would be foisted upon their lending agencies. They were principally responsible for inserting a phrase in the original G20 statement on the Common Framework that reserved a possibility for “official creditors’ collective assessment,” along with the DSA, to determine the need for a debt treatment and the size of the restructuring envelope.²⁵ As it has turned out in practice, however, only the DSA has guided the debt treatment; collective assessment has never been applied.

The Chinese perspective tends to offer three further critiques. First, analysts with this perspective contend that debt analysis examines only the liability side of the sovereign balance sheet and ignores the asset side; taking sovereign assets into consideration would substantially improve prospects for sustainability in many cases.²⁶ Some state assets might be illiquid but could be harnessed to debt servicing if given sufficient time for economic development. Second, they propose a distinction between “productive debt,” which finances investment and growth, and “consumptive debt,” which finances consumption and fiscal deficits. Differentiating between the two would provide a better picture of sustainability, they argue.²⁷ Moreover, in the context of a debt restructuring, consumptive creditors should be required to provide greater debt relief than productive ones. Finally, the restructuring perimeter is generally drawn too broadly and captures debt that is not public or publicly guaranteed.²⁸ Chinese officials argue that requiring restructuring of the debts of non-guaranteed state-owned enterprises that are commercially viable disincentivizes investment unnecessarily. Central bank swaps should also be excluded entirely from restructuring.²⁹ The scope of the DSA should thus be circumscribed accordingly. The development community and some UN agencies share many of these critiques (see more in the next section).

With respect to procedure, in restructuring cases, rising creditor countries chafe at having to provide financing assurances for an IMF program before having important information in the DSA and thus knowing the amount of debt relief required. Such assurances are required before the program documents are circulated to the IMF’s executive directors for approval in the board meeting. Historically, Paris Club creditors accepted this arrangement because they could block approval during the board meeting if the amount of relief required in the DSA turned out to be unacceptable. But, with smaller shares of votes in the board, the non–Paris Club countries are less confident of being able to do the same. Chinese officials, among others, advocate early sharing of the DSA with the official creditors and more collaboration with them over the course of the DSA process.³⁰

Development Community

Guzmán and Stiglitz (2024) offer several trenchant critiques. Although informed partly by the experience with Argentina’s 2020 restructuring, their criticisms apply to low-income countries as well. They say DSAs, as they have been conducted, contribute to rather than ameliorate the so-called too little, too late syndrome in debt restructuring. They reiterate the earlier, widely shared criticism by Guzmán and Heymann (2015) that DSAs lack realism with respect to the effects of fiscal austerity on GDP and repayment capacity. Whereas a primary fiscal surplus facilitates repayment, the shrinkage of the GDP denominator raises the debt and service ratios.³¹ So, first, this negative effect must be more forthrightly acknowledged and accurately modeled.

Second, in the context of a debt restructuring that is intended to restore sustainability, they say the DSA should use a discount factor that is close to the risk-free interest rate, whereas the IMF and the World Bank have been using interest rates that are premised on substantial risk of default afterward. Third, pooling domestic and external debt in analysis, and in the measure of gross financing needs in particular, is problematic owing to marked differences in, for example, rollover capacity of the two types of liabilities. They warn that the IMF’s treatment of domestic debt affects how capital markets will develop in low-income countries.

Fourth, Guzmán and Stiglitz also advocate the publication of the DSA at an early point, before Executive Board consideration, to frame: 1) public discourse within the borrower country, 2) creditors’ expectations, and 3) the first stage of negotiations among creditors and between them and the borrower. They suspect that doing so would weaken the influence of creditors over borrowers on the margin compared to delayed disclosure.

UNCTAD has nurtured a parallel, alternative approach that contains explicit recommendations for DSA reform.³² The organization is motivated to advance the sustainable development goals (SDGs) and the broad social and humanitarian objectives underlying them. While these are generally shared by the Bretton Woods institutions, emphasis and operationalization differ substantially. UNCTAD’s framework is more broadly encompassing in terms of the financial resources for development and the time horizon over which sustainability is assessed. Provided enough resources are brought to bear on investment, a country that breaches thresholds in the medium term according to the IMF might instead be sustainable under UNCTAD’s long perspective. Growth generated by such investment relaxes fiscal constraints on sustainability in ways that could be better integrated into the LIC-DSF.

Climate Community

A yet more profound, perhaps radical, critique has risen from the community of scholars and activists devoted to climate change mitigation and adaptation and environmental protection. Their arguments are premised on the conviction that the SDGs and the decarbonization targets of the Paris Agreement should be taken seriously, as binding commitments prior to those to creditors, in the analysis and restructuring of sovereign debt.

Members of the climate community argue, simply put, that climate change both places demands upon sovereigns to invest in adaptation and mitigation that must be prioritized and severely constrains their debt-carrying capacity. The analytical frameworks used by the World Bank and the IMF, they add, completely miss the changing climate’s threat to sustainability. Many otherwise sustainable countries would not be so rated if the analysis incorporated vulnerability to extreme weather events, their impact on output, growth, and exports, and the stranding of carbon-emitting assets as consequences of mitigation policies.

An exceptional country might benefit economically from climate change, but the effects are likely to be overwhelmingly negative for the large majority of low-income countries and emerging markets. Because traditional analytical frameworks virtually ignore climate risk, they overestimate debt sustainability by a substantial margin.

Because traditional analytical frameworks virtually ignore climate risk, they overestimate debt sustainability by a substantial margin.

Sri Lanka’s debt, for example, is deemed to be sustainable after its current restructuring, with a total debt-to-GDP ratio of 95 percent and a foreign-exchange-debt-service-to-GDP ratio of 4.5 percent during 2027–2032. With eleven tropical cyclones having hit the country during 2000–2020, though, it was quite likely to be struck by another over the course of its IMF program and repayment period. But officials in Colombo would not countenance the complexities that climate considerations would have injected into their debt restructuring negotiations during 2023 and 2024, which were already fraught and prolonged.³³ Cyclone Ditwah, the country’s deadliest natural disaster in more than twenty years, struck Sri Lanka in late November 2025, required a recalibration of the IMF program, and prompted calls to renegotiate the restructuring agreement.

It is fair to say that there is broad sympathy for the substantive claim that climate change impacts debt sustainability, including among the staffs of the IMF and the World Bank. But there is equally widespread uncertainty about how such considerations should be introduced into a DSA analytically. Marina Zucker-Marques, Kevin P. Gallagher, and Ulrich Volz (2024) provide an “enhanced DSA” that represents one conscientious effort to accomplish this difficult analytical task. Accounting for the costs of climate change by their method, they calculate that forty-seven of the sixty-six low-income countries in their sample breach critical thresholds by 2028.³⁴

Other analysts might challenge the method by which these climate adjustments are made to the framework and the (many) underlying assumptions. But studies such as Zucker-Marques, Gallagher, and Volz (2024) and Volz (2022) make the case that introducing climate into debt projections harms prospects for sustainability gravely. The World Bank and the IMF, in their recent reviews and guidance notes, acknowledge the importance of climate risk to the realism of forecasts.³⁵ It is clear that, going forward, leaving serious consideration of climate risks out of debt sustainability analysis would be decidedly risky from a financial point of view.

An expert task force on integrating climate into DSAs for low-income countries recommends that the World Bank and the IMF enhance the relevant methodology by improving data on climate risk and climate finance, employ climate risk scenarios that anticipate damage and costs, introduce these into a macro-financial model, and adopt a risk management approach when assessing fiscal and financial impact.³⁶ The Bretton Woods institutions should also revise the LIC-DSF to reflect the importance of climate investment–led growth, and the IMF should draw on the World Bank’s work on public sector investment and country-specific climate expertise in debt analysis. This agenda is more easily mandated than executed, but it lays down worthy markers.³⁷

Ishac Diwan, Martin Kessler, and Vera Songwe (2024) diagnose the financial situation of most lower- and lower-middle-income countries (LLMICs) as illiquid,³⁸ but not strictly unsustainable. They recognize that LLMICs must invest in climate mitigation and adaptation and propose a “liquidity bridge,” principally in the form of a quantum increase in financing from multilateral development banks. Climate investments, debt rescheduling, and the liquidity bridge thus constitute the three legs of their proposal. The rules governing such a program would be addressed by the G20 in the context of the Common Framework, suitably broadened to include the LMICs.

Daniel D. Bradlow, Rosa M. Lastra, and Stephen K. Park (2024), taking a legal approach, advocate introducing climate along with other environmental and social considerations into debt sustainability analysis. They recommend using legal norms and instruments to specify countries’ obligations in these areas, as agreed in international accords and treaties, and incorporating them into the IMF’s calculations of debt sustainability. Such obligations should be recognized to be of equal normative legal standing with countries’ contractual obligations to creditors. They emphasize that the Fund itself has acknowledged a relatively wide array of “new” issues to be “macro-critical,”³⁹ and thus these should be considered relevant to sustainability and program design. Bradlow, Lastra, and Park advocate that the IMF consult with the other international organizations, including UNCTAD, the UN Environment Program, and the UN Development Program, when revising the sustainability frameworks; extend participation in analyses to civil society organizations and other groups; and establish a mechanism to receive and investigate complaints by parties that claim to have been harmed by mistakes in the DSAs.⁴⁰

One conceivable alternative solution to the threats to debt sustainability posed by climate risks, complex as those are to incorporate into analysis, would be to simply wait until climate shocks occur and then restructure debt at that time. Such an approach would avoid having to calculate probabilities of such events and avoid false alarms. But it would only be feasible if sovereign debt restructuring were made far faster and easier than is presently the case. Current procedures, whether inside or outside the Common Framework, are not nearly expeditious enough given the size and frequency of anticipated climate shocks.

Selected Issues

Before turning to the agenda for improving the frameworks, this section considers six aspects of the DSAs: country income classification, bureaucratic organization and collaboration, assessment of debtors’ institutional strength, problems raised by endogeneity of the analysis, the role of staff judgment, and analysis of domestic political risk.

Country Income Classification

The assignment of countries to their income category is a critically important step procedurally prior to sustainability analysis and not always straightforward. In the World Bank classification, the cutoff point between “low-income” countries and “lower middle-income” countries is a gross national income (GNI) of $1,145 per capita as of 2023, below which a country is classified as low-income.⁴¹ Staff still hold some discretion in the assignment for countries that sit on the margin, and vulnerability to debt distress can be one basis for holding a country (such as Ghana) back from graduating from the low-income to lower middle-income category.⁴² Sri Lanka was temporarily “reverse graduated” from middle-income status to IDA-eligible in 2022,⁴³ although it nonetheless took the SRDSF and thus restructured its debt outside the Common Framework.⁴⁴

For countries near the threshold, the assignment can be very consequential, because the two frameworks differ among other things, as discussed, with respect to consideration given to external debt as opposed to the total public debt burden. Ghana, for example, would have been assessed to be capable of carrying considerably more debt if it had been classified as a lower middle-income rather than a low-income country, because it would have fallen under the SRDSF as opposed to the LIC-DSF. Under the SRDSF, it is possible that Ghana’s creditors would have been called upon to deliver perhaps only half as much debt relief as they actually provided.⁴⁵

Bureaucratic Collaboration and Organization

The LIC-DSF is the joint product of the World Bank and the IMF. The mandate for the two institutions to collaborate dates from the origin of the framework, with the launch of the Highly Indebted Poor Countries Initiative (HIPC) in 2005.

The rationales for the World Bank’s involvement in the case of low-income countries are that it is often the largest single creditor, its advice and lending also depend on the country’s sustainability rating, and the substantive concerns of development, structural policy, and capacity, in which the Bank tends to have comparative advantage, loom large in macroeconomies of low-income countries relative to market-access countries. With respect to lending, a country’s debt-carrying capacity will determine the mix of grant and concessional financing offered by the Bank. With respect to expertise, the Bank possesses particular knowledge in public debt management, domestic debt markets, and project finance, which can be macro-critical in low-income countries. Mutual responsibility for the analysis facilitates consistency in the advice given to low-income countries by the two institutions. When the LIC-DSF was originally approved, the Executive Boards of both institutions approved the joint proposal (IMF 2004 and 2005). Since then, the Bank both shares the review of the LIC-DSF framework and weighs in on individual DSAs.⁴⁶

The World Bank’s level of involvement in the LIC-DSF has grown over time. It used to be quite limited, but now a small team of financial economists conducts debt analysis on its behalf. About twenty-five professionals staff the debt office, one of three offices within the Bank’s Macroeconomics, Trade, and Investment Global Practice, which was recently renamed Economic Policy.⁴⁷ The debt team consists of debt management experts, often from national debt management offices, and financial and macro economists. In general, the Bank has a greater interest and say on matters of domestic debt, transparency, and long-term growth, for example, than the IMF. Shorter-term issues are the province of the IMF, including the sustainability of the debt in the context of program lending, especially if restructuring is warranted.

On the IMF side, the Strategy, Policy, and Review Department (SPR) takes the lead in the establishment of the debt frameworks and reviews the DSAs that are prepared by country teams for annual surveillance, regular programs, and programs requiring debt restructuring. The Debt Policy Division within SPR is dedicated to these tasks. The area departments of the Fund are responsible for sending the country mission and composing the DSA using the template mandated by the LIC-DSF. The DSA and the assumptions that are entered into it are managed and overseen by the country mission chief based in the relevant area department.

Regular DSAs are conducted for all countries, and, in many cases, disagreements between the World Bank and the IMF are fairly readily resolved. But cases involving sovereign debt restructuring can be fraught.

Regular DSAs are conducted for all countries, and, in many cases, disagreements between the World Bank and the IMF are fairly readily resolved directly between point people in the debt office at the Bank and SPR and the country team at the IMF.⁴⁸ But cases involving sovereign debt restructuring can be fraught, principally because they have distributive consequences for the stakeholders rising into the billions of U.S. dollars. Institutionally speaking, the Bank is concerned that its lending terms reflect the rating, whereas the Fund needs assurances that the program is fully financed.⁴⁹ In such cases, intense attention is focused on the debt and debt service ratios that define sustainability for the country concerned.

Overall, the IMF is in the lead, owing to a couple of factors. First, and foremost, the large creditor countries that sit on the Executive Boards of both institutions have chosen the Fund as the location to discuss and resolve disagreements over the DSAs. Whereas executive directors on the Fund side occasionally challenge staff work on a DSA, the World Bank’s Executive Board almost never interjects itself. Second, even though both institutions are generally exposed, the Fund’s lending is more immediate, and the amount of its near-term assistance is typically larger than the Bank’s.

 Institutional Analysis

When determining the debt-carrying capacity of a country, the LIC-DSF (and the SRDSF) draw upon a Composite Indicator, which is in turn substantially based on a World Bank product called the Comprehensive Policy and Institutional Assessment (CPIA). The intuition behind using CPIA is that countries with strong institutions should be able to take on and service more debt with greater reliability than countries with weak or missing institutions. The CPIA does not measure the policies and institutions of the debtor directly; rather, it collects information on the perceptions of World Bank staff about them. It aggregates ordinal rankings (one to six) by staff for each member country on sixteen criteria in four broad areas.⁵⁰ These areas and criteria are set out in box 1.

As is apparent from the list, the scope of these institutional criteria is very broad. Although results are mixed, several studies show positive correlations between the strength of institutions by this measure and sustainability.⁵¹ Qualms about the CPIA nonetheless led to its reconsideration during the last review of the SRDSF framework in 2021. The CPIA is now subsumed within the still-new Composite Indicator, although it constitutes a relatively large 45 percent share of it. The Composite Indicator also incorporates real GDP growth (20 percent), remittances as a share of GDP (10 percent), international reserves (15 percent), and world growth (10 percent). The new criteria were added at the instigation of the IMF and are more direct determinants of debt-carrying capacity. But in the comingling of the institutional and economic criteria, it is possible to detect a bureaucratic compromise between the two institutions.

Although CPIA might be useful as a comprehensive indicator for other purposes in the World Bank, its contribution to analyzing debt sustainability is questionable.⁵² As far as predicting debt distress or default is concerned, the CPIA is a triumph of empiricism over theory. Although CPIA scores are indeed correlated with debt distress and defaults, the relationship does not look strong.⁵³ While institutions certainly matter, the measure is too blunt to fit the purpose of debt sustainability analysis.

Consider the individual criteria in the indicator. Those in the economic management cluster are clearly related to debt accumulation, servicing, and redemption. A couple of criteria in the public sector management area can be expected to bear on debt sustainability, such as budget and financial management, revenue mobilization, and even central bank credibility.⁵⁴ But most of the other criteria are linked in a causal sense only tenuously to countries’ capacity to carry debt, and some extremely so. The social inclusion and equity criteria, while very important objectives, are decidedly ambiguous with respect to debt distress. Given the range of relevance of these sixteen criteria, it is remarkable that there is any detectable correlation between the aggregate indicator and debt-carrying capacity or distress. The debt frameworks would do better to restrict themselves to using the subset of criteria that are causally relevant.

It is worth noting that the Multilateral Investment Guarantee Agency (MIGA), another arm of the World Bank Group, does not use CPIA when underwriting political risk insurance. It refers to some of the same criteria, as do private credit-rating agencies, but incorporates them instead into its own risk rating framework.⁵⁵

Endogeneity

The determination of sustainability with or without high probability is subject to a substantial degree of endogeneity. This means that the sustainability of the debt is endogenous to the IMF program, its financing (including from supplementary sources), and conditionality. A particular stock of debt might be serviceable with such a program but not without one, and this must be taken into account in the DSA. However, some analysts worry that such endogeneity can get the IMF into trouble if the program places too much stock in the policy reforms and resolve of the government to implement them. Worse would be for officials to take advantage of this circularity to dismiss well-founded concerns about sustainability.

In 2016, Ghana was on an explosive debt path, but it was spared findings of unsustainability by a statistical rebasing of GDP, which dropped the reported debt ratio by 20 percent in a single year, and oil discoveries, whose exchange-rate effects were one-off and revenue effects slow in coming.⁵⁶ Despite a subsequent further buildup of debt, the 2021 DSA found that the overall risk of external and overall debt distress was high but nonetheless assigned a rating of sustainable, based on the government’s policy package. The DSA said:⁵⁷

Public debt is sustainable provided that the authorities’ medium-term consolidation plan (reflected in the baseline) is rigorously and credibly implemented to improve the primary balance, put debt on a declining trajectory, and ensure continued market access. However, inadequate implementation of the plan, growth slowdowns, and exogenous shocks, including longer exit from the pandemic, worsening global risk sentiment, and adverse terms of trade movements, pose significant risks to the outlook. Therefore, a deeper consolidation effort is crucial to make sustainability more robust to setbacks in policy implementation and exogenous shocks.

Within a year, however, Ghanaian authorities had conceded the need for a comprehensive debt restructuring. Gill and Pinto (2023), among others, argue that this outcome was foreseeable in 2021 and earlier.

Similarly, in 2021, Kenya’s external position was deteriorating after the pandemic, and the country was assessed to be at high risk of debt distress. But it was nonetheless rated sustainable on the strength of its promise to implement expenditure and revenue reforms under a new EFF/ECF program. A popular uprising against the revenue measures threw the program off track in 2024, leaving the country in a precarious financial position.⁵⁸

The solution to the dilemmas posed by endogeneity would be to retain it as a feature of the DSA—as policy adjustments undeniably affect sustainability—but to avoid unjustified optimism about the political viability of implementation, shift the political assessment into judgment, where it can be more explicit, and clarify the basis for it, as unsparing as that might sometimes have to be.

Judgment

Although there is skepticism regarding the role of judgment in debt sustainability analysis, there does not seem to be any realistic prospect of eliminating it.⁵⁹ Indeed, as the task of predicting distress becomes more complicated—with increases in debt, the number of instruments, the diversity of markets, and political and institutional change—the number of factors that bear on sustainability but escape effective modeling is likely to increase rather than decrease. The task is compounded by introducing climate risk (and climate opportunities) into assessment of sustainability.

The challenge is to account for such non-modelable factors in a systematic, realistic way while minimizing bias. The frameworks have made some progress in this direction by requiring explicit statements about when judgment is used and the basis for it. But, first, this can and should be done more explicitly. Second, some of the ostensibly technical choices, for example, of assumptions that underly debt projections, embody (sometimes heroic) judgments about implementation and the borrower’s national politics. By presenting these more clearly, the IMF and World Bank can lay the basis for their own learning over time as experience is accumulated and analyzed internally by staff, the IMF’s Independent Evaluation Office, and outside experts.

Realistically, occasional misjudgments should be expected. But procedures should ensure that mistakes do not happen because staff dismiss or downplay the posture of the political opposition to government when elections are approaching, fail to incorporate important available information, or adopt a focus that is partial rather than comprehensive.

Domestic Political Analysis

A good deal of the DSAs’ opaqueness results from the IMF’s reluctance to fully embrace the extent to which sustainability and default are, as historian Niall Ferguson once quipped, “always and everywhere a political decision.”⁶⁰ Politics pervades sustainability in that a decision not to service creditors inheres in the ability and willingness of governments to extract the necessary resources from the economy and population. Politics are present when payment is due and over many years prior to that moment, when the legal and institutional capacity to service debt is established, or fails to be established, and borrowing decisions are made and scrutinized, or left unscrutinized. Such decisions are inescapably political, and assessment of debt sustainability should take the full measure of this fact.

The current frameworks take such factors into account, but they do so partially, unsystematically, and less than transparently. The SRDSF defines sustainability as there being “no politically and economically feasible policies that stabilize the debt-to-GDP ratio and deliver acceptably low rollover risk without restructuring and/or exceptional bilateral support, even in the presence of Fund financing.”⁶¹ Political feasibility is thus explicitly acknowledged as a basis for sustainability, and this applies to the LIC-DSF as well. However, in its internal deliberations and public country reports, IMF and World Bank staff are more comfortable and explicit when analyzing the economic risks.

It is a widely understood and generally accepted that the IMF should be nonpartisan in its engagement with member countries, although the Articles of Agreement tread softly on the matter.⁶² Longstanding policy holds that “the IMF’s judgments should not be influenced by the nature of a political regime of a country, nor should it interfere in domestic or foreign politics of any member.”⁶³ Internal regulations prohibit staff from speaking publicly or publishing on “any national political questions” or engaging in political activity that is “inconsistent with, or reflects adversely upon, the independence and impartiality required by their position as international civil servants.”⁶⁴ Basically, formal political neutrality derives from guidance from the Executive Board and is a preference of senior management. Staff and management have thus assiduously avoided domestic political questions in documents, public statements, and discourse about members.

However, staff and management’s rightful adherence to partisan neutrality within member countries has fed a larger aversion to things that could fall under the term domestic politics, even things that are nonpartisan and do not otherwise constitute interference. It is perfectly legitimate, and indeed necessary from the standpoint of program effectiveness and safeguarding the Fund’s resources, for IMF staff to analyze domestic political developments insofar as they relate to governments’ willingness and ability to implement programs and service debt. Undertaking such assessments is not partisan interference. Fund staff can also meet with opposition officials and gather information about political commitments to policy options, which they do with some frequency, without transgressing the boundary between information gathering and interference.

Considerable domestic political analysis is already done, as staff and management are generally reasonably savvy about internal politics of borrowing states. But much of this analysis is sotto voce and not recorded in public accounts of the risks to programs or sustainability, hidden to avoid accusations of partisanship or interference on domestic political questions. But the aversion now constrains candid, systematic examination of the domestic political risks to sustainability.⁶⁵ Deepening political analysis of debt sustainability would follow similar developments in other areas of the IMF’s operations, as with respect to fragile and conflict-affected states and anti-corruption.

Airing staff analysis of the domestic political risks for sustainability and programs more explicitly would have several benefits. First, doing so would define competing internal views more sharply and, in principle, facilitate better assessment with greater credibility externally. Second, being more explicit would clarify the assumptions behind the analysis and facilitate a mid-course correction if evidence emerges that they were mistaken. Third, doing so would enable better differentiation between economic and financial analysis on the one hand and political analysis on the other. (It reduces, for example, temptation to introduce expectations about election outcomes or regime survivability into the economic analysis implicitly via underlying assumptions about growth, productivity, exchange rates, and so forth.) Finally, cleaner separation of economic and financial analysis from political analysis would facilitate better ex post evaluation of each and improvements in analytical procedures over time.

The question remains as to how, in the context of DSAs, to conduct analysis of domestic political risks that is impartial and realistic.

The question remains as to how, in the context of DSAs, to conduct analysis of domestic political risks that is impartial and realistic. Staff’s first instinct might be to look to econometric work that relates political variables to default in the same way that IMF models treat macroeconomic and financial variables. Such work is not likely to be helpful in generating country-specific assessments, however, and political analysis will require greater reliance on qualitative methods. In the case of fragile and conflict-prone states, the relevant guidance directs staff to address a specific set of questions when considering the domestic political circumstances of countries.⁶⁶ Such checklists provide useful guidance, but the relevant questions vary considerably across countries undergoing assessments, depending on the nature of political development, regime type, age, and stability; Argentina requires a different set of questions than Chad, Laos, and Suriname.

A common method is to assign ordinal rankings to particular aspects of countries’ national political systems, such as the quality of the legislative process, bureaucratic capacity, debt repayment culture, security risks, data transparency, and corruption. The rankings are assigned by experts and aggregated into indexes of overall political risk. Such is the general approach used by most credit rating agencies for sovereign ratings, for example.⁶⁷

With its country expertise, resident representatives, regional offices, and networks of collaborative international organizations, the IMF should be able to improve on such simple, aggregated indexes. It should draw more heavily on country and regional political experts as consultants for these exercises. Staff can sharpen these findings by using interrater reliability tests, thereby better controlling for subjective variation among raters.⁶⁸

The IMF can and should subject the mechanical signals generated by indexes to a broader qualitative review and potential correction or override. Such review cannot of course be arbitrary but reasoned, explicit, and stated in such a way as to be potentially falsifiable when time has run its course and outcomes are known.

In high-stakes cases, of large exceptional access programs, for example, the IMF should think creatively and invest additional time and resources in political risk assessment. Scenario analysis, whereby default and non-default scenarios are laid out and probabilities assigned to each by an expert group, should be part of the tool kit. Gaming and simulations can be enlightening. Teams assigned to making the cases for optimism and pessimism respectively should debate the sustainability of the country before judges selected from management. Such exercises might not be definitive, but participants generally learn from them about vulnerabilities, determining factors, and plausibility of alternative scenarios.

The political component of the debt sustainability assessment would, like the other components, be led by the mission chief or country director with the assistance of the relatively small group of staff but drawing upon a wide array of functional and area expertise. Breadth of expertise should be mandated by the new guidance. The staff write-up should not only make the political assessment clear, but the basis for it should also be clear, so that country teams can correct any forecast error early and better diagnose mistakes after the fact.

Agenda for the Current and Future Reviews

If there were the opportunity to design debt sustainability analysis anew from whole cloth, it would ideally have several characteristics. It would be a single framework that could accommodate varying levels of development and thereby avoid large threshold effects arising from assigning countries to income groups. It would be transparent in the economic assumptions and policy parameters underlying the analysis. It would cleanly separate mechanical application of indicators and forecasts, on the one hand, from the exercise of staff judgment, on the other. And it would inspire confidence in the objectivity of analysis, specifically that insiders have not massaged findings to achieve a particular result. An ideal framework would not eliminate disputes, because actors suffering losses will raise objections and have a right to voice those based on evidence. Nor would a good framework eliminate false alarms or missed defaults, because such analysis is inherently probabilistic. But it could minimize them.

As a practical matter, however, in the current review of the LIC-DSF, the IMF and the World Bank must amend the framework as it now stands, within the mandates of the two Executive Boards, which are evolutionary. This section addresses eight reforms that can be achieved within the current review as well as reforms that might have to wait for subsequent reviews, including the upcoming review of the framework for market-access countries.

1. Threshold effects. The rationale for separating the debt frameworks for low-income and market-access countries rests on the different resource pools from which they draw at the World Bank and the IMF. But the separation has generated large threshold effects for countries that are close to one another yet fall on opposite sides of the cutoff level, as well as for countries that graduate from low- to middle-income status. Such effects should be ameliorated, if not eliminated altogether, by de-emphasizing categorial assignments in favor of continuous measures. Doing so would require a degree of coordination of the LIC-DSF review with ongoing reconsideration of the SRDSF—a departure from the practice of evaluating and reforming the two frameworks separately. Joint consideration would likely test, in turn, collaboration between the IMF and the World Bank.
2. Total versus external public debt. Perhaps the most important difference between the two frameworks is their focus on total debt versus external debt. That domestic and external debt distress are highly correlated is well taken. Nonetheless, when a crisis strikes, the foreign exchange window closes firmly; there seems to be no escaping the need to analyze both separately. Even advocates of total debt as the focus of analysis for low-income countries acknowledge the need for analysis of the foreign exchange window’s capacity to support outflows.⁶⁹ Moreover, for reasons of political economy, domestic and external debt remain very different in their amenability to restructuring. 
3. Political risk. IMF staff and management should request and the Executive Board grant a clear mandate as part of the debt sustainability framework to gather information on and analyze domestic political developments insofar as they might affect a government’s ability and willingness to service debt, control further debt accumulation, and implement any accompanying program. Staff and management should be instructed to be realistic about political support for and against debt service and program implementation, as well as about borrowers’ incentives to discriminate among creditors when servicing debt and creditors’ abilities to condition further financial support on political considerations. (See the preceding section on domestic political analysis for further suggestions.)
   
   For example, the DSA should not simply assume that governments will fully implement the fiscal measures in the program, which biases analyses optimistically. Von Luckner et al. (2024, 36) recommend dispensing with forecasting the fiscal variables in favor of simply using the debt and fiscal information available at the time of the DSA. If the IMF has difficulty realistically acknowledging implementation risk, then it would be better not to premise the debt forecasts on out-year fiscal projections.
4. Institutional analysis. The frameworks should retain analysis of the strength of institutions, as there are both theoretical reasons and evidence to suggest that they are important. But the criteria for institutional strength should be honed to a subset of those now in the CPIA that have a more direct causal bearing on sustainability, and the institutional indicators should not be subsumed within a composite that conflates them with economic indicators; rather, they should be separated.
5. Cleanly separate objective analysis from political decisions. Maintain the separation principle in staff independence and analytical objectivity, on the one hand, and the policy decisions, which are the responsibility of the Executive Boards, on the other. Ultimately, the board decides on lending and, notwithstanding previously adopted policies, can simultaneously choose to make an exception to policy or amend the policy to address the case at hand.⁷⁰ These should be cleanly separated: that is, members of the board should not pressure staff to tweak calculations to achieve a desired result, and staff should not massage the DSA to provide cover for a political decision that is not otherwise supported by technical analysis. The board cannot, and arguably should not, act apolitically; after all, it must bear political responsibility for decisions taken. Nonetheless, when the board chooses to override a staff recommendation out of strategic or systemic considerations, the staff analysis should be as objective as possible—both for the analytical integrity of the forecasts and so that each can be accountable for its role in the process.
6. Judgment. Notwithstanding the economic objectivity of the analysis and the independence of the staff, and staff preference at both institutions for relying on mechanical signals where possible, retaining judgment seems unavoidable. As the assessment of risk becomes more complicated—notwithstanding simplification of the model—omission of important variables becomes more likely and mechanical signals are likely to become less rather than more reliable. There will be a continuing need for discretion and judgment involving factors and circumstances that cannot be modeled. It will be important to further delineate and separate the judgment of staff on technical matters (including underlying assumptions) from the more strategic judgment of the board. The basis for judgment at both levels should be clearly laid out.
7. Climate. Climate advocates who recommend transformational change to the frameworks are likely to be disappointed by the current LIC-DSF review, because introducing climate risk into DSAs is analytically challenging and owing to backlash against the climate agenda on the part of the administration of U.S. President Donald Trump and other governments. Such advocates are engaged in a long game and can advance this agenda in subsequent reviews. Nonetheless, constructive, though more modest, steps should be taken now.
   
   The Fund and Bank should begin by distinguishing between the continuous (through rising) costs of climate change and shocks in the form of extreme weather events. The former are more predictable than the latter, notwithstanding the uncertainties; DSAs should incorporate the continuous costs and treat the costs arising from shocks separately. To address shocks, the IMF, the World Bank, and the G20 should accelerate the debt restructuring process, so that debt can be treated more expeditiously when such shocks occur, and promote the introduction of climate-contingent debt instruments into official and private lending. The Fund and Bank should promote the use of these and other state contingent debt instruments, such as GDP and commodity-linked bonds, notwithstanding the computational complications they pose for DSAs. Changes to DSAs do not have to address the costs of climate alone; they can work in conjunction with accelerated restructuring and contingent instruments.
8. Institutional collaboration. Joint ownership of the LIC-DSF is perhaps the most cited example of World Bank-IMF collaboration and is generally thought to be successful.⁷¹ But some aspects of their joint work on the LIC-DSF should be reexamined. First, the continued use of the CPIA within the Composite Indicator, despite the tenuous relevance of perhaps half of the institutional criteria, renders assessment of debt-carrying capacity, at a minimum, unnecessarily complex and opaque (see recommendation 4, above). Second, the mandate to collaborate is an obstacle to co-evolving the debt sustainability frameworks for low-income and market-access countries to reduce threshold effects. The two institutions should address these problems collaboratively; but if they cannot overcome them, perhaps the mandate should be reassessed more comprehensively.

The pace of change in the global economy and domestic and international financial markets calls upon the IMF and World Bank to be ambitious in the current review of the LIC-DSF. Nonetheless, participants in and observers of the review should recognize that the process of improvement is a long game: reforms that are desirable and perhaps necessary in the long run might not be achievable this year, but they can be revisited in future years. And it is well to emphasize that, while the staffs of the two institutions prepare the proposals, it is the member states that are ultimately responsible for guiding the work through the Executive Boards and securing a successful conclusion.

About the Author

C. Randall Henning is a professor of international economic relations at the School of International Service at American University. He specializes in international and comparative political economy, global governance, and regional integration. He has focused recently on international regime complexity, resilience of multilateral cooperation, and the future of the International Monetary Fund and regional financial institutions. Challenges to emerging-market and developing countries, debt sustainability and restructuring, Europe’s monetary union, and the global financial system are topics of special interest. He has authored or co-authored numerous books and journal articles on these issues.

Acknowledgments

The author acknowledges Marialena Athanasopulou, James Boughton, Liu Dongming, Tamar Gutner, Martin Mühleisen, Sun Jingying, and Jeromin Zettelmeyer, among others, for comments on earlier drafts, and thanks the many people who made themselves available for conversations on a not-for-attribution basis. The author alone is responsible for any errors in fact or judgment. Comments are welcome at henning@american.edu.

References

Abrams, Alisa, and Cyrus Rustomjee. 2024. “Toward an IMF Framework for Engagement with Partners.” IEO Background Paper BP/24-01/04. Washington, D.C. May 9.

Allen, Mark, and Danny Leipziger. 2005. “Operational Framework for Debt Sustainability Assessments in Low-Income Countries—Further Considerations.” Washington, D.C.: IMF and IDA, March 28.

Bradlow, Daniel D., Rosa M. Lastra, and Stephen K. Park. 2024. “Re-thinking the Sustainability of Sovereign Debt.” Journal of International economic Law 27: 336-52.

Bradlow, Daniel D. 2024. “A New Conceptual Framework for African Sovereign Debt: Finding an Optimal Outcome that Addresses 5 Challenges.” SSRN Paper. August.

Caliari, Aldo. 2024. “Debt Concerns of Low-Income and Emerging Market Countries.” In Oxford Handbook of the International Monetary Fund, eds Mark Hibben and Bessma Momani, pp. 608-29. Oxford, UK: Oxford University Press.

Chen, Muyang. 2024. The Latecomer’s Rise: Policy Banks and the Globalization of China’s Development Finance. Ithaca, NY: Cornell University Press.

Diwan, Ishac, Martin Kessler and Vera Songwe. 2024. A Bridge to Climate Action: A Tripartite Deal for Times of Illiquidity. Finance for Development Lab, Policy Note 14. Paris, January.

Dreher, Axel. 2022. Banking on Beijing: The Aims and Impacts of China’s Overseas Development Program. Cambridge, UK: Cambridge University Press.

Erce, Aitor. 2024. “Exceptional Access Criteria Part II: Debt Sustainability and Market Access.” IEO Background Paper to the EAP Evaluation, BP/24-02/04. Washington, D.C.: IEO, December.

European Commission. 2024. Post-Programme Surveillance Report: Greece. Brussels: European Commission. Autumn.

European Stability Mechanism. 2023. “Assessing Debt Sustainability: Modelling Challenges and the Way Forward.” Luxembourg. 30 November.

Fitch. 2023. “Sovereign Rating Criteria.” Available at https://www.fitchratings.com/research/sovereigns/sovereign-rating-criteria-06-04-2023

G20 Finance Ministers and Central Bank Governors. 2020. “Statement, Annex I: Common Framework for Debt Treatments beyond the DSSI.” Virtual. November 13.

Gallagher, Kevin, and Yan Wang. 2020. “Sovereign Debt through the Lens of Asset Management: Implications for SADC Countries.” Boston University Global Development Policy Center, GEGI Working Paper 042. Boston, November.

Gaudin, Antoine, Brendan Harnoys-Vannier, and Martin Kessler. 2024. “How Do Debt Forecasts Go Wrong? Insights and Takeaways for Future Reforms.” Finance for Development Lab Working Paper. Paris, December.

Gelpern, Anna, Sebastian Horn, Scott Morris, Brad Parks, Christoph Trebesch. 2021. How China Lends: A Rare Look into 100 Debt Contracts with Foreign Governments. AidData.

Georgieva, Kristalina, and Rhoda Weeks-Brown. 2023. “The IMF’s Evolving Role within a Constant Mandate.” Journal of International Economic Law 26 (1): 17-29.

Gill, Indermit, and Brian Pinto. 2023. “Making the Low-Income Country Debt Sustainability Framework Fit for Purpose.” World Bank Group, Policy Research Working Paper No. 10602. Washington, D.C., November.

Gill, Indermit, and Brian Pinto. 2024. “Making the Market Access Countries’ Debt Sustainability Framework Relevant for Emerging Markets.” World Bank Group, Policy Research Working Paper No. 10831. Washington, D.C., June.

Government of Zambia, Ministry of Finance and National Planning. 2023. “Debt Sustainability Analysis Report.” Lusaka, July.

GSDR [Global Sovereign Debt Roundtable]. 2023. “Cochairs Progress Report.” Washington, D.C.: IMF. October 12.

GSDR. 2024. “3^rd Cochairs Progress Report.” Washington, D.C.: IMF. October 23.

Gündüz, Yasemin B. 2017. “Debt Sustainability in Low-Income Countries: Policies, Institutions, or Shocks?” IMF Working Paper 17/114. Washington, D.C. March.

Gutner, Tamar. 2023. “Collaboration, Cooperation, Coordination: A History of the Bretton Woods Twins’ Efforts to Work Together.” Review of International Political Economy 30 (3): 965-990.

Gutner, Tamar. 2025. “Conceptualizing and Evaluating How International Organizations Collaborate.” World Development 186: 106789.

Guzmán, Martín, and Daniel Heymann. 2015. “The IMF Debt Sustainability Analysis: Issues and Problems.” Journal of Global Development 6 (2): 387–404. DOI 10.1515/jgd-2015-0034.

Guzmán, Martín, and Joseph E. Stiglitz. 2024. “The Practice of Sovereign Debt Sustainability Analysis.” Debt Sustainability Assessments & Their Role in the Global Financial Architecture series. Bonn, Germany: Jubilee USA Network and Friederich-Ebert-Stiftung. August.

Guzmán, Martín, and Joseph E. Stiglitz. 2024b. “The Practice of Sovereign Debt Sustainability Analysis.” Slide presentation to the Finance for Development Lab, Paris School of Economics. Paris. April 4. [collapse]

Gwet, Kilem L. 2020. Handbook of Inter-Rater Reliability, 4^th ed. New York: AgreeStat Analytics.

Hagan, Sean. 2025. “The IMF’s Exceptional Access Policy: It’s Not Working.” SciencesPo Chair in Sovereign Debt and Finance Policy Note 21, and Finance for Development Lab. Paris, January.

Henning, C. Randall. 2017. Tangled Governance: International Regime Complexity, the Troika, and the Euro Crisis. Oxford: Oxford University Press.

Henning, C. Randall. 2023. “How to Make Sovereign Debt Restructuring Work Now.” Centre for International Governance Innovation (CIGI) Policy Brief No. 179. Waterloo, Ontario. October.

Henning, C. Randall. 2024. “The IMF’s April 2024 Policy Reforms and Sovereign Debt Restructuring.” CIGI Papers No. 304. Waterloo, Ontario. September.

Huang, Yufan, and Deborah Brautigam. 2025. “Socialisation, Policy Opportunity, and Bureaucratic Bargaining: Explaining China’s Zig-Zag Engagement with Multilateral Debt Restructuring.” Review of International Political Economy 32 (4): 1027-50.

Hurley, Gail. 2024. “How Transparency Makes Debt Sustainability Analyses a Trusted and Effective Tool.” Jubilee USA Network and Friederich Ebert Stiftung. New York: May.

IEO-IMF. 2016. “The IMF and the Crises in Greece, Ireland, and Portugal.” Evaluation Report. Washington, D.C.

IEO-IMF. 2020. “IMF Collaboration with the World Bank on Macro-Structural Issues.” Washington, D.C.: IMF.

IEO-IMF. 2024. “The IMF’s Exceptional Access Policy: Evaluation Report”. Washington, D.C.: IEO, December.

IMF. 2004. “Debt Sustainability in Low-Income Countries—Proposal for an Operational Framework and Policy Implications.” Prepared by the IMF and World Bank. Washington, D.C. February 3.

IMF. 2005. “Operational Framework for Debt Sustainability Assessments in Low-Income Countries—Further Considerations.” Prepared by the IMF and World Bank. Washington, D.C. March 28.

IMF. 2012. “Staff Guidance Note on the Fund’s Engagement with Countries in Fragile Situations.” Washington, D.C. April 25.

IMF. 2014. “Guidance Note on the Treatment for Fragile States in Conflict Situations.” Washington, D.C.

IMF. 2017. “Review of the Debt Sustainability Framework in Low-Income Countries: Proposed Reforms.” Washington, D.C., September.

IMF. 2018a. “Guidance Note on the Bank-Fund Debt Sustainability Framework for Low Income Countries.” Washington, D.C.: IMF, February.

IMF. 2020. “Eligibility to Use the Fund’s Facilities for Concessional Financing.” Washington, D.C. March.

IMF. 2021. “Ghana: Article IV Consultation.” IMF Country Report No. 21/165. Washington, D.C. July.

IMF. 2022. “Staff Guidance Note on the Sovereign Risk and Debt Sustainability Framework for Market Access Countries.” Washington, D.C. August.

IMF. 2024a. “Kenya: Seventh and Eight Reviews under the EFF and ECF . . . – Debt Sustainability Analysis.” Washington, D.C.: IMF, October 18.

IMF. 2024b. “Supplement to 2018 Guidance Note on the Bank-Fund Debt Sustainability Framework for Low Income Countries.” Washington, D.C.: August.

IMF. 2025. “Argentina: Ex-Post Evaluation of Exceptional Access under the 2022 Extended Fund Facility.” Country Report No. 2025/003. Washington, D.C. January.

Independent Evaluation Group, World Bank. 2023. The World Bank’s Role in and Use of the Low-Income Country Debt Sustainability Framework. World Bank Group: Washington, D.C. April 26.

Kaufmann, Daniel, and Aart Kraay. 2024. “The Worldwide Governance Indicators: Methodology and 2024 Update.” World Bank Group Policy Research Working Paper No. 10952. Washington, D.C., November.

Kincaid, Russell G. 2016. “The IMF’s Role in the Euro Area’s Crisis: What Are the Lessons from the IMF’s Participation in the Troika?” Background paper for the IEO Evaluation on the IMF programs for Greece, Ireland and Portugal. Washington, D.C.

Knack, Stephen. 2013. “It’s Only Words: Validating the CPIA Governance Assessments.” World Bank Policy Research Working Paper 6526. Washington, D.C. June.

Kozack, Julie. 2005. “Considerations in the Choice of the Appropriate Discount Rate for Evaluating Sovereign Debt Restructurings.” IMF Policy Discussion Paper 05/9. Washington, D.C., December. 

Kraay, Aart, and Vikram Nehru. 2006. “When Is External Debt Sustainable?” World Bank Economic Review 20 (3): 341-65.

Lazard. 2024. “Climate and Sovereign Debt Vulnerabilities: Some Practical Solutions.” Policy Brief. Paris. February.

Makoff, Gregory. 2023. Default: The Landmark Court Battle over Argentina’s $100 Billion Debt Restructuring. Washington, D.C.: Georgetown University Press.

MIGA. 2021. Investment Guarantee Guide: Insuring Investments, Ensuring Opportunities. World Bank Group, Washington, D.C.

Ministry of Finance of [the] People’s Republic of China. 2019. “Debt Sustainability Framework for Participating Countries of the Belt and Road Initiative.” Beijing: CMOF, 25 April.

Moody’s Ratings. 2022. “Rating Methodology: Sovereigns.” London, 22 November.

Setser, Brad W. 2024. “The IMF Needs to Focus on Setting Good Targets for External Debt Sustainability.” Follow the Money blog. Council on Foreign Relations. February 22.

Setser, Brad W. 2025. “Focusing the IMF’s Role in Development Finance on Debt Sustainability,” In ICE, Revista de Economía, edited by José Antonio Alonso and Marta Mulas. Madrid. https://doi.org/10.32796/ice.2025.939.7904.

Siriwardana, Mahinda K. M. 2025. “Sri Lanka’s Debt Restructuring: Key Lessons, Global Creditor Engagement and the Way Forward.” Paper prepared for the SIS-FDL workshop on sustainable development finance, Shanghai, December 4-5.

Standard & Poor’s. 2019. “How We Rate Sovereigns.” S&P Global Ratings paper. February 15. Available at www.spglobal.com/ratingsdirect.

Task Force on Climate, Development and International Monetary Fund. 2024. Room to Grow: Integrating Climate Change in Debt Sustainability Analyses for Low-Income Countries. Global Development Policy Center, Boston University. Boston, September.

UNCTAD. 2015. “Sovereign Debt Workouts: Going Forward – Roadmap and Guide.” Geneva: United Nations. April.

UNCTAD. 2022. “UNCTAD Sustainable Development Finance Assessment (SDFA) Framework: Linking Debt Sustainability to the Achievement of the 2030 Agenda.” DA-COVID-19 Project Paper 16/22. Geneva: United Nations. 

UNCTAD. 2023. Trade and Development Report. Geneva: United Nations.

Virketis, Pijus. 2024. “The Product of All Fears: IMF’s USD GDP Forecast Quality during EM Bond Restructurings.” Manuscript. HBK Capital Management. September 8.

Volz, Ulrich. 2022. “Climate-Proofing the Global Financial Safety Net.” Journal of Globalization and Development 13 (1): 1-30.

Von Luckner, Clemens G., Sebastian Horn, Aart Kraay, and Rita Ramalho. 2024. “Predicting Debt Distress in Low-Income Countries.” World Bank, manuscript. Washington, D.C., November. 

Willems, Tim, and Jeromin Zettelmeyer. 2022. “Sovereign Debt Sustainability and Central Bank Credibility.” IMF Working Paper WP/22/16. Washington, D.C. January.

Wang, Yan, Ying Qian and Kevin Gallagher. 2022. “Debt Distress and Development Finance in the COVID-19 Era.” Boston University, Global Development Policy Center. GCI Policy Brief 013. Boston, May.

World Bank Group, Operations Policy and Country Services. 2012. “CPIA Criteria 2021.” Washington, D.C.: World Bank. At https://thedocs.worldbank.org/en/doc/69484a2e6ae5ecc94321f63179bfb837-0290032022/original/CPIA-Criteria-2021.pdf.

Ye Yu. 2023. “The G20 Reciprocal Socialization for global Coordination on Sovereign Debt and Development: Perspective of China.” International Politics 26: 1-22.

Zucker-Marques, Marina, Kevin P. Gallagher, and Ulrich Volz. 2024. Defaulting on Development and Climate: Debt Sustainability and the Race for the 2030 Agenda and Paris Agreement. Boston, London, Berlin: Boston University Global Development Policy Center; Centre for Sustainable Finance, SOAS, University of London; Heinrich Böll Foundation. April.