Questions and Answers on the New Sovereign Risk and Debt Sustainability Framework for Market Access Countries
Last Updated: February 19, 2021
Why has the framework been renamed “Sovereign Risk and Debt Sustainability Framework for Market Access Countries” (MAC SRDSF)? Is it because it introduces an additional layer of analysis (sovereign risk) compared to the original framework?
The IMF debt sustainability analysis framework has always provided information on sovereign risk, even if it did not provide a specific risk classification as it does in the new framework. The framework has also been used to inform judgment on debt sustainability, and will continue to be used for that purpose. The new name, therefore, is better suited to describe the full purpose and use of the framework.What are the key reforms of the framework and how are these expected to contribute to the IMF core functions of surveillance and lending?
Staff has introduced several innovations to ensure that the new framework is a state-of-the-art toolkit for the analysis of sovereign stress. The three most relevant reforms are: (i) the time-horizon-based analysis of risks, (ii) accounting for a broader set of country-specific characteristics and their interaction, and (ii) moving towards a stochastic analysis of sovereign risks.
- A time-horizon-based analysis of risks—as also used in the frameworks of some other institutions, for instance, the European Commission—is crucial both because some models do best at predicting sovereign stress at a particular horizon, and to devise appropriate policy responses, e.g., to understand whether crisis prevention or crisis resolution might be needed (when policies do not have sufficient time to be effective), or to calibrate the pace of fiscal consolidation.
- Accounting for a broad set of country-specific characteristics and their interaction is critical to reflect countries’ differences in their capacity to meet the same level of debt obligations. Staff has captured this heterogeneity in several ways: introducing proxies of debt carrying capacity in the near-term risk tool (e.g. index of institutional quality and history of stress), using information on investor base to assess country-specific vulnerability to rollover risk in the GFN tool, and introducing trigger stress test for specific types of countries.
- The move towards a stochastic analysis of risks, including through a probability model for the analysis of near-term sovereign risks and the replacement of macro-fiscal stress-tests with enhanced debt fancharts that more accurately capture the actual uncertainty surrounding the baseline, acknowledges the fact that sovereign risks and debt sustainability are probabilistic concepts that must account for countries’ full range of uncertainty.
Will general government coverage be mandatory under the new framework?
General government would be the default and expected coverage under the new framework. This said, there are circumstances where a narrower or wider coverage may be appropriate and can be allowed. More details can be found in the Board Paper and Annex II.
Given that the new framework no longer includes fixed debt and gross financing needs (GFN) targets, does this mean that the IMF no longer considers debt and GFN levels relevant for sovereign risk and debt sustainability analysis?
Debt and GFN levels remain relevant in the MAC SRDSF: the debt level is accounted for in both the near-term tool and the medium term risk index derived from the fanchart, while the GFN level enters the medium term risk index through the GFN module. However, the framework no longer assumes that the same levels of these variables are relevant for all advanced and emerging economies, respectively. Debt levels enter in the context of models that also account for many other country characteristics, including debt structure, quality of institutions, international reserves, cyclical indicators, and global financial conditions. Furthermore, the evolution of the debt level as described in the fanchart does of course reflect interest rates and growth rates in addition to the primary fiscal deficit path and financing assumptions.
Could you provide additional detail on how the framework would be adapted to support sustainability assessments (particularly the three-zone sustainability assessments)?
In broad terms, the sustainability assessments will be based on three inputs: the SRDSF’s GFN and fanchart modules, as well as a multivariate econometric model which is analogous to the logit model used in the near term module, but specified and estimated on a dataset that focuses on “unsustainable events” (such as debt restructuring and hyperinflations) rather than merely episodes of sovereign stress (most of which can be resolved without debt restructuring).
The output of the three modules is summarized in a composite index, where greater index values indicate higher risk. The index is then divided into three risk zones, based on the probability that debt is not sustainable. Index values in the lowest risk zone (probability of unsustainable debt strictly below 20 percent) are interpreted as indicating that debt is sustainable with high probability, index values in the risk zone between 20 and 50 percent as indicating that debt is sustainable but not with high probability, and index values above 50 percent as indicating that debt is unsustainable. These “mechanical signals” are inputs in sustainability assessments that will ultimately remain judgement-based, in particular, because there could be relevant information that is not reflected in the models.What role is left for team judgement in the new framework?
Since any analytical model is a simplification of reality and cannot reflect all elements relevant for sovereign risk and debt sustainability analysis, judgement will retain an important role in the new framework and will be used by teams to derive the final, horizon-based risk assessments, as well as an overall risk assessment. The new framework, however, introduces enhanced transparency in the use of judgement, thus supporting evenhandedness. Clear guidance will be given to IMF staff on how to use judgment to reflect country-specific information not captured by the model. In addition, the application of judgment will need to be justified in the DSA write-up when it leads to risk assessments that are different from the risk signals derived by the models.
What is the logic of introducing an “Early Warning System” in the MAC DSA?
It is critical for the Fund, in its surveillance capacity, to assess crisis risks both in the short- and medium term, and advise country authorities on steps to lower these risks. Where warranted, staff risk assessment will take into account the authorities’ efforts, which could lead to an upgrade of the near-term risk assessments compared to the signal that is produced by the near-term model. The near-term risk signal and assessment will not be disclosed to the public during the first 12-months after the switch to the new methodology, to give the authorities time to familiarize themselves with it.
The new framework introduces clear sovereign risk signals. Does this mean that the IMF is moving towards becoming a rating agency?
No. For surveillance cases, the IMF’s and rating agencies’ models focus on different concepts: sovereign stress in the case of the IMF; and defaults/restructurings in the case of rating agencies. IMF models analyze the risk of “stress”, which is a different and broader concept than unsustainability of debt. In the majority of cases, sovereign stress can be resolved through additional financing or economic reform, without inflicting losses on creditors. Furthermore, the IMF does not attempt to rate risk on a finely graded scale, like the rating agencies to. The mechanical signals involve three risk levels: high, moderate, and low. The reason for this is to urge the authorities to take action when the risks are classified as high.
It is also relevant to note that the IMF has already moved towards a framework with mechanical risk signals in 2017 for low-income countries with the LIC DSF review. Introducing mechanical signals for MAC supports the IMF’s commitment to evenhandedness across its entire membership.
Finally, there are other public institutions with a surveillance mandate, such as the European Commission, that have developed frameworks that group countries according to sovereign risk levels. This does not turn the European Commission or the IMF into rating agencies. It supports the surveillance role of these institutions and their policy advice.
What is the logic of the 12-month transition period for publication requirements? And what information will be dropped from the published version of the staff reports during this 12-month period?
On the issue of publication, there were different views in the IMF Board, with some Directors favoring full disclosure of the proposed mechanical signals and risk assessments, and others concerned about the market implications of such disclosure.
Staff’s proposal – to have limited disclosure to the public (but full disclosure to the Board) for a twelve-month period – seeks to strike a middle ground, by allowing countries’ officials to familiarize themselves with the new tools and the new DSA reporting modalities. After twelve months, disclosure to the public would be reconsidered based on the experience gained with the new framework. This proposal was accepted by the Board.
During the 12-month period, the following information will be dropped from published staff reports: the three-zone sustainability assessment in non-exceptional access program cases, and the near-term risk signal and assessment in surveillance and precautionary program cases.
What are key issues for the Guidance Note?
The Guidance Note on the implementation of the new framework will include critical information on interpreting the models correctly, handling special country cases, evaluating conflicting signals, and identifying key situations where staff judgment would be essential. The Guidance Note will also pin down some of the more detailed operational issues, including: defining the types of financial assets to incorporate in the mechanical framework; identifying detailed criteria for including or excluding certain central bank liabilities in the public debt concept and when to use a consolidated public sector concept (that includes the central bank); and further specifying the parameters of the tailored stress test and long-term modules.