Following the release of our external liquidity scorecard last week, we have developed a debt sustainability index to provide a top-down look at which emerging and frontier markets are at the greatest risk of default. Again, we have sourced our indicators from centralised sources like the IMF and World Bank for availability and comparability across 46 emerging and frontier markets (six more than our external liquidity scorecard) and exclude countries that are already in default.
With 27 variables, we have divided the scorecard into three distinct categories: 1) fiscal policy and debt dynamics; 2) debt stock and flow; and 3) debt composition. For each category, we display the raw data in separate tables below. We then devise a composite debt sustainability index.
We score each variable in terms of standard deviations better (worse) than the sample median and take the simple average across variables to arrive at a composite debt sustainability index, as well as sub-indexes for fiscal policy & debt dynamics, debt stock & flow, and debt composition (we remove 10 countries due to insufficient data – see table notes). The resulting output is a quick and dirty way to quantify debt sustainability risk for each country on a relative basis.
Results
After excluding countries with insufficient data, we find that the most vulnerable countries are Jamaica, Sri Lanka, Mongolia, Pakistan, Tunisia, El Salvador, and Kenya. On the other hand, the least vulnerable countries are Uzbekistan, Nigeria, Vietnam, Kazakhstan, Peru, Cote d’Ivoire, and Serbia.
Moreover, a simple scatter plot shows that there is a negative relationship between the overall debt sustainability score and the country risk premium (measured by the EMBI), as we should expect. Adding a log-linear line of best fit (statistically significant at the 99% confidence level with an R2 =23.1%), we can use this model to make inferences about bond valuations.
On that basis, Sri Lanka, Ecuador, Angola, Nigeria, El Salvador and Tunisia all stand out as notably undervalued, despite elevated external vulnerabilities. Meanwhile, Jamaica, Mongolia, Colombia, Indonesia and Serbia appear to be the most overvalued credits. As we emphasised before, idiosyncratic factors may be at play and these results should be supplemented with traditional country analysis before drawing any conclusions.
Methodological issues
While we appreciate our model is highly stylised, the simplified and transparent approach is part of its appeal and we find it mostly offers intuitive results (and where it does not, this can be a signal for further investigation). Of course, we urge our readers to take this data and its conclusions with a pinch of salt, and caution that it should be used in conjunction with traditional country risk analysis.
We also recognise some drawbacks from this approach. Just like the external liquidity scorecard, this simple approach fails to account for potential non-linearities and threshold effects within variables, while equal weighting may ignore potential differences in importance. In addition, more timely and thorough data can be found for many countries using official sources, providing a more complete snapshot and oftentimes adding some useful context. Nor have we backtested the model and assessed its predictive power; any model is likely to give false signals (Type 1 and Type 2 errors).
Data availability and the vintage of the available data is another challenge, with most of the World Bank IDS indicators pertaining to external debt stock and composition only updated through 2019. The longer the lag, and less contemporaneous the data is, the less useful it is as an early warning indicator. We have tried to compensate for this by deflating 2019 debt values by the 2020 estimate for GDP and exports to account for the drastic changes seen across these indicators in 2020. This assumes debt is unchanged, which is unlikely given huge deficits last year, but is the best we can do given existing data constraints.
Moreover, we omit other indicators that might be a cause or signal of distress, including institutional factors and qualitative differences in public financial management and debt management capacity. Nor have we included a variable to capture default experience or payment track record, which can be an important indicator of willingness to pay and predictor of default. However, we have chosen a more general approach to allow for cross-country comparison, and think our scorecard serves as a useful warning light for debt distress.
In the Appendix, each variable is detailed with sources and relevant metadata. We welcome feedback from our readers on methodology and coverage and remain available to answer any questions.
Appendix: Data explanations and sources
Budget balance / GDP: Proj. budget deficit (surplus) as % of GDP for 2021. Source: IMF October 2020 WEO. Frequency: Annual (2021 proj.)
Primary balance / GDP: Proj. primary deficit (surplus) as % of GDP for 2021. Source: IMF October 2020 WEO. Frequency: Annual (2021 proj.)
Debt-stabilising primary balance / GDP: Debt stabilising primary balance implied from IMF WEO for 2021-25 period. Source: IMF October 2020 WEO. Frequency: Annual (2021-25 proj.)
Actual – stabilising primary balance: Difference between proj. 2021 primary balance and implied debt stabilising balance (eg amount of consolidation needed to stabilise the debt stock). Source: IMF October 2020 WEO. Frequency: Annual (2021-25 proj.)
r – g differential: Difference between proj. 2021-25 real GDP growth and proj. 2021-25 real interest rate (effective interest rate minus inflation) implied from IMF WEO. Note: positive (negative) differential means it increases (decreases) debt, all else equal. To the extent this is due to high inflation, benefit will likely be offset, in part or full, by corresponding currency depreciation (see next variable). Source: IMF October 2020 WEO. Frequency: Annual (2021-25 proj.)
Annual FX forecast (%): Constant annual appreciation (depreciation) of currency over 2021-25 implied by IMF WEO (derived by dividing GDP proj. in local currency by proj. in US$ terms). Source: IMF October 2020 WEO. Frequency: Annual (2021-25 proj.)
Public debt / GDP: Est. public debt / GDP ratio in 2020. Source: IMF October 2020 WEO. Frequency: Annual (2020 est.)
Interest / GDP: Proj. interest payments (derived by subtracting budget balance from primary balance) as % of GDP in 2021. Source: IMF October 2020 WEO. Frequency: Annual (2021 proj.)
Interest / revenue: Proj. interest payments % of government revenue in 2021. Source: IMF October 2020 WEO. Frequency: Annual (2021 proj.)
Total external debt / GDP: External debt (public and private) in 2019 as % of GDP in 2020 (to account for contraction). Source: World Bank International Debt Statistics for external debt and IMF October 2020 WEO for GDP. Frequency: Annual (2019 and 2020 est.)
Public external debt / GDP: Public and publicly guaranteed (PPG) external debt in 2019 as % of GDP in 2020. Note: Subtract from prior variable for private external debt. Source: World Bank International Debt Statistics for external debt and IMF October 2020 WEO for GDP. Frequency: Annual (2019 and 2020 est.)
Total external debt / exports: External debt (public and private) in 2019 as % of 12-month trailing exports of goods (to account for contraction). Source: World Bank International Debt Statistics for external debt and IMF Direction of Trade Statistics (via Bloomberg) for goods exports. Frequency: Annual (2019) for debt and monthly (Sep 2020) for exports
Public external debt / exports: Est. public and publicly guaranteed (PPG) external debt in 2019 as % of 12-month trailing exports of goods. Source: World Bank International Debt Statistics for external debt and IMF Direction of Trade Statistics (via Bloomberg) for goods exports. Frequency: Annual (2019) for debt and monthly (Sep 2020) for exports
Total external debt service / exports: Proj. external debt service (public and private, principal and interest) in 2021 as % of 12-month trailing exports of goods (to account for contraction). Source: World Bank International Debt Statistics for external debt and IMF Direction of Trade Statistics (via Bloomberg) for goods exports. Frequency: Annual (2021) for debt service and monthly (Sep 2020) for exports
Public external debt service / exports: Proj. public and publicly guaranteed (PPG) external debt service (principal and interest) in 2021 as % of 12-month trailing exports of goods. Source: World Bank International Debt Statistics for external debt and IMF Direction of Trade Statistics (via Bloomberg) for goods exports. Frequency: Annual (2021) for debt and monthly (Sep 2020) for exports
Public external debt service / revenue: Proj. public and publicly guaranteed (PPG) external debt service (principal and interest) in 2021 as % of proj. government revenue in 2021. Source: World Bank International Debt Statistics for external debt and IMF October 2020 WEO for revenue. Frequency: Annual (2021)
Effective interest rate (%): Proj. 2021 interest payments / 2020 public debt stock. Source: IMF October 2020 WEO. Frequency: Annual (2020 and 2021)
FX-denominated (%): Est. % of public debt that is denominated in foreign currency. Calculated by dividing PPG external debt / GDP (2019 data from World Bank IDS in numerator, 2020 data IMF WEO in denominator – see earlier methodology) by public debt / GDP (2020 data from IMF WEO). We acknowledge this is imprecise due to date mismatch, but the ratio should be relatively stable from year to year. Source: World Bank International Debt Statistics for PPG external debt and IMF October 2020 WEO for public debt and GDP. Frequency: Annual (2019 and 2020)
ST external (% of total): % of PPG external debt as of 2019 that is short-term. Source: World Bank International Debt Statistics. Frequency: Annual (2019)
Concessional (% of external): % of PPG external debt as of 2019 that is on concessional terms per World Bank definitions. Source: World Bank International Debt Statistics. Frequency: Annual (2019)
Bilateral (% of external): % of PPG external debt as of 2019 held by bilateral creditors. Source: World Bank International Debt Statistics. Frequency: Annual (2019)
Multilateral (% of external): % of PPG external debt as of 2019 held by multilateral creditors. Source: World Bank International Debt Statistics. Frequency: Annual (2019)
Bonds (% of external): % of PPG external debt as of 2019 corresponding to publicly traded eurobonds. Source: World Bank International Debt Statistics. Frequency: Annual (2019)
Other private sector (% of external): % of PPG external debt as of 2019 held by private creditors, excluding publicly traded eurobonds. Source: World Bank International Debt Statistics. Frequency: Annual (2019)
China (% of external): % of PPG external debt as of 2019 held by Chinese creditors (public or private). Source: World Bank International Debt Statistics. Frequency: Annual (2019)
Related reading
External liquidity scorecard, 11 January