GLOBAL INSIGHT: Covid Debt of Emerging Nations — Who’s at Risk?

JOHANNESBURG (Capital Markets in Africa) — In 2020, emerging market debt rocketed higher. Has necessary pandemic stimulus planted the seeds of future fiscal crisis? Bloomberg Economics’ analysis shows South Africa is vulnerable but Brazil, India, Indonesia, Mexico, and Russia are less exposed. Persistently low-interest rates partially offset the impact of higher borrowing.

An increase in interest rates would raise the pressure, but from current low levels, the impact is likely manageable. Most of the debt is in the local currency, pays fixed coupons, and has medium- or long-term maturity — features that reduce its risk profile.

Under our base case, using market expectations for interest rates as well as our assumptions for primary deficits and economic growth, the debt-servicing burden will start rising marginally in the mid-2020s but will remain within historical bounds for most countries.

South Africa appears particularly risky — its debt-servicing ratio will rise to dangerous levels before the end of the decade, reflecting both higher debt levels and rising interest costs.

We also consider the case in which yields rise by 100 basis points above the baseline. Aside from South Africa, the other five countries remain robust, allowing their governments to withstand possible unwinding of some of the extraordinary monetary easing in advanced economies.

A 200-basis-point shock could create wobbles in India as servicing their large stock of debt becomes more burdensome. South Africa’s risks may crystallize sooner in this scenario.

Brazil, Indonesia, Mexico and Russia have room to provide more support to counter the Covid recession and avoid long-term scars to their growth potential. Ample global liquidity will help. Fiscal consolidation can arrive later when the economy is back near full health.

The pandemic has stretched the finances of emerging economies. Budget deficits widened as tax revenues slumped and health costs and stimulus outlays soared. Public debt in the six emerging markets covered here rose to an average 68.8% of GDP in 2020 from 56.8% in 2019, a bigger increase than that seen in the 2009 financial crisis.

Is this sustainable? We consider the path of interest costs as a share of GDP. The metric takes into account the size of borrowing as well as interest rates paid on debt. It’s a better gauge of sustainability than the more widely used debt-to-GDP ratio, which looks only at the former.

Take Brazil for example. Its debt ballooned from 62% of GDP in 2014 to over 100% in 2020. But the cost of servicing this debt last year was the same as in 2014 (5% of annual GDP) as falling interest rates offset the impact of higher debt. An increasingly credible central bank and a drop in inflation were instrumental in bringing borrowing costs down.

To calculate debt servicing costs out to 2030, we use DDIS to estimate refinancing needs and our macroeconomic assumptions to project new debt issuance. Interest on newly issued debt is derived from the forward curve FWCM. Further detail on the methodology is available at the end of the piece.

Base Case: All Safe Except South Africa
For the six EMs covered here, we estimate that their average debt interest payments will gradually rise from 3.6% of GDP in 2019 to 4.3% in 2030. Driving the evolution of the debt-servicing burden is a stabilization in the stock of debt over the next decade after a rise in 2020. Financial markets also expect interest rates to remain low and push average borrowing costs down, helping to keep debt-servicing costs under control.

The aggregate data hide significant variation between countries. South Africa stands out as particularly exposed. Its interest-to-GDP ratio is set to rise above 5% this year and top 11% in 2030, a sharp increase from an average 3.1% in 2005-2020. The level is higher than some recent cases of sovereign defaults such as Argentina (4% in 2001 and 2019) and Lebanon (about 10% in 2019). The combination of higher expected interest rates, blowout borrowing and stagnant economic growth is toxic for debt sustainability.

At the other end of spectrum, Russia appears particularly robust. Its debt-servicing costs will rise after the pandemic but remain low at around 1% of GDP, thanks largely to low debt levels. Brazil’s servicing costs are higher than most other countries in our sample, but will remain below 2005-2020 levels.

The above assumes that the yield curve will behave according to market expectations captured in the forward curve matrix FWCM, which on average points to interest rates remaining low through 2030. Only in South Africa are forwards anticipating significantly higher interest rates.

At current low levels there is scope for surprises, especially on the upside. We consider two shocks: a 100-basis-point and a 200-basis-point upward shift in the yield curve above the baseline.

What could trigger such shocks? There are a number of plausible possibilities: a rise in interest rates in advanced economies as they withdraw monetary stimulus — similar to the taper tantrum episode of 2013; a decline in commodity prices hitting exporters; a spike in inflation; or a political shock that raises the credit risk premium.

If yields rise by 100 basis points, debt-servicing costs would increase relative to the baseline but remain manageable for most of the countries we examine. In aggregate, the ratio will rise to 5% of GDP in 2030, a moderate increase over the baseline. Most countries will remain close to historical bounds, but South Africa’s dire fiscal arithmetic will further deteriorate.

With a 200-basis-point shock, the aggregate ratio would rise to 5.7% at the end of our forecast period. Debt-servicing costs would increase above the historical average for most countries. In Brazil, the ratio would be near the 2015 peak. The rise in India’s interest payments close to 8% of GDP could become problematic.

EM currencies are typically more volatile than advanced economies and they’re subject to bouts of violent gyrations. Would this raise the debt-servicing costs? Yes, but not by much in the six countries we consider here.

That’s because the share of foreign-currency-denominated debt is small — ranging from practically zero in India to 23% in Indonesia. For example, if there was a one-time 20% depreciation, the average debt-servicing burden would be 4.4% in 2030 instead of 4.3% under the baseline. It would take a monumental collapse in the currencies to make a difference to public solvency.

This resilience to FX shocks is probably specific to the six countries we study here — others may be more exposed. Argentina and Lebanon, two recent examples of sovereign defaults, had a larger share of borrowing in dollars (53% in the former and 40% in the latter), and this composition contributed to their reneging on debt.

Brazil, Indonesia, Mexico and Russia are fiscally viable. Their debt dynamics are robust, even if interest rates rise from their current expectations. They have room to provide more support during the pandemic phase, and consolidate later when the economy is back in full health.

South Africa’s dynamics are worrying, made worse by stagnant economic growth. Adjustment is needed to lift growth or reduce deficits in order to avoid default over the coming decade, but such action doesn’t appear forthcoming.

The profile of debt provides some cushion against potential shocks. Most of the public debt is in local currency, pays fixed coupons and has medium- or long-term maturity. It limits the risks from currency depreciation, interest-rate hikes and sudden changes in investor sentiment.

How We Model Debt Interest Costs
Calculating interest-to-GDP has three inputs: the stock of debt, interest rates and nominal GDP. Below we describe in detail how we project each of these variables.

Debt has two components: outstanding debt and new borrowings to cover budgetary needs. For the former, we use the DDIS function to obtain the redemption profile of each country’s borrowings. The function also classifies securities by their currency of issue and coupon types: zero-coupon, fixed- and floating-rate, inflation-linked, local and foreign-currency bonds.

Our main assumption is that whenever a bond expires, it’s replaced with another that has the same currency, coupon type and tenor. The only thing that changes is the interest rate it pays. We obtain this from the FWCMfunction, which gives the forward rate. For example, it tells us what financial markets expect the yield on 10-year bonds will be three years from now.

How about new debt that results from budget deficits? We assume it’s issued to maintain the maturity structure of debt. If 20% of outstanding borrowing comes from 10-year, local-currency-denominated, fixed-rate bonds, then 20% of the new debt will be issued of the same type.

To be sure, there are instances when countries can deviate from our assumptions. For example, a country can fund its deficit with reserves instead of debt, issue debt on a shorter-term basis to reduce interest costs, or borrow at concessional rates. The combined effect would be lower projected borrowing costs than we currently assume. We’ll explore these possibilities for South Africa in a separate note.

For dollar-denominated bonds, we use the forward curve in the U.S. and assume the current credit spreads will persist. So if the spread between the five-year dollar bond in Brazil and the five-year U.S. Treasury bond is 100 basis points, we assume Brazil will refinance its five-year dollar bonds at 100 basis points above the five-year U.S. forward rate. We use the same assumption of bonds denominated in euros and yen.

For exchange rates, we assume the real exchange rate is stable, and the currency in nominal terms depreciates in line with inflation differentials.

Source: Bloomberg Business News