(Bloomberg) -- Emerging markets are under pressure from rising debt, slowing growth, and soaring yields. Sri Lanka has already joined the ranks of Lebanon and Zambia with an historic failure to pay. But the panic may be overdone. Our model shows that risks concentrate in small economies, and that most big developing countries will likely remain immune even if pressure on frontier markets intensifies.

• We’ve built a model to quantify the risk of default in 41 emerging countries over the next 12 months. Eleven countries have a default probability above 10%, including Argentina, Ecuador, and Ethiopia.• The most exposed nations constitute a small share (3%) of the world economy. Many, such as Pakistan and Ghana, are either seeking support from the International Monetary Fund or already receiving it. That should help contain contagion risks.• Larger economies such as Brazil, India, Indonesia, and Mexico are immune even if pressure on vulnerable countries intensifies. Turkey may not be so lucky. This is a stark difference from the 1980s when large emerging markets got into trouble.

Five ingredients are combining to ratchet up the risks from emerging-market debt.

First, the stock of borrowing in developing economies has risen from just over half of annual gross domestic product in 2019 to almost two-thirds this year. The pandemic led to an increase in public spending and a fall in tax intake in that span.

Second, global interest rates are rising at a speed not seen in four decades. The tightening will make it more burdensome to service debt denominated in foreign currency.

Third, weaker exchange rates in emerging markets are adding to the cost of external debt. Most currencies have depreciated by double digits against the dollar since the end of 2020. Governments receiving revenue in local currency and servicing debt in foreign currency will feel stretched.

Fourth, central banks in developing nations have been hiking interest rates even more aggressively than the US. This may tame inflation and prevent further weakening of the currency, but it’s adding to the public debt burden.

Last, but not least, defaults in Belarus, Russia, and Sri Lanka are raising the question of who could be next.

Countries to Watch

We took this question to our models. They combine the ingredients above (debt, interest rates, exchange rates) with other signals (financial market conditions, for example) to produce an estimate of the probability of sovereign default over the next 12 months for 41 countries.

The results? Excluding those already in default, 11 other economies have a nonpayment probability of 10% or higher in the coming year. The list includes Argentina, Ecuador, Ethiopia, Kenya, Pakistan, and Tunisia. They combine weak fundamentals—most have debt in excess of 60% of GDP—with a lack of support from financial markets, with most yields on dollar debt exceeding 20%.

At the other end of the scale, Indonesia, the Philippines, and Vietnam are least likely to default. Their stock of debt is more modest, with most at less than 45% of GDP. They’re experiencing relatively fast growth, most over 5% annually. And financial markets are rewarding them with relatively low interest rates.

To be sure, our analysis doesn’t capture all the drivers of default. One missing factor is politics, which can pressure countries into prioritizing imports of essential items like food and medication over debt servicing. Regulatory factors also can force financially viable countries into nonpayment, as was the case with Russia. Still, our models are a useful attempt to convert economic and financial data into default probabilities.

Our results are broadly consistent with data from financial markets, but there are notable differences. South American countries, such as Uruguay and Peru, are historically vulnerable to sovereign crises in their neighbors, such as Argentina and Ecuador. These economic linkages lead to higher default probabilities than financial markets would imply.

Another difference is Egypt, where we attach lower relative risk than markets do. Our country-level analysis agrees with this assessment: Egypt is more likely to address its problems through currency weakness than nonpayment of debt.

Not the 1980s

Our models show that rising debt, higher yields, and slowing growth are raising the risk of default in emerging markets. The consolation: A 1980s-style debt crisis still looks unlikely.

Fewer countries are at risk of default today compared with four decades ago. Our aggregate likelihood score, which weighs all countries equally, is currently near 10%—almost five times lower than its mid-1980s peak.

Risks also are concentrated in smaller emerging economies. A GDP-weighted index shows an even bigger gap between risks of default today vs. the early 1980s. Larger emerging economies like Brazil, Mexico, and Poland—which all defaulted in the 1980s—are more robust this time around.

Why? There’s an element of good policy here: These countries issue more debt in local currency, which is easier to service, and have accumulated more savings than before. There’s also an element of good luck: High commodity prices are providing a cushion for some big exporters.

Will larger emerging markets remain immune if smaller, distressed countries get into further financial trouble?

We construct a scenario in which borrowing costs in eight distressed countries rise by as much as they did in the first nine months of 2022, with spillover to all other economies using historical correlations. The aggregate GDP-weighted default probability will soar to 5% in 2023 from 1.5%. Default risks in bigger emerging markets such as Argentina and Turkey will rise above 40%. But other economies like Brazil and Poland will remain robust.

We also test a scenario in which US interest rates rise to 5% by the second quarter of 2023, in line with the call from our US team. The effect here is rather small compared with the baseline, in which rates rise to 3%. That makes sense: Going from 0% interest to 3% under the base case is painful, while the additional move from 3% to 5% will add only incremental damage.

Defaults can make a bad situation much worse, as the recent experience of Lebanon and Sri Lanka shows. In both countries, recessions have been deep, inflation very high, and politics unstable. There’s a silver lining: Recent IMF interventions, for example in Ecuador in 2020, have been effective in reducing the time a country stays in default by ensuring fast resolutions.

How We Model Default Risk

We estimate default probability by averaging the projections from two models.

Our economic model estimates default likelihood from three global variables (global industrial production, one-year US interest rate, US corporate bond spreads as a proxy for investment sentiment) and four domestic metrics (GDP growth, external debt, public debt, and the real effective exchange rate). The historical data are quarterly for 57 countries—advanced and emerging nations—spanning 1980 to the second quarter of 2022. The model is created by Ana Galvao, Michael McCracken, and Michael Owyang (2022).

Our financial model estimates default probability using yields on dollar debt and the share of interest payments to GDP. We estimate the model for 41 developing countries using data from 2011 to the second quarter of 2022.

Averaging two models allows us to extract more information from the available data. The alternatives—using either a shorter history combining both economic and financial market data into a single model, or ignoring financial market data altogether—are less efficient. Our models are dynamic, automatically generating forecasts for each underlying variable as well as the default probability for each country.

Our contagion analysis classifies countries as distressed if their dollar-borrowing costs exceed US rates by at least 10 percentage points. Eight nondefaulting countries are in this category: Argentina, Ecuador, El Salvador, Ethiopia, Ghana, Pakistan, Tunisia, and Ukraine. We don’t shock the yields of countries already in default.

©2022 Bloomberg L.P.