Just over a year ago, S&P Global Ratings published "Take A Hike: Which Sovereigns Are Best And Worst Placed To Handle A Rise In Interest Rates." In this article, we measured the direct fiscal cost to governments of two interest rate shock scenarios: a 100 basis point (bps) increase in the cost of refinancing central government debt, and a 300 bps increase. Since then, Russia has invaded Ukraine, oil prices have doubled, wheat prices have more than tripled, global inflation has surged, and the Fed has raised the funds rate by 150 bps. In contrast to May 2021, today all DMs except Japan, and all EMs except China, are refinancing maturing debt at higher rates than those they pay on their total outstanding debt. This means that the starting point (a central government's current market refinancing rates) for our 2022 interest rate shock exercise is notably higher than it was one year ago.

The cost of funding directly affects S&P Global Ratings' debt and fiscal assessments of sovereigns, and indirectly affects factors such as GDP growth. However, other factors such as external positions, monetary flexibility, the strength of institutions, and the absolute productivity of economies--as well as governments' and central banks' policy responses to unforeseen events--are greater drivers of sovereign ratings. Reflecting all these considerations, many of the sovereigns in this survey already have low credit ratings, incorporating the vulnerability of their public finances to interest rate shocks, among other risks.

Developed Market Sovereigns: For Many, A More Than 200 Bps Shock

Financing conditions have worsened over the last year, albeit real ex-post interest rates remain less than zero in many parts of the world. Still, the trend in real interest rates is ratcheting higher. For a majority of DMs, refinancing costs have increased by over 200 bps compared to one year ago. Only Japan is pricing new debt at below the average cost of total outstanding debt.

So what would be the fiscal cost of a 300 bps three-year rate shock for developed economy sovereigns? On average, it's a 1 percentage point (ppt) of GDP increase in interest spending by 2025 compared to this year's median interest expenditure to GDP of 2.2% (excluding the six DM sovereigns with central government debt of less than 50% of GDP). That is a significant pressure point on public spending in most cases, but still manageable, taking interest costs back to 2015 levels in France, for example.

However, a few DM sovereigns are more vulnerable to further rate shocks. Italy's elevated debt rollover rates (equivalent to more than 20% of GDP) and the recent run-up in its costs of new financing mean that even a 100 bps further increase in refinancing costs from today's levels would push its interest expenditure to GDP up by about 0.8 ppts, after three years of higher rates. Under this scenario, by 2025 Italy's interest expenditure to GDP would have returned to still-acceptable 2015 levels. However, a 300 bps three-year interest rate shock would move Italy's interest costs as a percentage of GDP to 5.5% (excluding second-round effects and assuming no change in Italy's debt structure including its average residual maturity). This is a level not paid since 2012.

To be clear, we see the probability of an additional 300 bps three-year rate shock for peripheral euro area sovereigns as unlikely. In our view, the European Central Bank is committed to ensuring a uniform monetary policy across the euro area. Over the next few weeks, it plans to launch a new anti-fragmentation mechanism, and reinvest maturing principal payments under the pandemic emergency purchase program until at least the end of 2024. Our current 'BBB' long-term sovereign rating on Italy already reflects the sovereign's high debt burden and limited fiscal flexibility, and our positive outlook is based on Italy's ongoing delivery of pro-growth reforms that we expect will improve its long-term debt dynamics. Italy's household and corporate sectors remain the least indebted in Europe, posting high savings rates, and the economy generates sizable annual external surpluses. This suggests that domestic residents have considerable capacity to absorb more domestic issuance in the event of further interest-rate volatility.

For Spain, with its lower rollover ratios and more advantageous starting point for refinancing rates, we estimate a 300 bps three-year rate shock would, nevertheless, increase interest payments to 3.0% of GDP by 2025, a level last paid in 2015 (and considerably above our estimate of 1.5% for 2022). Greece, whose 10-year yields have backed up an estimated 351 bps over the last year, has a modest amount of debt to refinance (just over 6% of GDP) per year in the commercial markets between now and 2025. Three-quarters of its sovereign debt is official lending at long maturities and low rates. This means that, even if Greece experienced yet another 300 bps rate shock, we would not project interest expenditure to GDP exceeding 3.0% in 2025 (versus 2.5% in 2022). Furthermore, Greece's Public Debt Management Agency is sitting on a cash position estimated at 17% of GDP, equivalent to just under three years of debt rollover requirements.

Our rollover ratios include short-term debt, which, by definition, has to be refinanced at least once a year. Given the high stocks of bills in both Japan and the U.S., the two countries are also in theory fiscally vulnerable to a rapid run-up in refinancing costs from where we are today, assuming no changes to their debt maturity profiles. It is important, however, to emphasize that at present Japan is refinancing its debt at slightly below 0%, or more than 100 bps below its average cost of financing. Moreover, Japan's very high domestic savings rates, and historic (though not current) tendency toward outright deflation would argue against the likelihood of a severe rate shock. The U.S. Treasury market remains the most liquid in the world, including the TIPS market (Treasury Inflation-Protected Securities), where yields begin to decline beyond 20-year maturities.

Emerging Market Sovereigns: Brazil And Hungary Join The Most Vulnerable

Unlike DM sovereigns, which borrow almost exclusively in their own currency, many EM sovereigns in our survey borrow in a mix of domestic and foreign currencies. For this reason, we have calculated their refinancing rates as the weighted average of domestic and foreign currency borrowing. For the purposes of this exercise, our estimates of "blended" domestic and foreign currency rates assume stable exchange rates.

Because of their higher rollover ratios and the rapid run-up in their cost of new debt, the five EM sovereigns most vulnerable to a 300 bps rise in refinancing costs are Ukraine, Brazil (BB-/Stable), Egypt (B/Stable), Ghana (B-/Stable), and Hungary (BBB/Stable). Ukraine is a particularly difficult case. As a consequence of Russia's invasion, the dollar value of Ukrainian GDP is set to decline by over 50% this year. This means the dollar share in total sovereign debt, as well as the amount of debt to refinance as a percentage of GDP, has soared (our long-term foreign currency rating on Ukraine is 'CCC+' with a negative outlook).

For Brazil, the difference between the cost of new and outstanding debt now exceeds 500 bps in light of the Brazilian central bank's rate hikes of 1,075 bps since March 2021; this, more than the rollover ratio, explains Brazil's vulnerability to any future rate hikes (though in Brazil's case, in our view, real interest rates have probably peaked, implying a lower likelihood of further rate shocks). We estimate that Egypt is due to refinance close to 32% of GDP in debt annually over the next three years (including short-term debt), the highest of all EMs. Elevated domestic inflation and rate hikes in Egypt have pushed up its refinancing costs to about 190 bps above the cost of outstanding debt. Ghana is also contending with a tightening central bank, a hawkish Fed, a volatile currency, and a fraught external environment, pushing up the cost of rolling over just over 15% of GDP in debt this year and next (though foreign currency redemptions are quite modest until 2025).

Like Brazil, Hungary is a new entrant in our top-five most vulnerable compared to one year ago. With a significant stock of government debt (unlike most other CEE peers) and an estimated annual rollover of 11.5% of GDP, the recent rise in the cost of new financing could lead to an increase in interest costs of 1.6 ppts of GDP by 2025 under the 300 bps rate shock scenario. This would still leave Hungary's projected cost of debt for 2025 at below the 4.5% of GDP average projected for all 19 EMs under the 300 bps scenario.

Nine of the 19 EM sovereigns in our survey would see their interest servicing costs increase by less than 1.2 ppts of GDP by 2023, even under a 300 bps rate hike scenario, considerably below the 16 out of 20 estimate of our 2021 survey. Sovereigns less vulnerable to an interest rate shock include Turkey (with very little debt to refinance, and no short-term debt to roll over), and all Asian sovereigns including India. The all-important caveat in the case of Turkey is that these results only hold in the absence of a further depreciation of the Turkish lira (which is currently trading at a real effective exchange rate 34% weaker than its 10-year average) or a normalization of its steeply negative real interest rates. The lira's fundamental undervaluation is not a guarantee against further depreciation--particularly in the face of the ongoing dollarization of the deposit base/government debt stock, Fed hikes, and balance-of-payments pressures connected to rising hydrocarbon and commodity prices.

Complicating their ability to control financing costs, Colombia, Egypt, Ghana, Kenya, Turkey, and Ukraine have sizable government liabilities in foreign currency. In contrast, other EM borrowers--Brazil, China, India, South Korea, the Philippines, and South Africa--finance themselves almost exclusively in local currency, giving them greater control over their cost of funding.

g-i: Silver Lining Or A Sign Of Further Tightening?

The difference between an economy's nominal GDP growth (g) and its average cost of financing (i) indicates whether debt to GDP can stabilize during periods when the underlying (primary) fiscal position is in balance. Tables 1 and 2 include our estimates of g-i for 2022. Where nominal GDP growth consistently exceeds average financing costs (India is a good example) governments can operate underlying deficits and still keep debt to GDP ratios flat (if domestic debt is principally denominated in domestic currency). Therefore, a positive g-i can usually be read as supportive of fiscal metrics. But this observation only applies in cases where g-i is not only high, but also stable over an extended period. On the back of geopolitical conflict, G20 central banks' tightening, high and volatile commodity prices, and de-globalization, g-i is anything but stable these days.

So while the first table shows 2022 g-i as positive for all 18 DM central governments, the question is--for how much longer? Over the next few years, as G20 central banks tighten and squeeze inflation out of the system, g-i will have to converge toward zero (or below) in most developed economies. This implies that to stabilize debt to GDP, governments would need to tighten underlying fiscal positions more than they appear willing or able to do at present, particularly given the tendency to subsidize energy in DMs and food and energy in EMs.

EM g-i readings for 2022 (table 2) are also interesting but for different reasons. We have previously flagged India's highly positive g-i calculation as confirmation that it can get away with loose fiscal policy, thanks to consistently high real and nominal growth outcomes, and policies resulting in financial repression (as well as the absence of foreign exchange risk given that essentially all of India's government debt is denominated in domestic currency). However, to the degree that inflation is increasingly a global problem, the g-i factor for 2022 appears to us to be a useful guide as to where EM central banks are in their tightening cycles. Where g-i is 0 (Mexico, Kenya), or negative (Brazil, and South Africa), we would argue that several major EM central banks appear ahead of their DM peers in the tightening cycle, meaning their debt markets are arguably less prone to future rate shocks than those DM peers where nominal GDP growth this year reflects high inflation more so than high real growth.

Stagflation And Its Discontents

Since our report one year ago, the outbreak of war between Russia and Ukraine has overtaken earlier indications that total factor productivity trends had survived the pandemic pretty much intact. Back in May 2021, we thought that while interest rates would inevitably normalize, so too would economic activity, employment, and tax receipts, leading to a gradual improvement in public finances across much of DM and EM. Unfortunately, since then, a far more inflationary scenario has prevailed, and pressure on global supply chains has persisted and in some sectors worsened. As a consequence, creditors are increasingly demanding that sovereigns pay a premium on new borrowing. In the case of EM borrowers, this premium also reflects forward-exchange-rate risk. Under such circumstances, governments would rather have as little debt to refinance as possible; for 22 out of the 37 sovereigns in our survey, both EM and DM, refinancing needs are less than 10% of GDP (10 out of the 18 in our survey). For the rest, however--the increasing number of governments with higher borrowing requirements, facing more volatile rates--the situation is undeniably more challenging than it was a year ago.

Appendix: Our Methodology

Table 1 ranks 18 developed sovereigns according to their vulnerability, over three years, to two new interest-rate shocks: a 100 bps increase in the cost of refinancing maturing debt; and a 300 bps increase. In table 2, we do the same for 19 of the most active rated EM sovereign issuers.

We define budgetary sensitivity as the increase in general government interest spending as a percentage of GDP. For the purposes of this publication, we look only at the gross cost of commercial debt, without considering that the net cost of debt for many sovereigns is even lower, reflecting significant dividend payments to governments from central banks on the back of their interest earnings from government bonds purchased under quantitative easing programs. We exclude official financing from our calculations, and we ignore cash positions. However, faced with a rate shock, most G20 governments would respond by drawing on cash balances, rather than by locking in higher rates. In a few cases where public finances are highly devolved (for example, Nigeria) we use central government budgetary data rather than general government data.

In our scenarios, we assume a shock across the curve rather than a shock that would steepen the curve. That's because, under the latter, the projections of interest costs would be over-reliant on assumptions about how treasuries would manage the maturity of their portfolios in the face of more-expensive longer-dated financing. This decision penalizes issuers with higher short-term debt, such as Japan and the U.S., but, given that our focus is on rollover risk, we see that as unavoidable. This exercise focuses only on first-order effects of higher market rates. It does not consider second-order effects on growth or financial stability.

Table 1

Table 2