Since the pandemic intensified in the spring of 2020, G-20 central banks have aggressively stepped up their purchases of government bonds in an effort to put a floor under sinking demand. This unprecedented monetary activism, alongside rising asset valuations, has raised concerns in the market that central banks may be falling behind the curve on their mission to contain core inflation. This would increase the risk that, by the end of this year or early next, they will need to hike policy rates more than they would have, had they started to tighten earlier. Others argue that their large holdings of government debt mean that central banks can no longer pursue their monetary policy objectives 100%, since they will also need to consider the fiscal implications of higher rates for their government shareholders. These considerations have driven yield curve steepening in the U.S., as well as an uptick in yields in the rest of the world.

S&P Global Ratings' view is that global inflation fears are overblown, and that orderly reflation is a positive development for the world economy. We believe global output will exceed 2019 levels by the middle of next year, as economies find a way to recover lost production and shift output toward new patterns of demand that underpin low inflation. Nevertheless, we are watching the rise in yields in the U.S. and other markets closely.

To better understand the sensitivity of public finances to potential hikes in rates, we have run a series of stress scenarios that measure the first-round effects of a rapid rise in governments' borrowing costs on budgetary deficits (see tables 1 and 2). The results indicate that even under a scenario of a 300 basis points (bps) rise in refinancing rates, 15 out of 18 developed sovereigns and 16 out of 20 emerging-market sovereigns would see their interest expenditure as a share of GDP increase by less than 1 percentage point of GDP relative to our base-case projections through 2023. The majority of sovereigns would pay either the same or less interest as a share of GDP than they did back in 2018. Our conclusion is therefore that the direct fiscal costs of higher rates are--with a few exceptions--manageable for sovereigns. We do not model the second-round effects of higher rates on growth, and this is where rate hikes would start to weigh on public finances. Against a backdrop of weaker GDP performance, pulled lower by stagnating productivity, consolidating very large government deficits would be extremely challenging for sovereigns and our ratings, from both an economic and a political perspective.

Our methodology

Table 1 ranks 18 developed sovereigns according to their vulnerability over a three-year time horizon to two interest rate shocks: 1. a 100 bps increase in the cost of refinancing maturing debt; and 2. a 300 bps increase. In table 2, we do the same for 20 of the most active emerging-market sovereign issuers.

We define budgetary sensitivity as the increase in general government interest spending as a percentage of GDP. For the purposes of this commentary, 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 G-20 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 scenario, we assume a shock across the curve rather than a shock that would steepen the curve. That's because under the latter situation, the projections of interest costs would be over-reliant on assumptions about how treasuries would manage the maturity of their portfolio 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 in this commentary is 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.

The starting point for cost of debt in advanced economies is strikingly low

The good news is that, thanks to expansive central bank balance sheets, developed sovereigns enter this uncertain period with notably low refinancing costs. Fifteen of the 18 largest developed sovereigns are currently rolling over maturing debt at a marginal rate more than 100 bps below their average cost of debt, the exceptions being Japan, the U.S., and Greece. As a consequence, for the vast majority of developed sovereigns, a 100 bps rise to their refinancing rates would represent no shock to public finances at all, at least in its first-round effects.

On the other hand, a 300 bps rate shock to governments' marginal cost of debt would lead to deficit widening via increases in interest expenditure for all 18 developed governments except Denmark (where the effective rate is over 300 bps above the current cost of refinancing). Of the 18 developed-market sovereigns in our report, only three would pay more than 1 percentage point of GDP in interest by 2024, compared with our baseline projections: Japan, the U.S., and Portugal. And during that year only two would pay 3.7% of GDP or more on interest payments (Italy and Japan), reflecting their high stocks of debt. The median interest expenditure to GDP for developed sovereigns under the 300 bps rate shock scenario is 1.8% of GDP. The explanation for this low sensitivity to higher rates is twofold: 1. The low starting point on refinancing rates. 2. Long-dated debt profiles in advanced economies mean that, on average, developed sovereigns are only refinancing debt of 10.5% of GDP per year over the next three years. The exceptions, the U.S. and Japan, have far higher short-term debt that has to be rolled over.

Both Japan and the U.S. are unlikely candidates for a rate shock, in our view. In Japan's case, over 90% of debt is held by domestic residents, including the Bank of Japan (BOJ). Since 2016, the BOJ has implemented a policy of yield curve control under which it has committed to keeping 10-year yields at zero percent. On its face, the U.S. is more vulnerable to higher rates for two reasons: At close to 25% of GDP, short-term debt is high, while, in contrast to other developed markets, refinancing rates are already very close to average rates, so the cost of debt is not as favorable. Nevertheless, we don't think a sustained rate shock of 300 bps would trigger a major fiscal deterioration, at least not through first-round effects. The one-year bill rate is currently 0.06%, and the Treasury would almost certainly further shorten its debt profile if it faced a shock at the longer end of the curve. Moreover, while U.S. dollar holdings are not as dominant as they once were, they still represent close to 60% of allocated reserves globally, and the institutional market for U.S. treasury bills remains the world's deepest.

Eurozone sovereigns including France, Italy, Portugal, and Spain have seen some spread widening over the last month, particularly at longer maturities, though their cost of debt remains an average of 155 bps below their average cost. A 300 bps increase in their refinancing rates would directly weigh on eurozone sovereigns' fiscal performance and reverse what has been a steady decline in interest expenditure to GDP since the European Central Bank (ECB) launched quantitative easing in 2015. Nevertheless, the cost of debt servicing, while increasing under the 300 bps rate shock, would essentially return to levels that prevailed in 2018, and hence, in the context of recovering growth, would not create any undue fiscal pressures. More than anything else, the lower vulnerability of peripheral eurozone sovereigns to rising market rates reflects progress they have made in terming out their debt maturities since the ECB launched quantitative easing in 2015. Since that year, the average maturity of Spain's debt profile has increased an estimated 1.7 years to 8.1 years; Italy's has increased by close to 1.5 years to just under 8 years. And France's has gone from 7.5 years to 9.4 years. As a consequence, debt rollover ratios will average about 10.5% of GDP over the next few years, including maturing short-term debt. Between 2021 and 2023, we estimate that Italy, Portugal, and Spain only need to refinance on average 17.4% of GDP per year, again including short-term rollovers.

Emerging sovereigns also quite resilient

The outcome of the stress test for emerging market countries is less conclusive, reflecting large differences in underlying creditworthiness between the 20 emerging markets included in our survey. Unlike developed sovereigns, which borrow almost exclusively in their own currency, many of the emerging markets in our survey borrow in a mix of domestic and foreign currency. For this reason, we have calculated their refinancing rates as the weighted average on domestic and foreign currency borrowing. These estimates of "blended" domestic and foreign currency rates assume stable exchange rates. Under this assumption, 16 of the 20 sovereigns in the group would see their interest servicing costs increase by less than 1% of GDP by 2023, even under a 300 bps rate hike scenario.

But there are outliers. Given their higher rollover ratios, the five emerging-market sovereigns most vulnerable to a 300 bps rise in refinancing costs would be Egypt (where interest expenditure would rise by 1.2 percentage points (ppts) of GDP over the first year), South Africa (1.3 ppts), Ghana and Kenya (both 0.9 ppts of GDP). Even under our baseline projections, six of the emerging-market governments in table 2 are already paying over 4% of GDP on interest this year. During 2021 Ghana is due to use 56.5% of its tax revenue to finance interest payments, and Egypt, India, and Nigeria are all facing interest costs that exceed 30% of state revenues. At the other extreme, four emerging markets--Chile, China, Korea, and Russia--will pay less interest as a share of GDP this year than most developed sovereigns, equivalent to less than 5% of their revenue.

Complicating their ability to control their cost of financing, Colombia, Egypt, Ghana, Kenya, Turkey, and Ukraine have relatively sizable government liabilities in foreign currency. In contrast, other emerging-market 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. Several emerging-market sovereigns (Hungary, India, Indonesia, the Philippines, Poland, South Africa, and Turkey) have already launched asset purchase programs targeting domestic government securities, either in the primary or secondary market, although the expansion of their central banks' balance sheets has been more restrained than for developed market peers.

What ultimately matters is why rates would rise

The outcome of our stress test is that the first-order effects of rising rates look manageable for nearly all developed sovereign borrowers, and 16 out of 20 emerging-market sovereigns. Four emerging market sovereigns--Egypt, South Africa, Ghana, and Kenya--are already refinancing at market rates above their average cost of debt, face high rollover ratios as a share of GDP, and would see interest costs increase substantially under the 300 bps shock scenario. For the majority of the other sovereigns in both surveys, a 300 bps rate shock to marginal rates would only return interest expenditure to GDP to levels seen in the recent past. This should make central banks more confident that once growth recovers, rate hikes will not upend public finances.

But what ultimately matters is the underlying rationale for rates to rise in the first place. If rates rise quickly to reflect rapid employment gains and buoyant GDP growth, against the backdrop of steady increases in productivity, the higher cost of debt servicing will almost certainly be offset by improving state revenue and more rapid consolidation of the primary (non-interest) government accounts. This is our baseline scenario of transitory inflation, which could indeed prove helpful to fiscal performance. On the other hand, if higher rates come about as a delayed central bank response to runaway inflation caused by stagnating post-pandemic productivity, then the interest rate shocks might be even stronger than the ones we present here, growth would falter, exchange rates weaken, and credit fundamentals suffer.

Table 1

Table 2