“Italy is a country that manages to have a growth rate that is below the interest rate, no matter how low the interest rate is.” Alberto Alesina, September 20, 2019.
“[D]ebt does not avoid hard choices. It only delays them.” N. Gregory Mankiw, The New York Times, March 26, 2011.
In the battle against the economic impact of COVID-19, governments around the world are pulling out all the stops. In advanced economies, leading central banks have pushed interest rates to zero or below. And, a recent IMF estimate puts the combination of discretionary spending and automatic fiscal stabilizers (including unemployment insurance and progressive income taxation) at $9 trillion―more than 10 percent of global GDP.
The limits to monetary policy are clear. With nominal long-term yields extremely low or negative and central bank balance sheets large and growing, tools like forward guidance and yield curve control have diminished potential to provide further economic stimulus (see our pre-COVID post). Fortunately, the prospect of persistently low interest rates amplifies the stabilizing impact of fiscal expansion, as there is little risk of “crowding out” private investment that occurs in normal times (see, for example, Eichenbaum).
How large is the scope for additional countercyclical fiscal policy? With sovereign yields so low, the cost of additional financial expansion looks to be minimal, at least for now (see, for example, Blanchard). And, as we will discuss, there are good reasons to expect the interest rate on the debt (r) to remain low relative to the economy’s growth rate (g), making it easier for governments to cap the surge of the debt.
Nevertheless, each time public debt-to-GDP ratios ratchet higher—as they did in the 2007-09 crisis and are now doing again—the question of “fiscal space” reemerges. Will governments have the budgetary capacity to respond to the next economic shock—regardless of its source—and provide relief and stimulus on the scale required to meet society’s needs?
In this post, we highlight recent fiscal developments in advanced economies, and review the factors affecting the sustainability of their high and rising levels of debt. To foreshadow our conclusion, the fact that many countries’ fiscal positions were precarious even before the COVID crisis does not weaken the current case for stimulus. But, doubts about fiscal space are growing. So, it is important that governments find a way to make a credible commitment to future fiscal consolidation when their economies have returned to full employment. Failure to do so could threaten confidence both in government finances and in economic performance.
We start by reviewing the COVID-related fiscal facts. In the chart below, the gray bars show the latest forecasts for the general government primary fiscal balance (the budget balance excluding interest payments) as a share of GDP for selected advanced economies in 2020 taken from the April 2020 IMF Fiscal Monitor. The black diamonds are the estimates for the same period, but from the IMF’s October 2019 Fiscal Monitor. For the G-20 advanced economies, in just six months, the projected fiscal balance for 2020 deteriorated by 8.1 percentage points of GDP to -10.1 percent. This is almost surely the most rapid peacetime decline on record. The April projected U.S. balance of -13.5 percent is both the largest among advanced economies and the biggest since 1945.
General government primary fiscal balance projection (2020, percent of GDP) and the average interest rate (r) – growth rate (g) differential (1870-2015)
Source: IMF Fiscal Monitor for April 2020 and October 2019, and authors’ calculations based on Tables 11 and 12 in Jordà et al (2019).
Naturally, this rapid fiscal deterioration is driving government debt upward. In the chart below, the gray bars depict the April IMF projection for 2020 general government gross debt as a share of GDP, while the red diamonds show the projections for this year made in October 2019. For the G-20 advanced economies as a whole, the projected debt ratio for 2020 increased by 19 percentage points to 132 percent of GDP. And, high deficits in future years probably will add to this increase. For example, the Committee for a Responsible Federal Budget projects that U.S. federal debt alone will rise gradually to over 110% of GDP in coming years. (In some cases, like Japan, net debt excluding government holdings of financial assets is notably lower. But, gross debt is more useful as an indicator of rollover needs.)
General government debt projections for 2020 (percent of GDP), April 2020 vs. October 2019
Source: IMF Fiscal Monitor for April 2020 and October 2019.
Nearly 40 years ago, Sargent and Wallace taught us that governments can issue public debt up to a certain threshold. Beyond that real (inflation-adjusted) limit, the consequences are either outright default or, if the debt is in domestic currency bonds that the central bank can acquire, a partial default in the form of inflation.
Unfortunately, we do not know how close we are to such a debt limit. There is no magic number for all countries at all times. Indeed, for nearly a decade, Japanese government bond yields have remained below one percent as the gross debt-to-GDP ratio rose well above 200 percent. By comparison, at a debt ratio under 50 percent of GDP, Argentina defaulted in 2001. Even with today’s high and rising public debt ratios, long-term advanced-economy bond yields are historically low, suggesting that investors generally are not concerned about default risk on what is widely viewed as “safe debt.” But, as we know from costly experience (like that of the euro area in 2010-12), investor attitudes can shift rapidly driving up market risk premia and making rollover of existing debt costly or impossible.
By making judgments about the largest primary balance a government could sustain in the presence of “fiscal fatigue,” researchers try to estimate debt limits (see, for example, Table 3 in Ostry et al). And, yet, as the IMF’s practices show, forming a view about a nation’s debt limit requires taking account of a wide range of factors. The list includes the government’s evolving fiscal reaction function (e.g. to what extent will the government raise its primary balance when interest costs rise); the rollover schedule of the debt; the scale and form of contingent liabilities (including the potential need to recapitalize banks or to make up for shortfalls in public pension or medical insurance schemes); the country’s external position; and the prospects for long-run economic growth (see the opening quote from Alesina).
While we do not know the trigger for a sovereign debt crisis, we do know that the lower the interest rate on the debt (r) relative to the economy’s growth rate (g), the smaller the primary balance needed to make public debt sustainable. In our primer, we derive the debt-sustainability condition from the government’s budget constraint. This weak standard requires that the primary balance (s) as a ratio to GDP be at least as large as the stock of outstanding sovereign debt (b) times the difference between the nominal interest rate the government has to pay (r) and the rate of growth of nominal GDP (g). If it is not, then the ratio of debt to GDP explodes.
That is, sustainability requires that
s ≥ (r-g) x b
A useful way to interpret this condition is that the right-hand-side coefficient (r-g) is the market risk premium on the outstanding debt (b). So long as the primary balance (s) is sufficient to cover this risk premium times the level of debt, debt remains in check. A government balance larger than this means debt is declining relative to GDP. More importantly, with a smaller balance, the ratio of debt to GDP rises—eventually leading to default or fiscal dominance and inflation (see our primer on helicopter money).
So, to figure out how much fiscal space a country has, the first step is to estimate r-g. It is common to assume that sovereign interest rates exceed aggregate growth rates, so r-g is positive. However, recent empirical work on advanced economies comes to the opposite conclusion. It appears that the long-run average for r-g is most likely negative, or at most only modestly above zero. Using the Jordà-Schularick-Taylor Macrohistory Database for the period from 1870 to 2015, we find that r-g was less than zero more than half the time for 14 of the 15 advanced countries in our charts above (Austria is not in the database). Similarly, in a broader data set covering 55 countries (including emerging economies) since 1800, Mauro and Zhou report that r-g was below zero in 61 percent of the advanced economy observations and in 75 percent of the emerging economy observations (see their Figure 2).
Perhaps most important, estimates of the common component of r-g across advanced economies in the latest IMF World Economic Outlook (Box 2.2) suggest not only that the gap is usually negative, but that it is likely to stay there (see the chart below). Indeed, we can think of a number of reasons to expect r-g to remain low for some time in the aftermath of the COVID crisis. Above all, the desire to pay down debt and build precautionary buffers will almost surely boost the propensity to save among households, firms and governments (the U.S. household saving rate reached a record 33 percent in April). In addition, the higher cost of doing business—combined with a plunge in the value of vintage capital (like skyscrapers, cruise ships and public transport)—in the post-COVID world could diminish incentives to invest. The same factors could slow productivity growth.
Advanced economies: estimates of the common component of the interest rate (r) – growth rate (g) gap, 1960-2030
Note: Gray-shaded area denotes 95-percent confidence range for the forecast (dotted red line).
Source: April 2020 IMF World Economic Outlook, Chapter 2, Figure 2.2.1.
If we are correct, it will be much easier to keep debt ratios from rising after the initial fiscal burst. To see why, consider the following table. Each cell reports the minimum primary balance necessary to steady the debt-to-GDP ratio at each combination of the debt ratio (the columns) and the r-g risk premium (the rows). So, for example, if r-g averages about -3 percent (in line with the IMF’s estimated common component of advanced economy r-g shown by the dotted black line in the chart above), then a primary balance of -7.5 percent of GDP is sufficient to stabilize a debt ratio of 250 percent of GDP. Looking back at the first two charts, this is almost precisely the 2020 primary balance-debt ratio mix that we currently observe in Japan. In contrast, the United States still needs to raise its 2020 primary balance by about 9 percentage points of GDP (to -4.5%) in order to stabilize its debt ratio at 150 percent.
Steady-state primary balance (percent of GDP) associated with different debt ratios and market risk premia
Source: Authors’ calculations.
With r-g so low, it looks as if most advanced economies’ debt levels are manageable. In fact, some countries still have considerable space for maneuver. But, counting on r-g to remain low is risky. First, highly indebted countries are on a knife-edge that is sensitive to the level of interest rates: if sovereign interest rates (r) rise, the fiscal contraction needed to steady the debt would be brutal. To see how severe austerity could be, consider what would happen in Japan if r-g were to rise from the long-run average of about -3 to 0. In that case, Japan would need to tighten its 2020 primary balance by a whopping 7.5 percentage points of GDP! For comparison, with a debt of less than 70 percent of GDP, Germany would need to tighten by less than 2 percentage points.
Second, while we are optimistic about r-g for the next few years, history counsels caution. Indeed, the standard deviation of r-g across advanced economies in the Macrohistory Database ranges from 7 to 14 percentage points. For the most part, these changes in r-g are associated with changes in r, rather than g (see Table 3 here). And, perhaps most important, Mauro and Zhou warn that it is the marginal interest cost, not the average interest cost measured by r, that typically poses rollover challenges to highly indebted governments. They report that marginal interest rates lead defaults by about two months (leaving virtually no time for a policy adjustment), that public debt ratios appear to rise (and primary balances to fall) in the year prior to default, and that average interest-growth differentials do not lead at all.
So, what to do? Considerations about long-run debt sustainability do not limit our support for the massive effort governments are making in the face of the pandemic. Fiscal policy is clearly the most effective stabilization tool at their disposal. Even with massive spending (and lending) designed to limit the short-run economic damage, the unprecedented COVID shock threatens to slow long-run growth (see, for example, Kozlowski, Veldkamp and Venkateswaran). But, without near-term fiscal support, the long-run damage could be far greater.
However, with few exceptions (like Germany), advanced economy policymakers did little in the “high employment period” following the global financial crisis (or the euro-area crisis) to create fiscal space. They refused to make the “hard choices” to which Mankiw refers in the opening citation. Instead, over the two years prior to the COVID shock, the federal government in the United States implemented the largest pro-cyclical fiscal expansion since the 1960s. Even when enacted, these policy actions looked very ill timed.
The lesson now is simple. As they manage the COVID crisis, governments should be thinking about frameworks that make consolidation the norm in good times, and that shine a very harsh light on any pro-cyclical abuses of fiscal space. Put differently, if we are to use fiscal space most effectively, supporting employment and growth during times of stress, we need both rules-based fiscal stimulus in recessions and rules-based consolidation in expansions.
Otherwise, when the next crisis hits, investors’ trust in advanced economy fiscal policymakers may not be sufficient to finance the massive deficits needed to speed the return to full employment.