Box 1.4. Can Countries Sustain Higher Levels of Public Debt?
The global decline in interest rates over the past Figure 1.4.1. InterestGrowth Rate three decades has dramatically reduced sovereign bor- Differentials in Advanced Economies, rowing costs in many countries. Some commentators 19902016 have argued that in this environment, governments (Percent) can sustain higher levels of public debt, particularly in Interestgrowth rate differentials have been generally advanced economies (Furman 2016; OECD 2016b; declining over the past 25 years or so but remain volatile. Buti and Carnot 2016). The argument is simple: lower France Germany Japan interest rates reduce the cost of debt service, so govern- 10.0 United Kingdom United States ments can afford to borrow more. This box examines an extended version of this argu- 8.0 ment: that debt sustainability is determined not by the interest rate alone, but by the differential between 6.0 the interest and growth rates. A smaller differential implies that the debt ratio increases more slowly (if the 4.0 differential is positive) or decreases more quickly (if the differential is negative) for a given level of the pri- mary balance, hence allowing a government to sustain 2.0 a higher debt ratio without the need for tighter fiscal policy. As the results in this box show, what matters 0.0 most for debt sustainability is not the short-term evo- lution of the differential, which reflects cyclical factors, 2.0 but its longer-term structural level. Figure1.4.1 shows the difference between the effec- 4.0 tive interest rate on government debt and the rate of 1990 94 98 02 06 10 14 92 96 2000 04 08 12 16 nominal growth since 1990 for a sample of advanced economies.1 During this time, there has been a marked downward trend in the interestgrowth rate differen- Source: IMF staff estimates. Note: The interest rate is computed as interest payments tial; even though interest and growth rates have both divided by outstanding debt at the end of previous year. declined, the interest rate has fallen further than the growth rate. Rather than its being a recent phenom- enon, declines in the past five years (reflecting higher growth rather than lower interest rates) are simply the a result, future interestgrowth rate differentials are continuation of this trend. likely to be lower than they were on average in the What might have driven the persistent decline in past decade. This box reports on two experiments the interestgrowth rate differential? The likely causes to assess the impact of a transitory and permanent are structural. For example, this pattern would arise decline in the interestgrowth rate differential on if expectations about nominal growth took time to sustainable debt levels. The analytical framework, adjust to the lower rates seen in the 1990s and 2000s. which is an extension of the work done by Ghosh and Likewise, a worldwide reduction in safe assets or others (2013), produces a debt limitthe maximum decreasing global risk appetite would also have pushed debt level before defaultfor each country in the down the interest rate on government bonds. And sample. An important feature of this approach is that demographic changes may have increased the demand the evolution of the interestgrowth rate differential is for savings instruments, reducing the compensation partly unpredictable. In technical terms, the differen- governments must offer to public debt holders. tial follows a persistent stochastic process.2 The model Given the structural nature of these factors, this is calibrated to important aspects of public finance trend is unlikely to be reversed in the near term. As data for seven countries: Canada, France, Germany, Japan, the Netherlands, the United Kingdom, and the
1Ongoing structural changes in emerging market economies 2Further details of this framework are discussed by Barrett
make it harder to identify similar trends there. (forthcoming).
International Monetary Fund | April 2017 33
FISCAL MONITOR: Achieving More with Less
Box 1.4 (continued)
United States. The process governing the evolution of decline in the interestgrowth rate differential is much the interestgrowth rate differentials is estimated from larger than that for a transitory decline. A permanent the data since the early 1990s. decline of 1 percentage point increases the maximum Transitory decline in the interestgrowth rate differen- sustainable debt level by an average of 25 percent tial. This experiment simulates the impact the recent of GDP in the sample. Across countries, this figure decline in the interestgrowth rate differential has ranges from a low of 10 percent of GDP to a high of had on the debt limit for each country in the sample 40 percent of GDP. Of course, in reality, it is difficult by assuming that the observed decline is a draw from to assess whether the decline in the interestgrowth the estimated distribution of the interestgrowth rate rate differential is transitory or permanent. But even if distribution. Specifically, the model-generated debt only a portion of the decline is permanent, the impact limit in 2012 is compared to that consistent with the on debt limits is likely to be large. World Economic Outlook forecast for 2022 (to allow for The intuitive explanation for the larger sensitivity the dissipation of expected monetary policy changes). to structural changes is that public debt issued today Between these two points, the interestgrowth rate is rolled over and repaid over long periods of time. differential is forecast to fall by an average of 1.6 Thus, the sustainability of debt is driven principally percentage points for the countries in the sample. The by future interestgrowth rate differentials, which results of this experiment are quite small. They suggest ultimately depend on the shape of the distribution. that this fall could increase debt limits by about 2 The exact results also depend on the simplifying percent of GDP, on average. assumptions of the model, including that debt is short Permanent decline in the interestgrowth rate differ- term, growth is exogenous, and shocks to the surpluses ential. This experiment assumes that the decline in the are uncorrelated with growth. However, the results are interestgrowth rate differential is permanent. This is robust to various estimation periods of the process for implemented by shifting to the left the distribution of the interestgrowth rate differential and are of similar the differential by 1 percentage point. The key finding magnitude to those found in other studies, such as is that the sensitivity of debt limits to a permanent OECD 2016b.