April 5, 2019
Public debt and low interest rates or rather the high cost of public debt?
The view that deficits and higher public debt can be beneficial has received an important boost through the recent presidential address of Prof. Olivier Blanchard at the American Economic Association, entitled: “Public debt and low interest rates”. A new editorial by Daniel Gros.
The view that deficits and higher public debt can be beneficial has received an important boost through the recent presidential address of Prof. Olivier Blanchard at the American Economic Association, entitled: “Public debt and low interest rates”. He argues that an increase in government debt, for example through higher expenditure today, might be welfare enhancing if the growth rate of the economy is higher than the interest rate. The exact result depend considerably on the production technology, but a key condition is always the difference between the growth rate and the interest rate.
Blanchard also acknowledges that higher debt leads to higher risks and that experience shows that that the (market) interest rate on public debt, depends also on the debt level (as a % of GDP). He refers to a rule of thumb used by the IM that the risk premium (i.e. the difference between the riskless rate and the interest rate on public debt of any particular country) increases by 3- 4 basis points for every percentage increase in the debt to GDP ratio above 60 %.  This link between debt and risk premia is particularly important for the euro area where each government is responsible individually for its own debt. Recent research by the European Commission suggests that each increase in the debt to GDP ratio might lead to an increase in the risk premium close to 5 basis points.
A key implication of the empirical regularity that the risk premium depends on the debt level is that the marginal cost of public debt is much higher than the interest rate on public debt (which represents the average cost). The marginal cost is of course not directly visible and has little impact on political discussions, but it is very real. The high marginal cost arises from the fact that the higher interest rate applies not only to the additional debt, but to all of the stock of debt which has to be refinanced.
This difference between the average and the marginal cost of debt becomes particularly relevant for high debt countries, like Italy. At present it appears that the country has settled into a ‘new normal’ under which the risk premium fluctuates around 250 basis points. This is also approximately the rate on BTPs since the rate on long term German debt is close to zero. At an interest rate of 2.5 % the % Blanchard condition might be (borderline) satisfied for Italy if one assumes a growth potential of 0.5-1 % and an expected inflation rate of 2 %.
However, this perspective the present situation seems (borderline) sustainable hides the true marginal cost of public debt for Italy, which is much higher than the 250 basis points on BTPs.
The best way to illustrate the true cost of public debt is to compare the different paths of Italy and Portugal. Both countries have a debt level of about 130 % of GDP and both have grown very little over the last years. But even though Portugal remains much poorer than Italy (a factor which usually would militate for a higher risk premium) its risk premium round 150 basis points much lower than Italy. The reason for this difference is the different outlook: In Portugal the deficit is so low that the debt ratio is projected to decline rapidly, reaching approximately 100 % of GDP by 2023, according to the IMF. By contrast, the debt ratio of Italy is projected to remain approximately at its current level of close to 130 % of GDP.
By 2023 Portugal might thus pay only 150 basis points on 100 % of GDD, giving a total interest expense of 1.5 % of GDP, whereas the cost of debt for Italy might be 250 basis points on 130 % of GDP, giving a total of 3.25 % of GDP, more than twice as much as Portugal. The difference between these two scenarios is 1.75 % of GDP, and this difference should be related to the difference in the expected debt ratio of 30 % of GDP (130 versus 100, both in 2023). A difference in the debt ratio of 30 points can thus lead to a difference in the debt service ratio of 1.75 %, or about 30 billion euro per annum. This is equivalent to an implicit marginal cost of public debt of almost 6 %. This is much higher than any growth rate Italy could hope for.
The conclusion is clear: at a high level of debt the true cost of public debt is much higher than one would suspect looking only at interest rates. The present situation of Italy might look sustainable, and might be tempting to follow the advice of Blanchard to spend more today. But for Italy this temptation needs to be resisted: the implicit cost of increasing current spending instead of reducing public debt to safer levels is very high.
 This risk premium refers in general to long-term (i.e. usually 10 year) debt.
 The exact number would be 1.75/0.3 = 5.8 %.