Leonardo D’Amico, Francesco Giavazzi, Veronica Guerrieri, Guido Lorenzoni, Charles-Henri Weymuller 14 January 2022
The severity of the Covid-19 pandemic and its adverse effects on the economy have induced Europe to suspend the rules of the Stability and Growth Pact by activating the general escape clause. In January 2023, the escape clause will expire and, if nothing changes, the old rules will be back in place, possibly forcing painful fiscal adjustments in countries that are already struggling. Over the last few years, there has been a growing consensus that the current rules are outdated, too complicated, and not countercyclical enough. At the same time, there have been positive developments in the ability of Europe to pursue investment to address long-term challenges, such as the green transition and digitalisation process, through the EU recover plan NextGenerationEU (NGEU).
In light of these developments, in Giavazzi et al. (2021) we offer a proposal to strengthen the European fiscal framework based on two elements: a revision of the fiscal rules, and a plan to create a European Debt Agency to absorb the debt accumulated during the pandemic. It is important to think about these two elements in conjunction, but for exposition we divide our discussion here in two parts. In this first column, we discuss the fiscal rules side; in a second column, we discuss the debt management side. The overarching theme is that for the EU to live up to its ambitions, it is necessary to rely on the capacity of the Union to act as an effective central authority that can exercise effective supervision on member states finances and, at the same time, can take advantage of its enforcement powers to create additional fiscal space for all members and use it wisely.
A medium-term debt target
The main objective of a fiscal rule is to define a predictable path for public finances to ensure that investors are willing to purchase public debt, provide a reliable source of funds when the need arises, and avoid financial instability.
Let us begin with a compact description of the rule and then discuss its various elements, their pros and cons, and how they compare with other approaches.
The rule we propose is to set a ceiling on the growth rate of primary spending, with the ceiling revised over intervals of three years. The growth rate of spending is chosen so that, if followed over a ten-year horizon, under realistic macroeconomic scenarios, it would yield the following average reduction in the debt-to-GDP ratio dt:
The equation defines our medium-term debt target. Computing the expression on the right requires dividing the current level of debt-to-GDP in two parts dt = dF,t + dS,t, and applying different speeds of adjustment to the two parts. The part dF,t is fast-adjusting, the part dS,t is slow-adjusting. The coefficients β and γ determine the speed of adjustment, with β chosen larger than γ. The constant d* is the long-run debt target.
We define the slow-adjusting portion of debt as the sum of two components. First, it includes debt accumulated in response to ‘crises’, defined as periods when the EU general escape clause is active and deficits are needed to mitigate the adverse effects of a severe recession. Second, it includes the debt accumulated to finance ‘spending for the future’, which includes public investment that is beneficial for the long-run growth prospects of the country and expenditures that contribute to some European public goods that will benefit future generations. The fast-adjusting portion of debt is the residual stock of debt. A delicate issue is how to decide on categorising different types of spending as spending for the future. We will return to this below.
Let us be clear that the distinction between slow-adjusting and fast-adjusting debt is just for the purpose of computing the overall speed of adjustment. We do not propose issuing different types of debt – there is only one type of debt, and the two components are just accounting terms to determine how fast the country must adjust.
Both the choice of a medium-run debt target and of a spending rule as an instrument are present in different combinations in several proposals circulated in recent years (Andrle et al. 2015, Bénassy-Quéré et al. 2018, Darvas et al. 2018, European Fiscal Board 2018, 2019, 2020, Martin et al. 2021).1 The main difference is the two-speed component that comes in combination with the golden rule scheme discussed below. Before we get into that, let us discuss the general merits of a medium-term debt target.
Why choose a rule that explicitly targets a debt reduction objective?
The main reason is that keeping the debt-to-GDP ratio stationary is a basic property that ensures that investors are willing to absorb the debt issued by a given country and treat it as safe debt, under fairly general conditions (Bohn 1991, 1998).
An additional benefit of aiming for an objective of debt reduction is that it automatically responds to changes in the macroeconomic environment, namely, to changes in the real interest rate r and in the output growth rate g. Consider the law of motion of government debt and note that, for a given objective of reduction of the debt-to-GDP ratio in a given year, a lower value of r-g translates immediately into more space for a larger primary deficit.2
Why choose a ten-year target?
The main advantage of averaging over a ten-year period is that it helps to smooth the fiscal effort required to reach a certain amount of debt reduction. Numerical simulations show that increasing the horizon used for the projections – say, from 10 to 15 or 20 years – and keeping all else equal implies a smaller adjustment in the primary balance required in the initial year. The reason is that when we do projections with a constant growth rate of spending, primary surpluses tend to be increasing over time. Then, with a longer horizon, we end up averaging higher primary surpluses in future years with the current one. In other words, there is more back-loading of the fiscal effort, which allows for a more gradual adjustment.3
The benefits of smoothing are especially important for a country coming out of a recession causing revenue shortfalls, as a relatively long horizon prevents premature fiscal adjustments. On the other hand, for a country experiencing a period of temporarily booming revenues, the fact that the rule requires averaging over time only allows a gradual increase in spending. In sum, the smoothing properties of the medium-term target have counter-cyclical benefits.
Given the benefits highlighted above, why not use an even longer horizon?
For example, simulations of a suggested rule by the European Fiscal Board (2000) consider medium-term rules with a horizon of 15 or 20 years (see their Figure 5.11). The main downside of making the horizon too long is that we may rely too much on more uncertain forecasts farther into the future.
Why not make the spending rule simpler by not relying at all on forecasts and just following a rule like in Claeys et al. (2016), with the growth of public spending set equal to potential growth minus a mechanical debt correction term?
While we value simplicity, we think that modelling and forecasting are unavoidable parts of fiscal planning. The radical proposal of Blanchard et al. (2020) advocates dropping numerical rules altogether in favour of stochastic debt sustainability models, precisely on the grounds that each country is different and has different prospects at any point in time. While we prefer a rules-based approach, we believe that medium-term forecasts, with all their limitations, are a good way of tailoring a rule to countries’ specific situations.
Don’t existing rules already provide enough flexibility?
A very different objection could be that the current system of rules does partially address the trade-off between debt adjustment and fiscal smoothing in the part where compliance with the Medium-Term Budgetary Objective (MTO) is assessed, whereby the Commission can consider a country compliant as long as a certain amount of fiscal effort is made in the direction of satisfying the MTO. We think, however, that a more transparent approach, in which the smoothing is explicitly part of the way in which budget projections are evaluated, is preferable. Furthermore, the way in which the output gap is used in existing rules is widely acknowledged to give very imperfect countercyclical properties.
How to choose the parameters of the rule?
The choice of parameters is hard and, in particular, the choice of the long-run target d* is especially fraught. The value of 60% in the existing legislation is a conventional number that came from basically looking at an average across EU members in the late 1980s (Buti and Gaspar 2021). There are good economic reasons for revising that number, as low interest rates make higher debt levels sustainable for a given primary surplus. Some proposals (Martin et al. 2021) argue for country-specific debt objectives. While we sympathise with the economic argument behind those proposals, we see some risks in the exercise of computing these country-specific targets, especially insofar as they require identifying some maximum safe level of debt that requires taking a stance on the maximum politically sustainable primary surplus – something extremely hard to quantify – and can end up focusing market expectations on risky thresholds.4 A simpler, more practical approach seems to be the European Stability Mechanism’s recent proposal of just switching to 100% (Francová et al. 2021). In our numerical experiments, we have stayed with 60%, to minimise the need to revise the existing legislation, and have found speed parameters that yield reasonable adjustments. Moving to 100% would just make it easier to find acceptable speed parameters.
Turning to the choice of speed parameters, β and γ, our approach is to experiment with different values and look at simulations under different scenarios. In a baseline calibration, we have experimented with β = 0.05 and γ = 0.02, obtaining reasonable paths of debt reduction for highly indebted countries, which would be compatible with current budgetary plans.
Does this rule provide sufficient counter-cyclicality?
The elements of the rule that provide counter-cyclicality are (1) the fiscal smoothing implicit in the medium-term target, which we discussed above; (2) the three-year revision of spending growth paths; and (3) the fact that increases in debt during recessions go in the slow-adjusting component of debt. Nonetheless, no numerical rule is perfect, and it is possible that in some circumstances a country may be required an excessive fiscal effort. For this reason, we propose leaving room for a formal procedure by which a country can ask for a reduced speed of debt adjustment. Of course, the general escape clause can also be used.
An important part of the proposal is the presence of a golden rule that favours ‘spending for the future’ in two ways. First, some categories of spending are not subject to the spending ceiling on a flow basis. Second, the same categories of spending contribute to the computation of the slow-adjusting portion of debt. The second element is crucial, as it ensures that the flow exception does not translate one-to-one into a tighter constraint for other forms of spending. Some constraining element is present, as the speed γ is not zero, reflecting the overall objective of debt stabilisation. But it is weaker than in a system where the speed adjustment is absent.
Why not include only investment spending that explicitly raises future revenues in the golden rule?
Our view is that other forms of spending also have future benefits, although benefits that take more the form of insurance against future disasters. In an ideal system, a separate form of cost-benefit analysis would be done for this type of investment, and favoured-speed debt financing would be granted to different degrees. In fact, the same ideal system would allow for multiple speeds for any investment spending, depending on its long-term benefits. Choosing to have only two speeds is just to reduce the complexity of the system.
Is it not better to finance the costs of the transition with dedicated sources of fiscal revenue (e.g. a carbon tax)?
While finding dedicated sources of EU fiscal revenue is certainly desirable, we think it is likely that the fiscal resources needed for investment and to help in the short run the segments of the economy hit hardest by economic restructuring will exceed the additional revenues. We also think that it is correct to think of the green transition as an inter-temporal, inter-generational choice, for which partial debt financing is appropriate.
Won’t it be easy for countries to game the system and easily relabel as spending for the future all kind of programmes of dubious quality?
On this dimension, it is comforting that the recent experience of NGEU has so far been fairly successful. The Commission has specified clearly defined objectives, the spending plans submitted by governments have been thoroughly vetted, and the use of ‘milestones and targets’ has ensured a high degree of accountability. This is a useful blueprint that can be used to build a similar structure to define and monitor spending for the future.
Authors’ note: The debt management proposal grew out of conversations with Massimo Rostagno and we thank him deeply for that. The questions addressed in this column have been stimulated by useful comments and criticisms by Olivier Blanchard, Laurence Boone, Massimo Bordignon, Carlo Favero, Chiara Goretti, Silvia Merler, Gilles Moec, Jean Pisani-Ferry, Lucrezia Reichlin, Shahin Vallée, and Charles Wyplosz and we thank them all. The views expressed in this column are the responsibility of the authors and do not necessarily reflect the position of the French and Italian authorities.
Andrle, M, J Bluedorn, L Eyraud, T Kinda, P Koeva Brooks, G Schwartz, and Anke Weber (2015), “Reforming Fiscal Governance in the European Union”, IMF Staff Discussion Notes.
Bénassy-Quéré, A, M Brunnermeier, H Enderlein, E Farhi, M Fratzscher, C Fuest, P-O Gourinchas, P Martin, J Pisani-Ferry, H Rey, I Schnabel, N Véron, B Weder di Mauro, and J Zettelmeyer (2018), “Reconciling risk sharing with market discipline: A constructive approach to euro area reform”, CEPR Policy Insight No. 91.
Blanchard, O J, A Leandro, and J Zettelmeyer (2020), “Redesigning the EU Fiscal Rules: From Rules to Standards”, paper presented at the 72nd Economic Policy Panel Meeting.
Bohn, H (1991), “Budget balance through revenue or spending adjustments? Some historical evidence for the United States”, Journal of Monetary Economics 27(3): 333–359.
Bohn, H (1998), “The Behavior of U.S. Public Debt and Deficits”, Quarterly Journal of Economics 113(3): 949–963.
Buti, M and V Gaspar (2021), “Maastricht values”, VoxEU.org, 8 July.
Claeys, G, Z Darvas, and A Leandro (2016), “A proposal to revive the European Fiscal Framework”, Bruegel Policy Contribution.
Darvas, Z, P Martin, and X Ragot (2018), “European Fiscal Rules Require a Major Overhaul”, Conseil d’Analyse Économique.
European Fiscal Board (2018), Annual Report.
European Fiscal Board (2019), “Assessment of EU fiscal rules, with a focus on the six and two-pack legislation”.
European Fiscal Board (2020), Annual Report.
Francová, O, E Hitaj, J Goossen, R Kraemer, A Lenarčič and G Palaiodimos (2021), “EU fiscal rules: reform considerations”, ESM Discussion Paper 17
Giavazzi, F, V Guerrieri, G Lorenzoni, and C-H Weymuller (2021), “Revising the European Fiscal Framework.”
Martin, P, J Pisani-Ferry, and X Ragot (2021), “Reforming the European Fiscal Framework”, Conseil d’Analyse Economique.
1 Maduro et al. (2021) give a detailed analysis of the legal avenues by which reform efforts can be pursued.
2 This follows immediately from:
Blanchard (2021) has been making a forceful case of taking into account secular changes in r – g in thinking about fiscal adjustment, although he reaches different conclusions on the ideal shape of a fiscal rule.
3 Numerical simulations in a deterministic environment show that a multi-year debt-reduction target rule is approximately equivalent to a simple inertial rule for spending xt of the type xt = ρxt-1 – δdt for some values of the coefficients ρ and δ.
4 Whereas our medium-term approach implicitly allows for some degree of country-specific differentiation, as argued above, without focusing on the maximum debt level.