The COVID-19 crisis has led to a series of unprecedented fiscal interventions by governments worldwide (Baldwin and Weder di Mauro 2020). Even excluding conditional expenditures—such as loan guarantees, which may or may not be drawn upon—this additional public sector spending will nevertheless increase deficits by between a few to as much as 20 percentage points of GDP (Figure 1).1 Hence, while widely regarded as important and necessary, the implication of such massive increases in fiscal expenditures is that public debt is set to rise substantially in the near future.2
Figure 1 COVID-19 fiscal expenditure packages (selected G7, BRIC, and ASEAN economies, 2020)
Source: Author's calculations, from national authorities and IMF.
Notes: Spending only include expenditures at the general government/federal level and excludes local/state spending. Household share includes income supplements, food safety net, and unemployment and other social security payouts; firm share includes loans, grants, job support, and tax holidays (including to the self-employed sector), but excludes guarantees; government share includes healthcare system, transfers to state/local authorities, and international assistance. GDP data rely on latest year available (usually 2019). Grey bar indicates detailed breakdown not available.
Moreover, the increases in the government debt stock only constitute a part of the anticipated increase in the debt burden that each economy will need to confront. Households that face sharp declines in income, perhaps from involuntary unemployment, are likely to seek to smooth consumption by either running down savings or taking on debt. Firms—both cash-strapped small and medium enterprises, as well as larger corporations operating on thin margins—have experienced sudden, unprecedented collapses in demand, and will likely resort to seeking bank credit or, if possible, issuing debt.
This expected increase in private debt will compound the build-up of public debt. Going into the COVID-19 crisis, the total debt load in advanced and emerging economies was already significant. By the third quarter of 2019, total debt in the amounted to 246% of GDP in the former, and 226% in the latter. Globally, total debt was 235% of output (see Figure 2)3 Just as important, while rising public debt has been a bigger problem in emerging markets, developed markets have instead experienced pressure from rising private debt, emanating primarily from bloated corporate balance sheets. This combined burden may prove to be a challenge not just for post-lockdown economic recovery, but also growth prospects for the medium run.
Figure 2 Public and private debts
Panel A: Advanced economies, 1960Q1-2019Q3
Source: Author's calculations, from BIS.
Notes: Credit to private nonfinancial (series Q:..:P:A:M:XDC:A) and general government (series Q:..:G:A:N:XDC:A) from all sectors , market value for private, and nominal value for public, adjusted for breaks. Discrete jump at 1997Q4 is partially an artefact of the data, representing the inclusion of Japanese government debt into the database, exacerbated by the 1997 Asian crisis.
Panel B: Emerging economies, 1965Q1-2019Q3
Source: Author's calculations, from BIS.
Notes: Credit to private nonfinancial (series Q:..:P:A:M:XDC:A) and general government (series Q:..:G:A:N:XDC:A) from all sectors , market value for private, and nominal value for public (except Korea), adjusted for breaks. Jump at 1991Q1 is an artefact of the data, representing the inclusion of Russia into the database.
Some relevant theory
Economic theory has long implied that properly accounting for the effects of debt on growth requires working with changes in the aggregate debt stock, rather than either the public or private debt alone. At the most basic level, Ricardian equivalence arguments (Barro 1974) would imply substitutability between government debt and the future tax burden, which in turn detracts from saving available for allocation into private assets. If such government debt financing is directed toward consumption rather than investment, there would be a concomitant reduction in capital formation, and hence growth.
More articulated models make analogous points. On the supply side, straightforward extensions of the canonical Ramsey-Cass-Koopman model (e.g. Blanchard and Fischer 1989: Chapter 2) reconcile the distinct public and private components under a single intertemporal national resource constraint; since both forms of debt enter into the first-order conditions that define optimal growth, so long as taxation is distortionary, the balanced growth path will be altered as well. Endogenous growth models (e.g. Saint-Paul 1992) have also demonstrated that excess leverage can impose cost on growth. On the demand side, either overlapping generations (Diamond 1965) or New Keynesian (Eggertsson and Krugman 2012) settings admit the possibility of slower growth due to debt constraints.
The bottom line is that, at least in theory, it is the totality of the debt burden that matters for growth outcomes. Whether this applies in practice, of course, is an empirical question.
Debt accumulation reduces output growth
In a recent paper (Lim 2019), I embed total debt into a panel vector autoregressive (PVAR) model to examine how debt accumulation affects output dynamics. Working with quarterly data, I exploit the temporal lag between shocks to output and the rational responses by borrowers to identify the effects of debt on growth. The identification assumption is premised on the fact that neither private action nor public policy responds contemporaneously to innovations in aggregate economic activity, but do so after a quarter, since reacting to such shocks typically requires redesigning policies, restructuring plans, renegotiating terms, and amending contracts, all of which require time.
Based on an unbalanced panel spanning as many as 64 years and up to 41 advanced and emerging economies, I find that increases in the share of total debt to GDP leads to a contraction in the GDP growth rate.4 This effect attains its maximum after a quarter, before gradually fading over the course of around a year (Figure 3). While there is some evidence that this effect is mitigated in open economies by movements in the current account and exchange rate—depreciation helps contribute to a surplus in the balance of payments, which eventually offsets the growth drag after about two years—the negative effect of total debt expansion on output growth is undeniable.
Figure 3 Impulse response functions for debt on growth
Panel A: Orthogonalised impulse response function for debt on growth
Source: Author's calculations.
Notes: IRF calculated for a one standard deviation innovation in debt. The dark (light) gray areas indicate the 68% (95%) confidence intervals generated using Gaussian approximation of 200 Monte Carlo draws from a fitted panel VAR.
Panel B: Cumulative impulse response function for debt on growth
Source: Author's calculations.
Notes: Cumulative IRF calculated for one standard deviation innovation in debt, for 20 quarters after the shock. The gray dark dashed lines indicate the 95 percent confidence intervals generated using Gaussian approximation of 200 Monte Carlo draws from a fitted panel VAR.
More specifically, a country that faces a one standard-deviation bump in debt accumulation—of 2.2 percentage points faster than the average quarterly growth rate of 0.4%—is liable to experience an accompanying decline in GDP growth amounting to 0.2%. With an annualised real growth rate of around 2.9% across the countries in our sample, this amounts to a growth deceleration of around 7%. Prolonged spurts of even debt build-up—like what would be expected as a result of the COVID-19 shock—could see an even more severe cost in terms of medium-run growth performance.
Many of the proposals that have been advanced for dealing with the health crisis, especially for developing countries, focus on managing the costs and consequences of public debt accumulation alone. But the importance of simultaneously addressing the private debt burden is becoming ever more compelling, and is being increasingly recognised by proposals for dealing with COVID-19-related debt (Bolton et al. 2020, Krahnke 2020, Marchesi and Masi 2020). This column corroborates the arguments made, using empirical evidence that it is the totality of the debt burden that exacts a growth cost. Helping economies recover rapidly from this dramatic shock requires tackling both public and private elements of borrowing.
Baldwin, R and B Weder di Mauro (eds.) (2020), Mitigating the COVID Economic Crisis: Act Fast and Do Whatever It Takes, London, England: CEPR Press.
Barro, R (1974), “Are Government Bonds Net Wealth?”, Journal of Political Economy 82(6): 1095–17.
Blanchard, O and S Fischer (1989), Lectures on Macroeconomics, Cambridge, MA: MIT Press
Bolton, P, L Buchheit, P-O Gourinchas, M Gulati, C-T Hsieh, U Panizza and B Weder di Mauro (2020), “Necessity Is the Mother of Invention: How to Implement a Comprehensive Debt Standstill for COVID-19 in Low- and Middle-Income Countries”, VoxEU.org, 21 April.
Diamond, P (1965), “National Debt in a Neoclassical Growth Model”, American Economic Review 55(5): 1126–50.
Eggertsson, G and P Krugman (2012), “Debt, Deleveraging, and the Liquidity Trap: A Fisher-Minsky-Koo Approach”, Quarterly Journal of Economics 127(3): 1469–1513.
Krahnke, T (2020), “Doing More With Less: How the IMF Should Respond to an Emerging Markets Crisis”, VoxEU.org, 14 April.
Lim, J J (2019), “Growth in the Shadow of Debt”, Journal of Banking and Finance 103: 98–112.
Marchesi, S and T Masi (2020), “Debt Restructuring in the Time of COVID-19: Private and Official Agreements”, VoxEU.org, 4 May.
Saint-Paul, G (1992), “Fiscal Policy in an Endogenous Growth Model”, Quarterly Journal of Economics 107(4): 1243–59.
1 Admittedly, a number of governments have simply reclassified planned budgetary outlays and have not actually committed to little or no fresh spending. Nevertheless, spending has been in the order of around 4% in many emerging economies, and almost twice as much in advanced ones.
2 In principle, public debt may not increase if spending draws on reserves, or if spending is rapidly monetised by a pliant central bank, thereby leading to no increases in the share of debt to output. Still, the general principle that COVID-19-related fiscal spending would translate into a significant increase in public debt is hardly contested.
3 I have deliberately chosen to retain only total credit to the nonfinancial sector, which are the ultimate borrowers. Further including financial-sector debt in the total debt stock computation, as some have done, would generally entail double-counting of the potential burden of debt on the economy.
4 Although this result is obtained from the estimation of a Cholesky-identified PVAR, I also consider the robustness of the estimate using a variety of dynamic heterogeneous error-correction models, as well as impulse responses generated from local projections. The qualitative implications of the analysis are unchanged.