Not all government budget deficits are created equal: Evidence from advanced economies' sovereign bond markets

https://doi.org/10.1016/j.jimonfin.2021.102460Get rights and content

Highlights

  • Empirically, higher government budget deficits are associated with higher sovereign bond yields.

  • However, we find that for sovereign bond markets, not all budget deficits are created equal.

  • Investors penalise deficits due to higher current expenditure significantly more than deficits due to higher investment.

  • Some deficits are thus considered more “excusable” than others from a sovereign risk perspective.

Abstract

We find that bond markets charge significantly higher interest rates for deficits due to higher government current spending than for deficits due to higher government investment. Thus, from a sovereign risk perspective, not all government budget deficits are created equal. To show this, we use a panel regression approach on European Commission data for 31 advanced economies from 1990 onwards. Econometrically, we address potential endogeneity by using forecasts of fiscal variables and by instrumental variable methods. Based on our preferred specifications, a higher deficit solely due to higher government investment would in fact decrease long-term government bond yields. These findings suggest that the policy debate about fiscal sustainability and fiscal rules should, at the very least, distinguish budget deficits that are the result of investment from those that are not.

Introduction

On Wolfgang Schäuble’s last day in office as Germany finance minister in October 2017, his staff bid him farewell by dressing in black, assembling in the courtyard of the ministry and forming a big circle. They were alluding to Schäuble’s commitment to a “black zero”, an always-balanced budget, a policy that his successor maintained until the COVID-19 pandemic hit. Elsewhere too, finance ministries have sat on their hands, even in the face of negative borrowing cost for Germany and – for much of Europe – a significant public investment need, an cyclical slowing-down and monetary policy with little room to act. Hence, the vast fiscal support initiated in the course of 2020 by many EU and OECD countries in response to the COVID-19 pandemic is noteworthy not just for its sheer magnitude but also because it throws into stark relief just how restricted the role of fiscal policy had previously been.

As far as we are aware, there is no plausible economic theory according to which a budget that is always balanced, regardless of economic circumstance and government expenditure type, would constitute good economic policy. That is why this paper seeks to explore the matter empirically: we ask what bond markets “think” of fiscal policy or budgetary stances. We do so by analysing whether government bond yields in 31 EU and OECD countries since 1990 respond differently to deficits due to different types of government expenditure, most importantly government current expenditure and government investment. This means that we are hearing the verdict on fiscal policy stances “from the horse’s mouth”, given the feedback loop between fiscal policy decisions and government bond yields that became apparent again during the sovereign debt crisis. In a nutshell, our findings suggest that markets indeed distinguish deficits that are the result of higher government current spending from those that stem from higher government investment, and penalise the former significantly more than the latter. In this sense, we find that some deficits appear to be considered more “excusable” than others.

The existing empirical literature on the link between fiscal policy and interest rates has tended to disregard the reasons for which a government may increase its debt or run a budget deficit. A deficit may be the result of higher government current spending, higher government investment, or lower government revenue. Our analysis explicitly incorporates the current spending, investment and revenue components of the government budget deficit in an empirical analysis of the fiscal policy determinants of government bond yields: it thus takes seriously the intuitively plausible possibility that, from the perspective of bond markets, not all budget deficits are created equal.

Two key observations motivate the concern of this paper with government investment or asset accumulation and, hence, one of the underlying drivers of government budget deficits. First, in many OECD countries, government net worth ratios – that is, total government assets less total government liabilities, relative to GDP or national income – have declined substantially in recent decades (Atkinson, 2015, Buiter, 1985, IMF, 2018, Piketty and Zucman, 2014). This means that a considerable share of new debt in these countries was not used to finance investment, or that the process of asset accumulation through investment stalled. For example, government net worth was in negative territory in recent years in the US, UK, Germany and Japan, having stood at more than 60% of national income before 1990 in all four countries (Atkinson, 1995, Piketty and Zucman, 2014, IMF, 2018). In fact, in the sample of countries the IMF considered in its IMF, 2018 Fiscal Monitor, one third of countries had negative government net worth. While such developments would be expected to spell trouble for individuals, corporates or banks, they seem to have gone largely unnoticed – at least in academic and policy debates – in the case of many sovereigns.

Second, although trend levels of government investment vary substantially across EU and OECD countries – they are typically around or just below 4% of GDP in France and the United States, but only about half that in Germany – they have declined noticeably in recent years in many places, both in countries facing immediate domestic sovereign debt crises, such as Portugal, but also in those that did not, such as Japan or the United States. Proxying government investment with gross fixed capital formation of general government, Fig. 1 illustrates this fact. The IMF (2014) also notes a trend reduction of 25 percent in government investment in OECD countries, while the European Commission’s recent Economic Governance Review points to a widespread and persistent decline in public investment in Europe over the last decade (European Commission, 2020).

The declines in government net worth and government investment are linked: any part of a government budget deficit not used to finance government investment will have contributed to the decline in government net worth. It also seems likely that the focus of fiscal policy on austerity following the financial crisis played a part – in aiming to reduce debt and deficits, policy makers seldom appeared to explicitly consider implications for government asset accumulation and net worth. Consider the case of Germany, where a balanced-budget requirement was introduced in 2009. Germany’s budget deficit fell from 3.2% of GDP in 2009 to 0.1% of GDP in 2013 – that is, it decreased by 3.1 percentage points. Over the same period, government net worth as a share of national income fell by 3.6 percentage points, due to a corresponding fall of 3.6 percentage points in government non-financial assets, which comprise roads, bridges, buildings and so on. Thus, one way of looking at the matter is that Germany paid for balancing its budget by shrinking government non-financial assets and net worth – a process that the IMF refers to as a “fiscal illusion” (IMF, 2018) – all while facing the lowest borrowing cost on historical record.

The present paper hence explores the relevance of the composition of and reasons for government budget deficits for government bond yields, and thereby for government borrowing costs and fiscal sustainability. To do so, we use data on the investment, current spending and revenue components of the deficit drawn from the European Commission for a panel of 31 EU and OECD countries from 1990 onwards. To the best of our knowledge, the existing empirical literature on the links between fiscal policy and bond yields has not yet taken into account the underlying drivers of budget deficits through this kind of decomposition. In line with the literature, we confirm that higher government budget deficits (relative to GDP) increase government bond yields, with statistically significant yet quantitatively small effects. We then decompose the budget deficit to show that markets charge significantly higher interest rates for government budget deficits due to government current spending than for deficits due to government investment. Based on our preferred specifications, an increase in the deficit solely due to an increase in government investment would decrease bond yields, while an increase in the deficit solely due to an increase in government current spending would increase bond yields, ceteris paribus. This suggests that markets perceive government investment, on average, as favourable for growth or as having a positive return. Econometrically, we address the potential for endogeneity in a number of ways, including through timing shifts, instrumental variable approaches and the use of forecasts instead of actual values of fiscal variables.

Our findings reflect the perspective of government bond markets on sovereign risk, fiscal sustainability and government budget management. Given the need for fiscal policy to remain supportive for a few more years at least to stem the economic impact of the COVID-19 pandemic, these findings could have important policy implications. They imply that fiscal consolidation policies ought to focus more on government current (or non-investment) spending than government investment (or capital) spending, or at least distinguish the two in formulating deficit or debt reduction targets. They also suggest that fiscal rules in individual countries and monetary unions – such as the European Union’s Maastricht criteria, which have aimed to limit budget deficits to 3% of GDP irrespective of economic circumstance and underlying government expenditure composition – should reflect to what extent government budget deficits are the result of government investment. More generally, these findings suggest that the focus of fiscal policy in much of the EU and OECD until early 2020 on indiscriminate deficit and debt reduction regardless of economic circumstance and government expenditure type may not have been optimal, nor always required to contain government bond yields and borrowing costs.

The remainder of this paper proceeds as follows. The related empirical literature on the links between government budget deficits and interest rates are discussed in Section 2. Section 3 introduces the data and econometric approach. Our main results, robustness tests and extensions are presented in Section 4. Section 5 concludes and discusses policy implications.

Section snippets

Government budget deficits and interest rates

Fiscal policy directly and indirectly impacts a vast number of macroeconomic outcomes. One way of empirically evaluating fiscal policy is to identify its effects on a macroeconomic outcome of interest. The approach of Reinhart and Rogoff, 2010 is to bracket countries by the magnitude of their debt-to-GDP ratio, and compare different brackets’ mean and median growth and inflation rates. An alternative approach is to use government bond yields, which reflect investors’ perspectives on the

The data

The sample used for econometric analysis is based on annual data on government budget deficit components for the largest possible number of EU and OECD countries for which data are available, drawn from the European Commission DG ECFIN’s Annual Macroeconomic database (AMECO). This results in a sample covering 31 countries and, on average, 19 years per country.3

Main findings

This section presents key results from estimating a model of the type presented in Eq. 2. In essence, we investigate here whether bond markets accord importance to the composition of government budget deficits, in the sense of distinguishing deficits due to current spending and deficits due to investment.

Table 1 contains Fixed Effects (“FE”) estimates of a set of main specifications of interest. We report Driscoll-Kraay (Driscoll and Kraay, 1998) standard errors. Column (1) estimates a

Conclusion

Do sovereign bond markets account for the reasons underlying government budget deficits? We find that they do, and that in this sense, not all government budget deficits are created equal: markets charge significantly higher interest rates for deficits that are the result of higher current expenditure than for deficits that are the result of higher government investment. In this sense, we find that some deficits appear to be considered more “excusable” than others. Based on our preferred

Acknowledgements

I am greatly indebted and extremely grateful to Steve Bond, John Muellbauer and the late Sir Tony Atkinson for their guidance, inspiration and support. I would like to thank Daragh Clancy and seminar participants at Oxford, Barcelona, the ECB and ESM for helpful comments. Any errors remain my own.

This work was supported by the Economic and Social Research Council, UK.

References (45)

  • A.B. Atkinson

    The welfare state and economic performance

    National Tax Journal

    (1995)
  • A.B. Atkinson

    Inequality: what can be done?

    (2015)
  • E. Baldacci et al.

    Fiscal deficits, public debt, and sovereign bond yields

    IMF Working Paper.

    (2010)
  • M. Baxter et al.

    Fiscal policy in general equilibrium

    American Economic Review

    (1993)
  • S. Bazzi et al.

    Blunt instruments: Avoiding common pitfalls in identifying the causes of economic growth

    American Economic Journal: Macroeconomics

    (2013)
  • Bernoth, K., v. Hagen, J., Schuknecht, L., 2004. Sovereign risk premia in the european government bond market. ECB...
  • Blanchard, O., 2019. Public Debt: Fiscal and Welfare Costs in a Time of Low Interest Rates. Technical Report. Policy...
  • W.H. Buiter

    A guide to public sector debt and deficits

    Economic Policy

    (1985)
  • M. Canzoneri et al.

    Should the European Central Bank and the Federal Reserve be concerned about fiscal policy?

  • Dai, Q., Philippon, T., 2006. Fiscal policy and the term structure of interest rates. Working paper, New York...
  • J.C. Driscoll et al.

    Consistent covariance matrix estimation with spatially dependent panel data

    Review of Economics and Statistics

    (1998)
  • Eberhardt, M., 2013. Xtmg: Stata module to estimate panel time series models with heterogeneous slopes. Statistical...
  • Cited by (5)

    • Debt-financed fiscal stimulus in South Africa

      2024, Journal for Studies in Economics and Econometrics
    • Land-leasing behavior, local officials’ promotions, and chinese cities’ debt risks

      2021, International Journal of Strategic Property Management
    1

    The views expressed in this paper are those of the author and do not necessarily represent those of the ESM or ESM policy.

    View full text