Growth in a Time of Debt

Whether there is a systematic relationship between debt levels, growth and inflation – asked because the crisis had just pushed public debt sharply higher and the authors wanted to know the long-run macroeconomic consequences. “In this paper, we exploit a new multi-country historical data set on central government debt as well as more recent data on external (public and private) debt to search for a systematic relationship between debt levels, growth and inflation” (Section I). The timeliness argument is quantitative: “Government debt has been soaring in the wake of the recent global financial maelstrom, especially in the epi-center countries. This might have been expected. Using a benchmark of 14 earlier severe post-World-War II financial crises, we demonstrated (one year ago) that central government debt rises, on average, by about 86 percent within three years after the crisis.” The policy question is posed directly: “Outsized deficits and epic bank bailouts may be useful in fighting a downturn, but what is the long run macroeconomic impact or higher levels of government debt, especially against the backdrop of graying populations and rising social insurance costs?”

Gross central government debt for 44 countries over about two centuries, more than 3,700 annual observations, plus more recent data on total gross external debt. The authors stress that the dataset is the paper’s enabling contribution: “Prior to this data set, it was exceedingly difficult to get more than two or three decades of public debt data even for many rich countries, and virtually impossible for most emerging markets.” The observations cover “a wide range of political systems, institutions, exchange rate and monetary arrangements, and historic circumstances” (Section I). Definitions are stated narrowly: “‘public debt’ refers to gross central government debt. ‘Domestic public debt’ is government debt issued under domestic legal jurisdiction. Public debt does not include debts carrying a government guarantee. Total gross external debt includes the external debts of all branches of government as well as private debt that is issued by domestic private entities under a foreign jurisdiction.” The choice of the narrow public measure is explained by availability, and the authors flag what it leaves out: “the true run-up in debt is significantly larger than stated here, at least on a present value actuarial basis, due to the extensive government guarantees that have been conferred on the financial sector in the crisis countries and elsewhere.”

About 75 percent in real terms for the five systemic-crisis countries and about 20 percent for others, between 2007 and 2009. “For the five countries with systemic financial crises (Iceland, Ireland, Spain, the United Kingdom, and the United States), average debt levels are up by about 75 percent, well on track to reach or surpass the three year 86 percent benchmark that Reinhart and Rogoff (2009a,b) find for earlier deep post-war financial crises. Even in countries that have not experienced a major financial crisis, debt rose an average of about 20 percent in real terms between 2007 and 2009.” The authors attribute the rise to three causes and note the contrast with the immediately preceding years: it “stands in stark contrast to the 2003-2006 period of public deleveraging in many countries and owes to direct bail-out costs in some countries, the adoption of stimulus packages to deal with the global recession in many countries, and marked declines in government revenues that have hit advanced and emerging market economies alike” (Section II).

Barro’s distortionary-taxation channel for growth, and the erosion of the real value of nominal debt for inflation, with an explicit qualification about maturity structure. “The simplest connection between public debt and growth is suggested by Robert Barro (1979). Assuming taxes ultimately need to be raised to achieve debt sustainability, the distortionary impact imply is likely to lower potential output. Of course, governments can also tighten by reducing spending, which can also be contractionary. As for inflation, an obvious connection stems from the fact that unanticipated high inflation can reduce the real cost of servicing the debt. Of course, the efficacy of the inflation channel is quite sensitive to the maturity structure of the debt. Whereas long-term nominal government debt is extremely vulnerable to inflation, short term debt is far less so. Any government that attempts to inflate away the real value of short term debt will soon find itself paying much higher interest rates” (Section III). These are motivations rather than structures the paper tests; nothing in the empirical work identifies a channel.

No, and it says so explicitly, along with which direction the omission is likely to bias the results. “In principle, the manner in which debt builds up can be important. For example, war debts are arguably less problematic for future growth and inflation than large debts that are accumulated in peace time. Postwar growth tends to be high as war-time allocation of manpower and resources funnels to the civilian economy. Moreover, high war-time government spending, typically the cause of the debt buildup, comes to a natural close as peace returns. In contrast, a peacetime debt explosion often reflects unstable underlying political economy dynamics that can persist for very long periods. Here we will not attempt to discriminate the genesis of debt buildups, and instead simply look at their connection to average and median growth and inflation outcomes. This may lead us, if anything, to understate the adverse growth implications of debt burdens arising out of the current crisis, which was clearly a peace time event” (Section III). This matters for reading the results, because it means postwar high-debt, high-growth years enter the top bucket on the same footing as peacetime debt explosions.

For 20 advanced economies over 1946-2009, no obvious debt-growth link below 90 percent, then median growth about 1 percentage point lower and mean growth almost 4 points lower – with 96 observations in the top bucket. The country set is Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, the United Kingdom and the United States (Section III.A). The bucket counts are given as “443 for debt/GDP below 30%; 442 for debt/GDP 30 to 60%; 199 observations for debt/GDP 60 to 90%; and 96 for debt/GDP above 90%,” and the authors note “there are a significant number in each category, including 96 above 90 percent. (Recent observations in that top bracket come from Belgium, Greece, Italy, and Japan.)” The result: “it is evident that there is no obvious link between debt and growth until public debt reaches a threshold of 90 percent. The observations with debt to GDP over 90 percent have median growth roughly 1 percent lower than the lower debt burden groups and mean levels of growth almost 4 percent lower. (Using lagged debt should not dramatically change the picture.)” One internal discrepancy is worth a reader’s attention: the text refers to “the 1186 annual observations” while the figure note states “There are 1,180 observations,” which is what the four bucket counts sum to.

Yes, on the authors’ reading, with mean growth of 1.7 percent above 90 percent against 3.7 percent below 30 percent, across 2,317 country-year observations. “Table 1 provides detail on the growth experience for individual countries, but over a much longer period, typically one to two centuries. Interestingly, introducing the longer time series yields remarkably similar conclusions. Over the past two centuries, debt in excess of 90 percent has typically been associated with mean growth of 1.7 percent versus 3.7 percent when debt is low (under 30 percent of GDP), and compared with growth rates of over 3 percent for the two middle categories (debt between 30 and 90 percent of GDP)” (Section III.A). Bucket sample sizes differ sharply across brackets – 866, 654, 445 and 352 – and the table notes that “there are missing observations, most notably during World War I and II years.” The authors report country-level heterogeneity honestly: “there is considerable variation across the countries, with some countries such as Australia and New Zealand experiencing no growth deterioration at very high debt levels. It is noteworthy, however, that those high-growth high-debt observations are clustered in the years following World War II.”

Very similar on growth, and sharply different on inflation. “Perhaps surprisingly, the results illustrated in Figure 4 and Table 2 for emerging markets largely repeat the results in Figure 2 and Table 1. For 1900-2009, for example, median and average GDP growth hovers around 4-4.5 percent for levels of debt below 90 percent of GDP but median growth falls markedly to 2.9 percent for high debt (above 90 percent); the decline is even greater for the average growth rate, which falls to 1 percent. With much faster population growth than the advanced economies, the implications for per capita GDP growth are in line (or worse) with those shown for advanced economies” (Section III.B). The 24 countries are Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, Ecuador, El Salvador, Ghana, India, Indonesia, Kenya, Korea, Malaysia, Mexico, Nigeria, Peru, Philippines, Singapore, South Africa, Sri Lanka, Thailand, Turkey, Uruguay and Venezuela, with 1,142 postwar observations (502, 385, 145 and 110 across the four buckets) and 1,397 over 1900-2009. On inflation: “The similarities with advanced economies end there, as higher debt levels are associated with significantly higher levels of inflation in emerging markets. Median inflation more than doubles (from less than 7 percent to 16 percent) as debt rises from the low (0 to 30 percent) range to above 90 percent. Fiscal dominance is a plausible interpretation of this pattern.” Note that emerging-market data availability is itself a constraint: “while we have pre-1900 inflation, real GDP, and public debt data for many emerging markets, nominal GDP data is seldom available.”

Because emerging-market external debt is almost entirely foreign-currency denominated and because the public-private distinction dissolves in a crisis; the growth threshold is around 60 percent, well below the public debt threshold. “Combined public and private sector debt is of interest because in the case of crisis, the distinction between public and private often becomes blurred in a maze of bailouts, guarantees, and international hard currency constraints” (Section III.C). The result: “the growth thresholds for external debt are considerably lower than for the thresholds for total public debt. Growth deteriorates markedly at external debt levels over 60 percent, and further still when external debt levels exceed 90 percent, which record outright declines. In light of this, it is more understandable that over one half of all defaults on external debt in emerging markets since 1970 occurred at levels of debt that would have met the Maastricht criteria of 60 percent or less. Inflation becomes significantly higher only for the group of observations with external debt over 90 percent.” The sample is 20 emerging markets over 1970-2009 with 755 annual observations (252, 309, 120 and 74 across the four buckets) – a smaller and more recent sample than the public debt exercise, worth keeping in mind when comparing the two thresholds.

No – the data begin only in 2003, and the authors say explicitly that they cannot calculate one and that it is likely higher anyway. “We are not in a position to calculate separate total external debt thresholds (as opposed to public debt thresholds) for advanced countries. The available time series is too recent, beginning only in early 2000s as a byproduct of the International Monetary Fund efforts and creation of the Special Data Dissemination Standard” (Section I). The point is repeated at the end of Section III.C: “given the lack of sufficient long-dated historical data on advanced economies external debts, it is not possible to know whether they face similar thresholds to emerging markets. It is likely that the thresholds are higher for advanced economies that issue most external debt in their own currency.”

Very high, especially in Europe, on a 59-country snapshot covering 2003 to mid-2009. “External debt burdens are particularly high in Europe, with an average external debt to GDP ratio across advanced European economies of over 200 percent, and an average external debt to GDP across emerging European economies roughly 100 percent,” and a footnote adds that “if Ireland is added to the list, the average for advanced European economies rises to 266 percent.” The composition matters, and the authors say so: “(The fact that a sizable share of these debts are intra-European may or may prove a significant mitigating factor.)” Comparative points: “the United States’ gross debt liabilities are less than half of Europe’s as a share of GDP, despite the country’s epic sequence of trade balance deficits. Japan, despite having a gross public debt to GDP ratio approaching 200 percent, has much smaller gross external liabilities still, thanks in no small part to Japan’s famously strong home bias in bond holdings.” And the reversal relative to reputation: “Famously profligate Latin America, by contrast to the advanced economies, now has gross external debt liabilities averaging only around 50 percent of GDP. Moreover, in contrast to the advanced countries who added an average of 50 percent of GDP to gross external debt during the recent period, Latin American countries actually [deleveraged] external debt by over 30 percent of GDP” (Section III.C).

That post-crisis private deleveraging is a separate drag on growth, illustrated with the US record – and included because the public and external debt results give “only a partial picture” of the years right after a crisis. “Private debt, in contrast to public debt, tends to shrink sharply for an extended period after a financial crisis. Just as a rapid expansion in private credit fuels the boom phase of the cycle, so does serious deleveraging exacerbate the post-crisis downturn. This pattern is illustrated in Figure 7, which shows the ratio of private debt to GDP for the United States for 1916-2009. Periods of sharp deleveraging have followed periods of lower growth and coincide with higher unemployment. … In varying degrees, the private sector (households and firms) in many other countries (notably both advanced and emerging Europe) are also unwinding the debt built up during the boom years. Thus, private deleveraging may be another legacy of the financial crisis that may dampen growth in the medium term” (Section IV). The authors are candid about the data limitation: they focused on public and external debt “since reliable data on private internal domestic debts are much scarcer across countries and time,” so this section is explicitly an illustration from one country rather than a cross-country result.

It does not use causal language; the findings are stated as associations, and no identification strategy is offered. The abstract and the conclusion both speak of relationships and associations: “the relationship between government debt and real GDP growth is weak for debt/GDP ratios below a threshold of 90 percent of GDP”; “high debt/GDP levels (90 percent and above) are associated with notably lower growth outcomes” (Section V). The empirical method is a comparison of means and medians across debt buckets, with no regression, no controls, and no attempt at identification – the authors describe the exercise as simply looking “at their connection to average and median growth and inflation outcomes.” The single nod toward direction of causation is a parenthesis: “(Using lagged debt should not dramatically change the picture.)” A reader should treat the paper’s own framing as load-bearing here: it establishes patterns in a long historical dataset and speculates about why they might exist, and the policy conclusions others drew from it went further than the text does.

They say they cannot, and offer their earlier “debt intolerance” argument as speculation. “Why are there thresholds in debt, and why 90 percent? This is an important question that merits further research, but we would speculate that the phenomenon is closely linked to logic underlying our earlier analysis of ‘debt intolerance’ in Reinhart, Rogoff, and Savastano (2003). As we argued in that paper, debt thresholds are importantly country-specific and as such the four broad debt groupings presented here merit further sensitivity analysis. A general result of our ‘debt intolerance’ analysis, however, highlights that as debt levels rise towards historical limits, risk premia begin to rise sharply, facing highly indebted governments with difficult tradeoffs. Even countries that are committed to fully repaying their debts are forced to dramatically tighten fiscal policy in order to appear credible to investors and thereby reduce risk premia” (Section V). Two further vulnerabilities are named but not tested: reliance on short-term borrowing, where “countries that choose to rely excessively on short term borrowing to fund growing debt levels are particularly vulnerable to crises in confidence that can provoke very sudden and ‘unexpected’ financial crises,” and the foreign-versus-domestic currency composition of debt, which together lead the authors to conclude that “traditional debt management issues should be at the forefront of public policy concerns.”

The bucket boundaries, the absence of country-specific thresholds, the coverage gaps, and the fact that crisis-driven debt increases need not reach the top bucket at all. On the brackets: “The four ‘buckets’ encompassing low, medium-low, medium-high, and high debt levels are based on our interpretation of much of the literature and policy discussion on what is considered low, high etc debt levels. It parallels the World Bank country groupings according to four income groups. Sensitivity analysis involving a different set of debt cutoffs merits exploration as do country-specific debt thresholds along the broad lines discussed in Reinhart, Rogoff, and Savastano (2003).” On coverage: the long-run tables note “there are missing observations, most notably during World War I and II years,” and country sample periods vary widely – from 1790 for the United States to 1949 for Ireland in the advanced table, and from 1900 to 1990 in the emerging table. On the policy inference: “It is important to note that post crises increases in public debt do not necessarily push economies in to the vulnerable 90+ debt/GDP range.” Two features a careful reader should verify directly against the tables rather than infer: the top-bucket statistics are unweighted averages across countries with very unequal numbers of high-debt years, and the text’s observation count for the advanced postwar sample (1,186) does not match the sum of the four reported bucket counts (1,180). Editorial cross-reference, not a finding of this paper: a subsequent replication by Herndon, Ash and Pollin in the Cambridge Journal of Economics disputed this paper’s data handling and weighting; see that record for its findings.

That high debt is associated with notably lower growth across both country groups, that emerging markets face a tighter external-debt threshold, and that growing out of a deep debt burden is rare. “The sharp run-up in public sector debt will likely prove one of the most enduring legacies of the 2007-2009 financial crises in the United States and elsewhere. We examine the experience of forty four countries spanning up to two centuries of data on central government debt, inflation and growth. Our main finding is that across both advanced countries and emerging markets, high debt/GDP levels (90 percent and above) are associated with notably lower growth outcomes. In addition, for emerging markets, there appears to be a more stringent threshold for total external debt/GDP (60 percent), that is also associated with adverse outcomes for growth. Seldom do countries simply ‘grow’ their way out of deep debt burdens” (Section V). The concluding paragraph returns to the advanced-economy external debt position – “in the case [of] Europe, the advanced country average exceeds 200 percent external debt to GDP” – with the appropriate hedge: “Although we do not have the long-dated time series needed to calculate advanced country external debt thresholds as we do for emerging markets, current high external debt burdens would also seem to be an important vulnerability to monitor.”