- Shopping Bag ( 0 items )
Rong Qian, University of Maryland
Carmen M. Reinhart, University of Maryland and NBER
Kenneth Rogoff, Harvard University and NBER
This paper addresses the concept of "graduation" from external default, banking, and inflation crises. Employing a vast data set cataloging more than 2 centuries of financial crises for over 60 countries developed in Reinhart and Rogoff (2009), we explore the risk of recidivism across advanced economies versus middle- and low-income countries. We show that 2 decades without a relapse (falling into crisis) is an important marker. After 1800, roughly two-thirds of recurrences of external default on sovereign debt and three-quarters of recurrence of inflation crisis occur within 20 years. However, crisis recidivism distributions have very fat tails, so that it takes at least 50 and perhaps 100 years to meaningfully speak of "graduation." Indeed, in the case of banking crises in particular, it is hard to argue that any country in the world has truly graduated.
Given that graduation (with its companion question, will this ever happen again?) is arguably one of the most important issues in macroeconomics and development, there has been remarkably little theoretical or empirical investigation of the subject. For example, the large theory literature on sovereign lending and default, while producing many important insights into the fundamental distinction between willingness to pay and ability to pay, largely treats a country's basic developmental and political characteristics as parametric. There is very little on explaining the political, social, economic, and financial dynamics that ultimately lead a country to be less prone to certain types of crises.
We acknowledge that the concept of graduation is a hard nut to crack. Many advanced countries had enjoyed a long hiatus from systemic banking crises after World War II and yet had huge problems during the recent global financial crisis. After 90 years of serial default running from 1557 to 1647, Spain did not default again until 1809. Even the advanced countries had high inflation as recently as the 1970s and early 1980s, while many emerging markets had hyperinflation less than 2 decades ago. Is the advent of modern independent central banks sufficient to guarantee that fiscal dominance never again reasserts itself? Have the rich countries that have supposedly "graduated" from serial default on external debt shifted the locus of risk to de jure or de facto (via inflation or financial repression) default on domestic debt? Does the theory of sovereign default or of financial development tell us that we should expect richer and more advanced countries to be immune? Or is graduation a mirage, with the "graduates" really being at best "star pupils," and can graduates be distinguished from patients in remission?
Our goals in this paper are fairly narrowly circumscribed. Most of our analysis is based on data on the dates and duration of the crises themselves. We speculate on underlying causal factors but do not approach them empirically here. Although the various types of crises often occur in clusters, our quantitative analysis mainly treats individual crises separately.
We begin the paper in Section II by defining the crises that we will catalog. In Section III of the paper, we present a summary time line of crisis, followed by a brief overview of the early history of serial default on external debt. An interesting case is France, which defaulted on its external debt no fewer than nine times from the middle of the sixteenth century through the end of the Napoleonic War but has not defaulted on external debt since. France is a canonical case of what we define as an "external default graduate." (This did not stop France from having numerous severe banking crises in the past 2 centuries.)
In the main body of the paper, we provide a broad aggregative historical overview of the data across different types of crises, distinguishing between advanced countries and emerging markets, also taking into account the advent of International Monetary Fund (IMF) programs after World War II as another marker of a debt crisis.
In Section IX of the paper, we speculate on links between graduation and development and the possibility for recidivism among richer countries. The fact that the canonical theory of sovereign default does not strongly predict smaller problems in richer countries (it does not strongly predict graduation) might be considered a flaw in theory. But it might also be taken as warning sign that graduation can be more difficult and take even more time than our data of "just" a few centuries can reveal. On banking crises, the theory needs to better explain why countries never seem to graduate.
The main empirical results from our long-dated historical time series on financial crises may be described as follows. First, the process of "graduation," that is, emergence from frequent crisis suffering status, is a long process. False starts are common and recurrent. This is especially true in the case of banking crises, for both high-income countries and middle- and low-income countries.
Second, the vulnerability to crisis in high-income countries versus middle- and low-income countries differs mostly in external default crises, to a lesser extent in inflation crises, and surprisingly little in banking crises.
Third, the sequence of graduation for most countries is first to graduate from external default crisis, then from inflation crisis, and eventually from banking. The last stage of graduation is extremely difficult, even for high-income countries. Among high-income countries, even though most of them have graduated from external default crisis and inflation crisis, more than 20% recently experienced a banking crisis, and far more when weighted by size. Schularick and Taylor (2009) speculate that advanced countries continue to experience credit busts despite arguably advancing regulation and institutions, because as risks moderate, financial systems grow and restore them.
Finally, the role of IMF programs in crises in the modern period is important. The availability of IMF bridge loans certainly has increased countries' resilience to "sudden stops" but, even setting aside moral hazard problems, is by no means a cure-all. Countries entering IMF programs are still forced to undergo painful macroeconomic adjustments in an attempt to regain sound fiscal footing and regain access to private capital markets. The challenges of successfully implementing IMF programs are underscored by the fact that there are many significant cases in which countries default within 3 years of an IMF bailout. IMF programs may help facilitate orderly debt workouts but do not guarantee them. We also note that in its early history, many of today's rich countries regularly drew on IMF resources, although there has been a 3-decade hiatus.
II. Definition of Crises
External debt crisis. We distinguish between external and internal debt on the basis of the legal jurisdiction where the debt contracts are enforced. This is a convenient construct given the history and evolution of sovereign debt. Obviously it may be useful to parse the data in other ways for some exercises, and in principle our data set allows that.
Although there are exceptions and there has been some evolution in recent years, typically in our long-dated historical data set, external debt is denominated in foreign currency and held by foreign creditors. There are certainly important examples, such as Mexico's short-term Tesobono bonds in the mid-1990s, where the debt is domestic yet is denominated in foreign currency and held primarily by foreign creditors. Although we regard the U.S. abrogation of the gold clause in the early 1930s—when gold was revalued from $21 to $35 per ounce—to be a default on domestic debt, many non-U.S. residents were also holding the debt at the time. In general, following standard practice, we define an external debt crisis as any failure to meet contractual repayment obligations on foreign debts, including both rescheduling or repayments and outright default. (As both of these examples make clear, however, one ultimately needs to think carefully about whether graduation fromexternal default may sometimes just mean a shift to episodic de facto and de jure internal default.)
In practice, most defaults on external debt end up being partial, with creditors typically (but not always) repaying 30¢–70¢ or more on the dollar, admittedly not adjusting for risk. The rationale for lumping together defaults regardless of the ultimate "haircuts" creditors are forced to absorb is that, in practice, the fixed costs of external debt default (which include difficulties in obtaining trade credits and loss of reputation) tend to be large relative to the variable costs. In principle, one could parse episodes more finely here according to, say, output or tax revenue loss depending on data availability, although we do not undertake that exercise here. See, however, Tomz (2007) and Tomz and Wright (2007).
Inflation crises. Following Reinhart and Rogoff (2004), we define inflation crises as episodes in which annual inflation exceeds 20%. This threshold is lower than the 40% we and others have used in related studies on postwar data but is a compromise reflecting that prior to World War I, average inflation rates were much lower, and 20% inflation generally represented a significant level of dysfunction. Indeed, since we are particularly interested here in inflation as a vehicle for partial default, one clearly would also want to consider lower levels of sustained unanticipated inflation such as many advanced countries experienced in the 1970s. Depending on the maturity structure of debt, sustained 10% inflation can certainly be tantamount to de facto default. A proper calibration, however, would require detailed data on the maturity structure of debt (as in Missale and Blanchard 1994) and, ideally, also on the evolution of inflation expectations. We do not attempt this here, though again, this is an important caveat to interpreting the concept of graduation from external debt crises.
Banking crises. Our definition of banking crises follows standard practice (e.g., Kaminsky and Reinhart 1999; Caprio and Klingebiel 2003). Following our own earlier work, "We mark a banking crisis by two types of events: (1) bank runs that lead to the closure, merging or takeover by the public sector of one or more financial institutions and (2) if there are no runs, the closure, merging, takeover, or large-scale government assistance of an important financial institution (or group of institutions) that marks the start of a string of similar outcomes for other financial institutions" (Reinhart and Rogoff 2009, 11). We recognize that our listing of systemic (on a national scale) banking crises may be incomplete, especially prior to 1970, especially for crises outside the large money centers that attract the attention of the world financial press.
Having set out basic definitions, we are now ready to view some basic characteristics of the data. To provide context and motivation for the concept of graduation, we begin with a summary time line of financial crises since 1550, followed by a brief overview of the early history of sovereign defaults.
III. A Time Line of Financial Crises and the Early History of Sovereign Defaults
Table 1 provides a summary historical perspective that helps show how the three different varieties of financial crisis have spread over time and across country groups. Between 1550 and 1800, sovereign defaults on external debt were relatively common in Europe, but they were relatively rare elsewhere if only because (a) there were few other independent nations in a position to default and (b) given the crude state of global capital markets, relatively few countries were wealthy enough to attract international capital flows. Thus defaults were relatively insignificant in the regions that constitute today's emerging markets. Systemic banking crises, however, were relatively rare everywhere. The legal and technological underpinnings of modern private banking simply had not reached a stage of maturity and depth sufficient to cause systemic crises in most instances. (Of course, there are exceptions. Following Cipolla  and MacDonald , Reinhart and Rogoff  discuss how England's 1340 default to Florentine bankers triggered a financial crisis in Italy.) Similarly, inflation crises were relatively rare, although again there are many exceptions (see Reinhart and Rogoff 2009, chap. 12). Prior to the widespread adoption of paper currency, bouts of very high inflation were relatively difficult to engineer.
The end of the Napoleonic War in the early 1800s marks a significant transition. The largest advanced countries were increasingly able to avoid external default, albeit partly by their ability to issue an increasing share of their debt domestically. Default, however, became common in "peripheral" advanced countries such as Spain and Portugal, while newly independent emerging markets such as Greece and Latin America entered a long period of serial default. Over the same period, as advanced countries developed more sophisticated banking systems, banking crises became far more common. Emerging markets were certainly affected by advanced country banking crises but did not have so many of their own, if only because their financial systems were dominated by foreign banks.
By the turn of the twentieth century, emerging market financial institutions had developed to the point where domestic banking crises became more common. By the time of the Great Depression of the 1930s, banking crises were a worldwide phenomenon. Owing in no small part to the financial repression that followed in reaction to the Great Depression, banking crises were relatively rare during the period from the end of World War II until the early 1970s. As financial repression thawed, banking crises became more frequent in the advanced economies and serial in many emerging markets, bringing us to the recent financial crisis episode.
Finally, table 1 gives a time line of inflation crises, which of course were quite common in all countries in the 1970s and remained a problem in emerging markets until the past decade.
We thus focus our early history on sovereign external defaults. As Reinhart et al. (2003) and Reinhart and Rogoff (2009) emphasize, many of today's advanced economies had recurrent problems with default on sovereign debt during the period when they might arguably have been characterized as emerging markets. Table 2 illustrates the case of Europe for the 3-century period 1550–1850, with the years listed marking the beginning of a sovereign default episode.
As one can see clearly from the table, serial default was quite common among the major European powers during the sixteenth through nineteenth centuries, with France defaulting on its external debt nine times and Spain defaulting 10 times (with three more to follow in the second half of the nineteenth century). One important observation, immediately apparent from the table, is that there is typically a substantial interval between defaults, typically decades, but sometimes centuries. (Note that we require at least 2 years between default episodes to regard them as independent events.) After defaulting in 1683, Prussia's next default episode did not follow for more than a century in 1807. Portugal, after defaulting in 1560, did not default again until 1828, when the country lapsed into a period of serial default that did not end until 1890. At this writing, Portugal has not defaulted again since. (Importantly, during a significant portion of Portugal's quiescent period, it had effectively lost its independence.)
Figure 1 gives a measure of the duration of periods of recidivism during the pre-Napoleonic era for the independent (relatively) high-income countries of our sample. The figure captures the length of time between default episodes (including cases in which there was no recidivism). As one can see from the figure, fully half of all default recurrences occurred after a more than 20-year hiatus, with a significant percentage occurring even after a 60-year hiatus.
Advanced country external sovereign debt defaults have become much rarer events in the modern era. Germany's most recent default occurred in 1939, Austria's in 1940, and Hungary's in 1941 (Reinhart and Rogoff 2009). Especially interesting are the cases of Sweden and France. France, despite a near record level of defaults in its pre-Napoleonic era, has not defaulted on external debt since. Sweden, too, has not defaulted on external debt since its default at the end of the Napoleonic War in 1812. It would be interesting to explore whether wartime defaults are less damaging to reputation than peacetime defaults, though of course over many episodes, it is precisely the propensity to wage war that motivates many countries to build up large debts (as in the tax-smoothing model of Barro ). Later, we will consider the robustness of our recidivism results to the exclusion of wartime.
Reinhart and Rogoff (2009) also show that the kind of long cycles illustrated in table 2 are quite characteristic of some of today's emerging markets, many of which have defaulted at least once during the past 2-3 decades. The number of emerging markets that have experienced external debt crises expands considerably if one includes "near-default" episodes in which countries averted technical default thanks to IMF bridge loans. In virtually all these cases, the countries still suffered massive recessions as governments were forced to tighten fiscal policy as borrowing options dried up. Importantly, we do not include these in our calculations below, although arguably from the point of view of understanding macroeconomic volatility and the dangers of excessive debt accumulation, they are equally important. We return to this issue later when we study IMF programs.
Excerpted from NBER Macroeconomics Annual 2010 Copyright © 2011 by National Bureau of Economic Research. Excerpted by permission of The University of Chicago Press. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.