The widespread banking crises in advanced economies and more recently the “near” default of Greece have dashed the popular notion that rich countries have outgrown severe financial crises. Record or near-record declines in output accompanying these events signalled the end of the “great moderation era” (Bernake 2004), which turned out to be rather short-lived for an era. It began in the mid-1980s, but history shows that 20 years is a modest time frame when it comes to bank and debt crises (Tomz 2007). In fact, graduation from recurring sovereign external debt crises is a very tortuous process that sometimes takes a century or more. For banking crises, we simply do not know what it takes to graduate. Indeed, it is unclear whether any country has managed it.
In our recent paper (Qian et al. 2010), we address the elusive concept of “graduation” from external default, banking, and inflation crises. Employing a dataset cataloguing more than two centuries of financial crises for over sixty countries (developed in Reinhart and Rogoff 2009), we explore the risk of crisis reversals across advanced economies and middle- and low-income countries. We show that two decades without a relapse is an important marker – albeit far from a guarantee. Post 1800, roughly two-thirds of recurrences of external default, and three-quarters of recurrence of inflation crisis, have occurred within a twenty-year window. However, crisis relapse distributions have very fat tails, so that it takes at least fifty years and perhaps longer to meaningfully speak of “graduation”. (As in Reinhart and Rogoff (2009), inflation crises are defined as years where the annual inflation rate exceeds 20 percent.)
The main findings of our study can be summarised as follows:
- The process of “graduation,” defined as the emergence from recurrent crisis bouts, is a long drawn-out process. False starts are common and recurrent. These false starts are routinely (mis)interpreted by financial markets and policymakers as evidence the countries have really turned the corner quickly because this time is different. Indeed, false starts are more the norm than the exception in the case of banking crises, for both high- and middle-low-income countries.
- The vulnerability to crisis in high-income countries versus middle- and low-income countries differs most in external default crises, somewhat less for inflation crises, and is surprisingly similar for banking crises. Currency crashes are also a recurring phenomenon in advanced and emerging economies alike.
- The sequence for most of countries is first to graduate from external default crises, then from inflation crises.
Varieties of financial crises across the ages
In Table 1, we present a timeline of financial crises that underscores how the incidence of a particular strand of crisis (default, inflation, and banking) rises and declines across the centuries.
As the table indicates, between 1550 and 1800, sovereign defaults on external debt were relatively common in Europe. Defaults, however, were relatively rare elsewhere if only because
- there were relatively few independent nations outside Europe (sovereignty being an essential ingredient of sovereign default) and
- the crude state of global capital markets meant that relatively few countries were wealthy enough to attract significant international capital flows and have the “luxury” of external debts.
Before 1800, systemic banking and inflation crises were relatively rare everywhere because, in the case of the former, the legal and technological underpinnings of modern private banking had not reached a stage of maturity and, as to the latter, the printing press was not yet in wide use. Debasing coinage required more effort.
The end of Napoleonic War in the early 1800s marked a significant transition. The largest advanced countries were increasingly able to avoid external default partly by their ability to issue an increasing share of their debt domestically. Default, however, became common in “peripheral” or newly independent countries.
Over the same period, advanced economies developed more sophisticated banking systems. As the number of banks grew, so did the incidence of banking crises. The “typical” emerging market did not have a well developed domestic financial sector and relied on external bank and bankers; the co-existence during this period of a low incidence of domestic banking crises and a high incidence of default on external debts should not be surprising.
By the turn of the twentieth century, emerging market financial institutions had developed to the point where domestic banking crises became more common. Due to the financial repression that followed the Great Depression, banking crises were relatively rare. As financial repression thawed into the early 1970s, banking crises became more frequent in the advanced economies and serial in many emerging markets, bringing us to the recent financial crisis episode.
Table 1. Varieties of financial crises across the ages, 1550-2010
Source: Qian et al. (2010)
Crisis probabilities: Advanced versus emerging middle- and low-income countries
Table 2 documents the difference between the (advanced) high income countries and the emerging economies for the three types of crises. The largest gap between the two country groups is the likelihood of external default. The average external default crisis probability of high-income group is less than half that of middle and low income countries and almost one fifth that of Latin America. The difference across groups in the incidence of high inflation shows a similar pattern. The average probability of banking crises in the advanced and emerging countries, by contrast, are extremely similar.
Banking crises are “an equal opportunity menace” (Reinhart and Rogoff 2009, chapter 10). Banking crises picked up dramatically in emerging markets since 1980, and, as is well-known, these have flourished in rich countries since 2007 as well. Yet, at present, neither high- nor middle/low-income countries have imminent prospects of meeting the criteria for “graduation” from banking crises.
The risk of a reversal: Tranquil periods between crises
We next examine the frequency distribution of “tranquil” periods. How long, on average, is it before one crisis episode ends and the next crisis begins? Figure 1 gives the statistics separately for external default, inflation, and banking crises.
The frequencies are broadly similar across different types of crises, with a significant share of distribution falling between 10 and 20 years. However, the fat tails of the distributions highlight that 20 years without a default, banking or inflation crisis is hardly decisive evidence of “graduation”. Notice that:
- For external defaults, most emerging market recurrences happen within 20 years;
- Only a few countries that have once defaulted have avoided any further defaults, at least not long enough to pass the 50-year filter we use.
- For inflation and banking crises, the 20-year mark contains an even larger percentage of reversals and, cases where there are of no reversal (or recurrence) are scarce.
Figure 1. Duration of “tranquil time”
Source: Qian et al. (2010)
To gain a deeper insight into relapses – or its complement, graduation – we look at measures of distance since the last crisis. As shown in Figure 2:
- The median tranquil period for default is just over a century for the advance countries (105 years) versus only 14 years for the developing countries. (This 105-year tranquil spell for advanced economies is calculated including 211-year entries (1800-2010) for countries which never defaulted nor had an inflation crisis during this sample.)
- The world median is 23 years.
- For banking crises, the median for the advanced countries is 9 versus 12 years for the developing countries.
Figure 2. Years since last external default through 2010 (or since 1800 or year of independence if there are no defaults)
Source: Qian et al. (2010)
Because financial crises tend to occur only at very long periodicities, it is easy to forget that they happen at all. Across a large data set, we find that two decades without a crisis is an important market to signal the “first step” toward graduation, but also that there are still relatively high odds of relapsing into crisis even after several decades of tranquillity.
Premature declarations of victory are all too common
Argentina’s poster child status within the international community as late as 1998-1999 (on the eve of disaster) is an illustration of this tendency. Rich countries do seem significantly less likely to default on external debt, though of course part of their "graduation” is a by-product of being able to fund themselves internally in domestically issued debt, or, in the case of Greece recently, in being a part of a wealthy club that can help support a bailout.
When it comes to banking crises we state the obvious. The post-2007 resurgence of banking crises makes plain that, to date, it is not possible to meaningfully speak of “graduation” as having taken place in advanced countries. Indeed, our 211-year sample shows that the absence of banking crisis in the high-income economies from the end of World War II through to 2006 was the exception rather than the rule.
Bernanke, Ben (2004), “The Great Moderation”, remarks at the 2004 Eastern Economic Association meetings, Washington DC.
Qian, Rong, Carmen M Reinhart, and Kenneth S Rogoff (2010), “On Graduation from Default, Inflation and Banking Crises: Elusive or Illusion?”, NBER Macroeconomics Annual 2010, MIT Press.
Reinhart, Carmen M and Kenneth S Rogoff (2009), This Time is Different: Eight Centuries of Financial Folly, Princeton University Press.
Tomz, Michael (2007), Reputation and International Cooperation: Sovereign Debt Across Three Centuries, Princeton University Press.