This Time Is Different: Eight Centuries of Financial Folly Read online

  Finally, but worthy of discussion, our approach is systematic in documenting the incidence of default but is silent about the magnitude of default. Even though our database on public debt can provide valuable insight on the magnitude of the original default or restructuring, it would be a stretch of the imagination to suggest that these data provide a snapshot of the subsequent restructuring nuances or the actual recovery rates. With these caveats in mind, a number of results stand out.

  Real GDP in the Run-up to and the Aftermath of Debt Defaults

  First, how bad are macroeconomic conditions on the eve of a default? Unambiguously, output declines in the run-up to a default on domestic debt are typically significantly worse than those seen prior to a default on external debt. As highlighted in figures 9.1 and 9.2, the average cumulative decline in output during the three-year run-up to a domestic default crisis is 8 percent. The output decline in the year of the domestic debt crisis alone is 4 percent; the comparable average decline for external debt events is 1.2 percent. To compare the antecedents of domestic and external defaults, we performed a variety of tests for individual years, as well as for the cumulative change in the window prior to default. The latter test comprised a total of 224 observations for domestic crises (that is, the number of annual observations in advance of domestic crises) and 813 for external crashes (again, years multiplied by number of crises).

  As noted earlier, the results have to be interpreted with care, for many domestic episodes are twin default crises and, as a consequence, output is also suffering from limited access to external credit (if there is any at all).

  Inflation in the Run-up to and the Aftermath of Debt Defaults

  The comparable exercise for the inflation rate yields even starker differences (figures 9.3 and 9.4). Inflation during the year of an external default is on average high, at 33 percent.2 However, inflation truly gallops during domestic debt crises, averaging 170 percent in the year of the default.3 After the domestic default, inflation remains at or above 100 percent in the following years. Not surprisingly, default through inflation goes hand in hand with domestic default—before, during, and after the more explicit domestic expropriations. The extensive scholarly literature on inflation has been silent on this point.4 We conclude that overt domestic default tends to occur only in times of severe macroeconomic distress.

  Figure 9.1. Real GDP before, during, and after domestic and external debt crises, 1800–2008.

  Sources: Maddison (2004), Total Economy Database (2008), and the authors’ calculations.

  Note: Real GDP is indexed to equal 100 four years before the crisis.

  Figure 9.2. Domestic and external debt crises and real GDP, three years before crisis and year of crisis, 1800–2008.

  Sources: Maddison (2004), Total Economy Database (2008), and the authors’ calculations.

  Notes: The Kolmogorov-Smirnov test (K-S test) is used to determine whether two data sets differ significantly. The K-S test has the advantage of making no assumption about the distribution of data. The test is nonparametric and distribution free. Here Kolmogorov-Smirnov, 8.79; significant at 1 percent.

  Figure 9.3. Consumer prices before, during, and after domestic and external debt crises, 1800–2008.

  Sources: International Monetary Fund (various years), International Financial Statistics and World Economic Outlook; additional sources listed in appendix A.1; and the authors’ calculations.

  Note: Consumer prices are indexed to equal 100 four years before the crisis.

  Figure 9.4. Domestic and external debt crises and inflation, three years before crisis and year of crisis, 1800–2008.

  Sources: International Monetary Fund (various years), International Financial Statistics and World Economic Outlook; additional sources listed in appendix A.1; and the authors’ calculations.

  A more analytical approach in comparing real GDP growth and inflation across episodes of domestic and external debt defaults is shown in table 9.1. The columns of the table provide sample averages of economic growth and inflation in the run-up to and the aftermath of defaults on the two types of debt. The bottom row reports the results of a Kolmogorov-Smirnov test, which is a statistical test of the equality of two frequency distributions. Both real growth and inflation behave distinctly differently around domestic defaults than around external ones.

  TABLE 9.1

  Output and inflation around and during debt crises

  The Incidence of Default on Debts Owed

  to External and Domestic Creditors

  To shed some light on the incidence of expropriation of residents versus nonresidents, we constructed four time series for the period 1800–2006, showing the probability of external default (or the share of countries in our sample that are in external default in a given year); the comparable statistic for domestic default episodes; the probability of an inflation crisis (defined here as the share of countries in any given year during the more than two hundred years of our sample in which the annual rate of inflation exceeded 20 percent); and the sum of the incidence of high inflation and domestic default, which summarizes the expropriation of the holdings of domestic residents.5

  Figure 9.5 shows the probability of external default versus the comparable statistic for domestic default through either inflation or explicit default. Table 9.2 presents some summary statistics on the underlying data. During the early period and up to World War II, the incidence of external default was higher than it was later.6 Specifically, for 1800–1939, the probability of external default was about 20 percent versus 12 percent for domestic residents. For the entire sample, there was no statistically significant difference in the incidence of default on debts to domestic creditors versus foreigners. With the widespread adoption of fiat money, inflation apparently became the more expedient form of expropriation. As a result, the incidence of taxing domestic residents increased after World War II.7

  Figure 9.6 plots the probability of domestic default as a share of the probability of default. A ratio above 0.5 implies that domestic creditors do worse, while a ratio below 0.5 implies that foreigners do worse.

  Figure 9.5. Who is expropriated, residents or foreigners? The probability of domestic and external default, 1800–2006.

  Sources: International Monetary Fund (various years), International Financial Statistics and World Economic Outlook; Maddison (2004); Total Economy Database (2008); additional sources listed in appendix A.1; and the authors’ calculations.

  Figure 9.6. Composite probability of domestic default as a share of the total default probability, 1800–2006.

  Sources: International Monetary Fund (various years), International Financial Statistics and World Economic Outlook; Maddison (2004); Total Economy Database (2008); additional sources listed in appendix A.1; and the authors’ calculations.

  TABLE 9.2

  Who gets expropriated, residents or foreigners? Preliminary tests for the equality of two proportions (binomial distribution), 1800–2006

  Certainly, this admittedly very crude first pass at the evidence does nothing to dissuade us from our prior belief that domestic debt is often held by important political stakeholders in debtor countries and cannot always be lightly dismissed as strictly junior debt, a point highlighted by Allan Drazen.8

  Table 9.2 allows us to look more systematically at the difference in the treatment of domestic residents and foreigners over time. It reports sample averages of the probability of domestic and external defaults for the entire period from 1800 to 2006 and two sub-periods split at World War II. As is stated in the notes to the table, the probability that domestic residents would be expropriated was higher in the postwar period.

  Summary and Discussion of Selected Issues

  In the past few chapters we have provided an extensive new cross-country data set on a key macroeconomic variable that governments often manage to keep remarkably hidden from view: domestic public debt. We have also presented what we believe to be the first attempt at a cross-country international catalog of his
torical defaults on domestic public debt, spanning two centuries and sixty-six countries.

  Our first look at the data suggests that researchers need to revisit the empirical literature on the sustainability of external government debt and on governments’ incentives to engage in high inflation and hyperinflation, taking into account the newly uncovered data on domestic public debt and, where possible, broader definitions of government or government-guaranteed debt. Of course, how domestic debt impacts inflation and external default will vary across episodes and circumstances. In some cases, domestic debt is eliminated through high inflation; in other cases, governments default on external debt.

  How did domestic public debt in emerging markets fall off many economists’ radar screens? Many researchers, aware only of the difficulties of emerging markets in issuing debt in the ultra-high-inflation 1980s and 1990s, simply believed that no one would ever voluntarily lend in domestic currency to the kleptocratic government of an emerging market. Surely no one would trust such a government to resist inflating such debt down to nothing. The logical implication was that domestic currency public debt must not exist. True, a few researchers have contemplated the possibility. Alesina and Tabellini, for example, considered a theoretical case in which domestic debt was honored ahead of external debt.9 But absent any data or even any awareness of the earlier existence of significant quantities of domestic public debt in virtually every country (even during its emerging market phase), these isolated examples have had no great impact on the mainstream academic or policy literature.

  The lack of transparency exhibited by so many governments and multilateral institutions in failing to make time series on domestic debt easily available is puzzling. After all, these governments routinely tap domestic and foreign markets to sell debt. In general, uncertainty about a government’s past repayment performance is more likely than not going to raise risk premiums on new issuances. Even more puzzling is why global investors do not insist on historical information relevant to the value of securities they may purchase. Any credit card company will want to know a consumer’s purchase and repayment history, what kind of debt burdens the consumer has managed in the past, and under what circumstances. Surely historical information is equally relevant to governments.

  One can only surmise that many governments do not want capital markets to fully recognize the risks they are running by piling on debt and debt guarantees due to their fear of having to pay much higher financing costs. Publishing historical data will make investors ask why current data cannot be made equally available. Still, one would think a strong case could be made for less profligate governments to open up their books more readily and be rewarded for doing so by lower interest rates. This transparency, in turn, would put pressure on weaker borrowers. Yet today even the United States runs an extraordinarily opaque accounting system, replete with potentially costly off-budget guarantees. In its response to the most recent financial crisis, the U.S. government (including the Federal Reserve Board) took huge off–balance sheet guarantees onto its books, arguably taking on liabilities that, from an actuarial perspective—as evaluated at the time of the bailout—were of the same order of magnitude as, say, expenditures on defense, if not greater. Why so many governments do not make it easier for standard databases to incorporate their debt histories is an important question for future academic and policy research.

  From a policy perspective, a plausible case can be made that an international agency would be providing a valuable public good if it could enforce (or at least promote) basic reporting requirements and transparency across countries. Indeed, it is curious that today’s multilateral financial institutions have never fully taken up the task of systematically publishing public debt data, especially in light of these agencies’ supposed role at the vanguard of warning policy makers and investors about crisis risks. Instead, the system seems to have forgotten about the history of domestic debt entirely, thinking that today’s blossoming of internal public debt markets is something entirely new and different.10 But as our historical data set on domestic central government debt underscores with surprising force, nothing could be further from the truth. Indeed, we have reason to believe that with our data we have only touched the tip of the iceberg in terms of fully understanding public sector explicit and contingent liabilities.

  - PART IV -

  BANKING CRISES, INFLATION, AND

  CURRENCY CRASHES

  Countries can outgrow a history of repeated bouts with high inflation, but no country yet has graduated from banking crises.

  - 10 -

  BANKING CRISES

  Although many now-advanced economies have graduated from a history of serial default on sovereign debt or very high inflation, so far graduation from banking crises has proven elusive. In effect, for the advanced economies during 1800–2008, the picture that emerges is one of serial banking crises.

  Until very recently, studies of banking crises have focused either on episodes drawn from the history of advanced countries (mainly the banking panics before World War II) or on the experience of modern-day emerging markets.1 This dichotomy has perhaps been shaped by the belief that for advanced economies, destabilizing, systemic, multicountry financial crises are a relic of the past.2 Of course, the Second Great Contraction, the global financial crisis that recently engulfed the United States and Europe, has dashed this misconception, albeit at great social cost.

  As we will demonstrate in this chapter, banking crises have long impacted rich and poor countries alike. We develop this finding using our core sample of sixty-six countries (plus a broader extended sample for some exercises). We examine banking crises ranging from Denmark’s financial panic during the Napoleonic Wars to the recent “first global financial crisis of the twenty-first century.” The incidence of banking crises proves to be remarkably similar in both high-income and middle- to low-income countries. Indeed, the tally of crises is particularly high for the world’s financial centers: France, the United Kingdom, and the United States. Perhaps more surprising still are the qualitative and quantitative parallels across disparate income groups. These parallels arise despite the relatively pristine modern-day sovereign default records of the rich countries.

  For the study of banking crisis, three features of the data set underlying this book are of particular note. First, to reiterate, our data on banking crises go back to 1800. Second, to our knowledge, we are the first to examine the patterns of housing prices around major banking crises in emerging markets, including Asia, Europe, and Latin America. Our emerging market data set facilitates comparisons across both duration and magnitude with the better-documented housing price cycles in the advanced economies, which have long been known to play a central role in financial crises. We find that real estate price cycles around banking crises are similar in duration and amplitude across the two groups of countries. This result is surprising given that almost all other macroeconomic and financial time series (on income, consumption, government spending, interest rates, etc.) exhibit higher volatility in emerging markets.3 Third, our analysis employs the comprehensive historical data on central government tax revenues and central government debt compiled (as detailed in chapter 3). These data afford a new perspective on the fiscal consequences of banking crises—notably emphasizing the implications for tax revenue and public debt, which are far more substantive than the usual narrower focus in the literature, bailout costs.

  We find that banking crises almost invariably lead to sharp declines in tax revenues, while other factors leading to higher deficits can include the operation of automatic fiscal stabilizers, counter-cyclical fiscal policy, and higher interest payments due to elevated risk premiums and rating downgrades (particularly, but not exclusively, for emerging markets). On average, during the modern era, real government debt rises by 86 percent during the three years following a banking crisis. (That is, if central government debt was $100 billion at the start of a crisis, it would rise to $186 billion after three years, adjusted for inflation
.) These fiscal consequences, which include both direct and indirect costs, are an order of magnitude larger than the usual bank bailout costs. The fact that the magnitudes are comparable in advanced and emerging market economies is again quite remarkable. Obviously, both bailout costs and other fiscal costs depend on a host of political and economic factors, especially the policy response and the severity of the real shock that, typically, triggers the crisis.4

  A Preamble on the Theory of Banking Crises

  Banking Crises in Repressed Financial Systems

  Our sample includes basically two kinds of banking crises. The first is common in poor developing countries in Africa and elsewhere, although it occasionally surfaces in richer emerging markets, such as Argentina. These crises are really a form of domestic default that governments employ in countries where financial repression is a major form of taxation. Under financial repression, banks are vehicles that allow governments to squeeze more indirect tax revenue from citizens by monopolizing the entire savings and payments system, not simply currency. Governments force local residents to save in banks by giving them few, if any, other options. They then stuff debt into the banks via reserve requirements and other devices. This allows the government to finance a part of its debt at a very low interest rate; financial repression thus constitutes a form of taxation. Citizens put money into banks because there are few other safe places for their savings. Governments, in turn, pass regulations and restrictions to force the banks to relend the money to fund public debt. Of course, in cases in which the banks are run by the government, the central government simply directs the banks to make loans to it.