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

  9. Calvo (1998) credits the late Rudy Dornbusch, who quotes the old banking adage, “It is not the speed that kills you. It is the sudden stop” (Dornbusch et al. 1995).

  10. Reinhart, Rogoff, and Savastano (2003b) illustrated that countries with a history of external defaults also had a poor inflation track record.

  11. Fisher (1933).

  12. Domestic defaults produce even worse inflation outcomes; see chapter 9.

  13. See Calvo, Leiderman, and Reinhart (1993); Dooley et al. (1996); and Chuhan et al. (1998) for earlier papers quantifying the role of external factors influencing capital flows to emerging markets and their access to credit markets. On predicting defaults with domestic and some global factors, see Manasse and Roubini (2005).

  14. See Kaminsky, Reinhart, and Végh (2004) and Aguiar and Gopinath (2007).

  15. See the work of Reinhart and Reinhart (2009), who document the common occurrence of “capital inflow bonanzas” in the years preceding debt crises in emerging markets. It is of note that this analysis also shows that capital flow bonanzas precede banking crises in both advanced and emerging market economies.

  16. For a fuller account of the episode, see Baker (1994).

  17. Hale (2003).

  18. Box 5.3 summarizes some of the results in Reinhart, Rogoff, and Savastano (2003a), which presents empirical evidence of this “quick-to-releverage” pattern.

  19. Reinhart, Rogoff, and Savastano (2003a).

  Chapter 6 External Default through History

  1. See Reinhart, Rogoff, and Savastano (2003a); they thank Harold James for this observation.

  2. Winkler (1933), p. 29. One wonders if Thomas Jefferson read those words, in that he subsequently held that “the tree of liberty must be refreshed from time to time with the blood of patriots and tyrants.”

  3. Macdonald (2006).

  4. Ibid.

  5. Kindleberger (1989).

  6. Reinhart, Rogoff, and Savastano (2003a).

  7. For example, see Mauro et al. (2006).

  8. Latin American states were not the only ones borrowing at this time. Greece (still engaged in its struggle for independence), Portugal, and Russia were also issuing and placing pound-denominated bonds in London.

  9. For a fascinating fact-based account of this enterprise that reads like fiction, see David Sinclair’s 2004 book, The Land That Never Was: Sir Gregor MacGregor and the Most Audacious Fraud in History. Of the 250 settlers who crossed the Atlantic en route to Poyais (which was supposedly on the Bay of Honduras, site of the present Belize), only 50 survived to tell the tale.

  10. MacGregor was also able to raise an additional £40,000 through various channels for a total of £200,000, well above the £163,000 raised by the nonfictional Federation of Central American States during the 1822–1825 lending boom.

  Chapter 7 The Stylized Facts of

  Domestic Debt and Default

  1. Reinhart, Rogoff, and Savastano (2003a) drew on national sources to develop a data set for selected developing countries and emerging markets covering the years 1990–2002. More recently, Jeanne and Guscina (2006) provided detailed data on domestic debt for nineteen important emerging markets from 1980 to 2005. Cowan et al. (2006) provided data for all the countries in the Western Hemisphere from 1980 (or 1990) to 2004. Reinhart and Rogoff (2008a) described a companion database covering a broad range of related variables, including external debt, on which we also draw here.

  2. Domestic public debt has never amounted to much in a few Latin American countries (Uruguay stands out in this regard), and public debt markets are virtually nonexistent in the CFA African countries (originally the Colonies françaises d’Afrique).

  3. Of course, during the early years of the interwar period, many countries pegged their currencies to gold.

  4. It should also be noted that until the past ten to fifteen years, most countries’ external debt was largely public debt. Private external borrowing has become more significant only over the past couple of decades; see Prasad et al. (2003). Arellano and Kocherlakota (2008) developed a model of the relationship between private debt and external government default.

  5. See chapter 12 for a discussion of the aftermath and consequences of high inflation.

  6. See Calvo (1991) on these “perilous” practices.

  7. See Barro (1974).

  8. See Woodford (1995) on the former, Diamond (1965) on the latter.

  9. For example, Tabellini (1991) or Kotlikoff et al. (1988).

  10. Including a handful of very recent domestic defaults, the sample can in fact be extended to more than seventy domestic defaults.

  11. Brock (1989). The average reserve requirements for developing countries in Brock’s sample from 1960 to early 1980s ran at about 0.25, more than three times that for advanced economies.

  12. Another subtle type of default is illustrated by the Argentine government’s treatment of its inflation-indexed debt in 2007. Most impartial observers agree that during this period, Argentina’s official inflation rate considerably understated its actual inflation because of government manipulation. This understatement represented a partial default on index-linked debt by any reasonable measure, and it affected a large number of bondholders. Yet Argentina’s de facto domestic bond default did not register heavily in the external press or with rating agencies.

  Chapter 8 Domestic Debt

  1. One of the best standard sources is Maddison (2004).

  2. For instance, Bulow and Rogoff (1988b); Reinhart, Rogoff, and Savastano (2003a).

  3. See Reinhart, Rogoff, and Savastano (2003a).

  4. For example, the Kolmogorov-Smirnov test rejects the hypothesis that the two frequency distributions are equal at the 1 percent level.

  5. See Gavin and Perotti (1997) and Kaminsky, Reinhart, and Végh (2004) for evidence on procyclical macroeconomic policies. See also Aguiar and Gopinath (2007) for a model in which the procyclical behavior of the current account can be rationalized by the high ratio of permanent to transitory shocks in emerging markets.

  6. Of course, today’s rich countries experienced much the same problems in earlier eras—when they, too, were serial defaulters and they, too, exhibited highly procyclical fisal policy.

  7. See Cagan (1956). Seignorage revenue is simply the real income a government can realize by exercising its monopoly on printing currency. The revenue can be broken down into the quantity of currency needed to meet the growing transactions demand at constant prices and the remaining growth, which causes inflation, thereby lowering the purchasing power of existing currency. The latter effect is generally referred to as the inflation tax. In a classic paper, Sargent (1982) includes data on central banks’ holdings of treasury bills after World War I for five countries (Austria, Czechoslovakia, Germany, Hungary, and Poland). But of course, these debts are essentially a wash on the consolidated government balance sheet.

  8. See Dornbusch and Fischer (1993).

  9. Of course, the possibility of using unanticipated inflation to default on nominal debt is well understood in the theoretical literature, such as Barro (1983).

  10. The case of Brazil is exceptional in that some of the debt was indexed to inflation, although lags in the indexation scheme still made it possible for the government to largely inflate away the debt with a high enough rate of inflation. Indeed, this appears to be exactly what happened, for the country lurched in and out of hyperinflation for many years.

  11. Calvo and Guidotti (1992) developed a model of the optimal maturity structure of nominal debt whereby the government trades off flexibility (the option to inflate away long-term debt when under financial duress) versus its high credibility for maintaining a low inflation rate (achieved by having very short-term debt, which is more difficult to inflate away).

  Chapter 9 Domestic and External Default

  1. It should also be noted that other economic indicators (besides inflation and per capita GDP growth, which we examine in detail) would provide a richer answer to the broad question of ho
w bad conditions have to be before a country should contemplate default. (Specifically, the impacts of domestic versus foreign default on social indicators relating to poverty, health, income distribution, and so on, are, in principle, bound to be quite different.)

  2. It is hardly surprising that inflation rises in the aftermath of external default, especially given the typical massive exchange rate depreciation.

  3. We have excluded Bolivia’s 1982 domestic default from these averages, because inflation peaked at over 11,000 percent in the year before (t − 1) the domestic default.

  4. Reinhart and Savastano (2003) discuss the forcible conversion of foreign currency bank deposits (as was also seen in Argentina in 2002) during the hyperinflations in Bolivia and Peru.

  5. The United States, of course, is the modern exception. Virtually all U.S. debt is domestic (as the Carter bonds have matured). Yet about 40 percent is held by nonresidents (mostly central banks and other official institutions), and all is dollar denominated. Thus, inflation in the United States would also affect nonresidents.

  6. The huge spike in external defaults in the 1820s was due to the much-studied first wave of sovereign defaults of the newly independent Latin American countries. But Greece and Portugal also defaulted at that time.

  7. See figure 9.6.

  8. Drazen (1998).

  9. See Alesina and Tabellini (1990).

  10. Beyond simply reporting debt data, international financial institutions such as the International Monetary Fund and the World Bank can also help with disseminating information on best practices (see, for example, the institutional evolution discussed in Wallis and Weingast 1988).

  Chapter 10 Banking Crises

  1. See Gorton (1988) and Calomiris and Gorton (1991) on banking panics before World War II; Sundararajan and Baliño (1991) for several case studies of emerging markets; and Jácome (2008) on banking crises in Latin America.

  2. Studies that encompass episodes in both advanced and emerging economies include those of Demirgüç-Kunt and Detragiache (1998), Kaminsky and Reinhart (1999), and Bordo et al. (2001).

  3. See, for instance, Agénor et al. (2000).

  4. Reinhart and Rogoff (2008a, 2008c) show that output growth typically decelerates in advance of a crisis.

  5. See Diamond and Dybvig (1983).

  6. Ibid.

  7. See Bernanke and Gertler (1990).

  8. See Kiyotaki and Moore (1997).

  9. See, for example, Bernanke and Gertler (1995).

  10. See Bernanke et al. (1999).

  11. This refers to default on external debt. The widespread abrogation of gold clauses and other forms of restructuring—on domestic debt—by the United States and other developed economies during the Great Depression of the 1930s are considered sovereign defaults on domestic debt (debt issued under domestic law).

  12. As we have already acknowledged, our accounting of financial crises in poorer countries may be incomplete, especially for earlier periods, despite our best efforts.

  13. To be precise, the advanced countries experienced 7.2 crises, on average, versus 2.8 for emerging market countries.

  14. See Obstfeld and Taylor (2004).

  15. See Kaminsky and Reinhart (1999).

  16. See Demirgüç-Kunt and Detragiache (1998). See also Drees and Pazarbasioglu (1998) for an insightful discussion of the Nordic experience with financial liberalization.

  17. See Caprio and Klingebiel (1996).

  18. To define a capital flow bonanza, Reinhart and Reinhart (2009) settled on an algorithm that provided uniform treatment across countries but was flexible enough to allow for significant cross-country variation in the current account. Like Kaminsky and Reinhart (1999), they selected a threshold to define bonanzas that is common across countries (in this case, the 20th percentile of the sample). This threshold included most of the better known episodes in the literature but was not so inclusive as to label as a bonanza a more “routine” deterioration in the current account. Because the underlying frequency distributions vary widely across countries, the common threshold produces quite disperse country-specific cutoffs. For instance, in the case of relatively closed India, the cutoff to define a bonanza is a ratio of current account deficit to GDP in excess of 1.8 percent, while for trade-oriented Malaysia the comparable cutoff is a ratio of deficit to GDP of 6.6 percent.

  19. Reinhart and Reinhart performed comparable exercises for currency, debt, and inflation crises.

  20. See Reinhart and Reinhart (2009).

  21. See Mendoza and Terrones (2008). See also the work of Kaminsky and Reinhart (1999), who also examined the growth in real credit available to the private sector around both banking and currency crises.

  22. See Reinhart and Rogoff (2008b). Each year refers to the beginning of the crisis.

  23. See Bordo and Jeanne (2002).

  24. See Gerdrup (2003).

  25. Historical comparisons are hard to come by, because most real housing price series are of recent vintage, but we do include in this category two older episodes: that of the United States during the Great Depression and that of Norway at the turn of the century (1898).

  26. See the work of Ceron and Suarez (2006), who estimate its average duration at six years.

  27. For example, Agénor et al. (2000) provide evidence that output and real consumption are far more volatile in emerging markets; Kaminsky, Reinhart, and Végh (2003) present evidence that the amplitude of the cycle in real government spending is orders of magnitude greater in emerging markets.

  28. This notion is consistent with the fact that house prices are far more inertial than equity prices.

  29. See Philippon (2007).

  30. See Frydl (1999) and Norges Bank (2004). See also Sanhueza (2001), Hoggarth et al. (2005), and Caprio et al. (2005).

  31. A similar problem plagues work on determining the effectiveness of foreign exchange intervention by measuring the profitability of such market purchases or sales. The results depend significantly on the width of the time window and on implicit assumptions of the cost of financing. See Neely (1995).

  32. See Vale (2004).

  33. See, for instance, Frydl (1999), Kaminsky and Reinhart (1999), and especially Rajan et al. (2008), who examine the output consequences of the credit channel following banking crises using micro data. We note that, of the cases of output collapses studied by Barro and Ursúa (2008), virtually all are associated with banking crises.

  34. Revenues (from Mitchell, 2003a, 2003b) are deflated by consumer price indexes; the numerous sources of these data are given on a country-by-country and period-by-period basis in appendix A.2.

  35. Indeed, in some important cases, such as that of Japan, the accelerated debt buildup goes on for more than a decade, so the three-year cutoff grossly understates the longer-term consequences.

  36. Note that figure 10.10 gives the percentage of change in debt rather than debt relative to GDP in order not to distort the numbers by the large falls in GDP that sometimes accompany crises. However, the same basic message comes across looking at debt relative to GDP instead. Note that the calculations are based on total central government debt.

  37. An important question is how rare banking crises, through sudden changes in market liquidity, might amplify the effects on asset prices, as analyzed by Barro (2009).

  Chapter 11 Default through Debasement

  1. See, for example, Sargent and Velde (2003). Ferguson (2008) provides an insightful discussion of the early roots of money.

  2. See Winkler (1928).

  3. MacDonald (2006).

  Chapter 12 Inflation and Modern Currency Crashes

  1. Végh (1992) and Fischer et al. (2002) are essential readings for discussions of the literature on relatively modern inflation crises.

  2. The dates in table 12.2 extend back prior to the time of independence for many countries, including, for example, Malaysia.

  3. China, which invented the printing press well ahead of Europe, famously experienced episodes of high i
nflation created by paper currency in the twelfth and thirteen centuries. (For more on this subject, see, for example, Fischer et al. 2002.) These episodes are in our database as well.

  4. Reinhart and Rogoff (2002).

  5. Cagan (1956).

  6. At the time of this writing, the “official” inflation rate in Argentina is 8 percent; informed estimates place it at 26 percent.

  7. Reinhart, Rogoff, and Savastano (2003b).

  8. The following section examines in detail the experience of countries that have recorded large declines in their degree of domestic dollarization, including in the context of disinflations.

  9. Reinhart and Rogoff (2004) show that countries with dual exchange rate systems tend to have worse average growth performances and vastly worse inflation performances.

  10. This pattern was particularly common in the second half of the 1990s among the transition economies (e.g., Azerbaijan, Belarus, Lithuania, and Russia) but was also present in other countries and periods—for instance, in Bolivia and Peru in the early 1980s and in Egypt in the mid-1990s.

  11. See Bufman and Leiderman (1992).

  12. See Dornbusch and Werner (1994).

  Chapter 13 The U.S. Subprime Crisis

  1. As indicated in note 7 to the preamble, we use the term “Second Great Contraction” after Friedman and Schwartz’s (1963) depiction of the 1930s as “The Great Contraction.” See also Felton and Reinhart (2008, 2009), who use the term “First Global Financial Crisis of the 21st Century.”

  2. See chapter 10 for further discussion.

  3. We have explored this issue further in chapter 10.

  4. Although China’s heavy-handed capital controls shielded it from contagious currency crashes during Asia’s turmoil, they did not protect it from a systemic and costly banking crisis emanating primarily from large-scale lending to inefficient and bankrupt state-owned enterprises.