The Twin Crises The Causes of Banking and Balance-of-Payments Problems

This Version: November 9, 1998

The Twin Crises: The Causes of Banking and

Balance-of-Payments Problems

Graciela L. Kaminsky Carmen M. Reinhart*

AbstractIn the wake of the Mexican and Asian currency turmoil, the subject of financial crises has come to

the forefront of academic and policy discussions. This paper analyzes the links between banking andcurrency crises. We find that: problems in the banking sector typically precede a currency crisis--the

currency crisis deepens the banking crisis, activating a vicious spiral; financial liberalization often precedesbanking crises. The anatomy of these episodes suggests that crises occur as the economy enters a

recession, following a prolonged boom in economic activity that was fueled by credit, capital inflows andaccompanied by an overvalued currency. (JEL F30, F41)

* Graciela L. Kaminsky, George Washington University, Washington, D.C. 20552. Carmen M. Reinhart,University of Maryland, College Park, Maryland 20742. We thank two anonymous referees for veryhelpful suggestions. We also thank Guillermo Calvo, Rudiger Dornbusch, Peter Montiel, VincentReinhart, John Rogers, Andrew Rose and seminar participants at Banco de México, the Board of

Governors of the Federal Reserve System, Florida State University, Harvard, the IMF, Johns HopkinsUniversity, Massachusetts Institute of Technology, Stanford University, SUNY at Albany, University ofCalifornia, Berkeley, UCLA, University of California, Santa Cruz, University of Maryland, University ofWashington, The World Bank, and the conference on “Speculative Attacks in the Era of the GlobalEconomy: Theory, Evidence, and Policy Implications,” (Washington, DC, December 1995), for veryhelpful comments and Greg Belzer, Kris Dickson, and Noah Williams for superb research assistance.

Pervasive currency turmoil, particularly in Latin America in the late 1970s and early 1980s, gave

impetus to a flourishing literature on balance-of-payments crises. As stressed in Paul Krugman’s (1979)seminal paper, in this literature crises occur because a country finances its fiscal deficit by printing moneyto the extent that excessive credit growth leads to the eventual collapse of the fixed exchange rate regime. With calmer currency markets in the mid- and late 1980s, interest in this literature languished. Thecollapse of the European Exchange Rate Mechanism, the Mexican peso crisis, and the wave of currencycrises sweeping through Asia have, however, rekindled interest in the topic. Yet, the focus of this recentliterature has shifted. While the earlier literature emphasized the inconsistency between fiscal andmonetary policies and the exchange rate commitment, the new one stresses self-fulfilling expectations andherding behavior in international capital markets.1 In this view, as Guillermo A.Calvo (1995, page 1)summarizes “If investors deem you unworthy, no funds will be forthcoming and, thus, unworthy you willbe.”

Whatever the causes of currency crises, neither the old literature nor the new models of

self-fulfilling crises have paid much attention to the interaction between banking and currency problems,despite the fact that many of the countries that have had currency crises have also had full-fledged domesticbanking crises around the same time. Notable exceptions are: Carlos Diaz-Alejandro (1985), AndresVelasco (1987), Calvo (1995), Ilan Goldfajn and Rodrigo Valdés (1995), and Victoria Miller (1995). Asto the empirical evidence on the potential links between what we dub the twin crises, the literature has beenentirely silent. The Thai, Indonesian, and Korean crises are not the first examples of dual currency andbanking woes, they are only the recent additions to a long list of casualties which includes Chile, Finland,Mexico, Norway, and Sweden.

In this paper, we aim to fill this void in the literature and examine currency and banking crises

episodes for a number of industrial and developing countries. The former include: Denmark, Finland,Norway, Spain, and Sweden. The latter focus on: Argentina, Bolivia, Brazil, Chile, Colombia, Indonesia,

1

Israel, Malaysia, Mexico, Peru, the Philippines, Thailand, Turkey, Uruguay, and Venezuela. The periodcovered spans the 1970s through 1995. This sample gives us the opportunity to study 76 currency crisesand 26 banking crises. Out-of sample, we examine the twin crises in Asia of 1997.

Charles Kindelberger (1978, page 14), in studying financial crises, observes: “For historians each

event is unique. Economics, however, maintains that forces in society and nature behave in repetitive ways.History is particular; economics is general.” Like Kindelberger, we are interested in finding the underlyingcommon patterns associated with financial crises. To study the nature of crises, we construct a chronologyof events in the banking and external sectors. From this timetable, we draw inference about the possiblecausal patterns among banking and balance-of-payments problems and financial liberalization. We alsoexamine the behavior of macroeconomic indicators that have been stressed in the theoretical literaturearound crisis periods, much along the lines of Barry Eichengreen et. al. (1996a). Our aim is to gaugewhether the two crises share a common macroeconomic background. This methodology also allows us toassess the fragility of economies around the time of the financial crises and sheds light on the extent towhich the crises were predictable. Our main results can be summarized as follows:

First, with regard to the linkages among the crises, our analysis shows no apparent link between

balance of payments and banking crises during the 1970s, when financial markets were highly regulated. In the 1980s, following the liberalization of financial markets across many parts of the world, banking andcurrency crises become closely entwined. Most often, the beginning of banking sector problems predatethe balance of payment crisis; indeed, knowing that a banking crisis was underway helps predict a futurecurrency crisis. The causal link, nevertheless, is not unidirectional. Our results show that the collapse ofthe currency deepens the banking crisis, activating a vicious spiral. We find that the peak of the bankingcrisis most often comes after the currency crash, suggesting that existing problems were aggravated or newones created by the high interest rates required to defend the exchange rate peg or the foreign exchangeexposure of banks.

2

Second, while banking crises often precede balance of payments crises, they are not necessarily the

immediate cause of currency crises, even in the cases where a frail banking sector puts the nail in the coffinof what was already a defunct fixed exchange rate system. Our results point to common causes, andwhether the currency or banking problems surface first is a matter of circumstance. Both crises arepreceded by recessions or, at least, below normal economic growth, in part attributed to a worsening of theterms of trade, an overvalued exchange rate, and the rising cost of credit; exports are particularly hard hit. In both types of crises, a shock to financial institutions (possibly financial liberalization and/or increasedaccess to international capital markets) fuels the boom phase of the cycle by providing access to financing.The financial vulnerability of the economy increases as the unbacked liabilities of the banking system climbto lofty levels.

Third, our results show that crises (external or domestic) are typically preceded by a multitude of

weak and deteriorating economic fundamentals. While speculative attacks can and do occur as marketsentiment shifts and, possibly, herding behavior takes over (crises tend to be bunched together), theincidence of crises where the economic fundamentals were sound are rare.

Fourth, when we compared the episodes in which currency and banking crises occurred jointly to

those in which the currency or banking crisis occurred in isolation, we find that for the twin crises,economic fundamentals tended to be worse, the economies were considerably more frail, and the crises(both banking and currency) were far more severe.

The rest of the paper is organized as follows. The next section provides a chronology of the crises

and their links. Section II reviews the stylized facts around the periods surrounding the crises while SectionIII addresses the issues of the vulnerability of economies around the time of the crisis and the issue ofpredictability. The final section discusses the findings and possibilities for future research.I. The Links Between Banking and Currency Crises

This section briefly discusses what the theoretical literature offers as explanations of the possible

3

links between the two crises. The theoretical models also guide our choice of the financial and economicindicators used in the analysis.A. The links: theory

A variety of theoretical models have been put forth to explain the linkages between currency and

banking crises. One chain of causation, stressed in James Stoker (1995), runs from balance of paymentsproblems to banking crisis. An initial external shock, such as an increase in foreign interest rates, coupledwith a commitment to a fixed parity will result in the loss of reserves. If not sterilized, this will lead to acredit crunch, increased bankruptcies, and financial crisis. Moreover, Frederic Mishkin (1996) argues that,if a devaluation occurs, the position of banks could be weakened further if a large share of their liabilities isdenominated in a foreign currency. Models, such as Velasco (1987), point to the opposite causal direction--financial sector problems give rise to the currency collapse. Such models stress that when central banksfinance the bail-out of troubled financial institutions by printing money, we return to the classical story of acurrency crash prompted by excessive money creation.

A third family of models contend that currency and banking crises have common causes. An

example of this may be found in the dynamics of an exchange rate-based inflation stabilization plan, suchas that of Mexico in 1987. Theory and evidence suggest that such plans have well-defined dynamics: 2 Because inflation converges to international levels only gradually, there is a marked cumulative realexchange rate appreciation. Also, at the early stages of the plan there is a boom in imports and economicactivity, financed by borrowing abroad. As the current account deficit continues to widen, financialmarkets become convinced that the stabilization program is unsustainable, fueling an attack against thedomestic currency. Since the boom is usually financed by a surge in bank credit, as banks borrow abroad,when the capital inflows become outflows and asset markets crash, the banking system caves in. Ronald I.McKinnon and Huw Pill (1996) model how financial liberalization together with microeconomic

distortions--such as implicit deposit insurance--can make these boom-bust cycles even more pronounced by

4

fueling the lending boom that leads to the eventual collapse of the banking system. Ilan Goldfajn andRodrigo Valdés (1995) show how changes in international interest rates and capital inflows are amplifiedby the intermediating role of banks and how such swings may also produce an exaggerated business cyclethat ends in bank runs and financial and currency crashes.

So, while theory does not provide an unambiguous answer as to what the causal links between

currency and banking crises are, the models are clear as to what economic indicators should provideinsights about the underlying causes of the twin crises. High on that list are international reserves, ameasure of excess money balances, domestic and foreign interest rates, and other external shocks, such asthe terms of trade. The inflation stabilization-financial liberalization models also stress the boom-bustpatterns in imports, output, capital flows, bank credit, and asset prices. Some of these models alsohighlight overvaluation of the currency, leading to the underperformance of exports. The possibility of bankruns suggests bank deposits as an indicator of impending crises. Finally, as in Krugman (1979) currencycrises can be the byproduct of government budget deficits.B. The Links: Preliminary Evidence

To examine these links empirically, we first need to identify the dates of currency and banking

crises. In what follows, we begin by describing how our indices of financial crises are constructed.Definitions, dates, and incidence of crises

Most often, balance of payments crises are resolved through a devaluation of the domestic currency

or the floatation of the exchange rate. But central banks can and, on occasion, do resort to contractionarymonetary policy and foreign exchange market intervention to fight the speculative attack. In these lattercases, currency market turbulence will be reflected in steep increases in domestic interest rates and massivelosses of foreign exchange reserves. Hence, an index of currency crises should capture these differentmanifestations of speculative attacks. In the spirit of Eichengreen, et. al. (1996 a and b), we constructed anindex of currency market turbulence as a weighted average of exchange rate changes and reserve changes.3

5

With regard to banking crises, our analysis stresses events. The main reason for following this

approach has to do with the lack of high frequency data that capture when a financial crisis is underway. Ifthe beginning of a banking crisis is marked by a bank runs and withdrawals, then changes in bank depositscould be used to date the crises. Often, the banking problems do not arise from the liability side, but from aprotracted deterioration in asset quality, be it from a collapse in real estate prices or increased bankruptciesin the nonfinancial sector. In this case, changes in asset prices or a large increase in bankruptcies ornonperforming loans could be used to mark the onset of the crisis. For some of the earlier crises inemerging markets, however, stock market data is not available.4 Indicators of business failures andnonperforming loans are also usually available only at low frequencies, if at all; the latter are also madeless informative by banks’ desire to hide their problems for as long as possible.

Given these data limitations, we mark the beginning of 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 financialinstitutions (as in Venezuela 1993); and (2) if there are no runs, the closure, merging, takeover, orlarge-scale government assistance of an important financial institution (or group of institutions), that marksthe start of a string of similar outcomes for other financial institutions (as in Thailand 1996-97). We relyon existing studies of banking crises and on the financial press; according to these studies the fragility ofthe banking sector was widespread during these periods. This approach to dating the beginning of thebanking crises is not without drawbacks. It could date the crises too late, because the financial problemsusually begin well before a bank is finally closed or merged; it could also date the crises too early, becausethe worst of crisis may come later. To address this issue we also date when the banking crisis hits its peak,defined as the period with the heaviest government intervention and/or bank closures.

Our sample consists of 20 countries for the period 1970-mid-1995. The countries are those listed in

the introduction and Appendix Tables 1 and 2. We selected countries on the multiple criteria of beingsmall open economies, with a fixed exchange rate, crawling peg, or band through portions of the sample;

6

data availability also guided our choices. This period encompasses 26 banking crises and 76 currencycrises.

As to the incidence of the crises (Table 1 and Figure 1), there are distinct patterns across decades.

During the 1970s we observe a total of 26 currency crises, yet banking crises were rare during that period,with only 3 taking place. The absence of banking crises may reflect the highly regulated nature of financialmarkets during the bulk of the 1970s. By contrast, while the number of currency crises per year does notincrease much during the 1980s and 1990s (from an average of 2.60 per annum to 3.13 per annum, Table1, first row), the number of banking crises per year more than quadruples in the post-liberalization period. Thus, as the second row of Table 1 higlights, the twin crisis phenomenon is one of the 1980s and 1990s.

Figure 1 also shows that financial crises were heavily bunched in the early 1980s, when real

interest rates in the United States were at their highest level since the 1930s. This may suggest that,external factors, such as interest rates in the United States, matter a great deal as argued in Calvo, et. al.(1993). Indeed, Jeffrey Frankel and Andrew K. Rose (1996) find that foreign interest rates play asignificant role in predicting currency crashes. A second explanation why crises are bunched is thatcontagion effects may be present, creating a domino effect among those countries that have anything lessthan immaculate fundamentals. Sara Calvo and Reinhart (1996) present evidence of contagion in capitalflows to Latin American countries while Eichengreen, et. al.(1996b) find evidence that knowing there is acrisis elsewhere increases the probability of a domestic currency crisis.

Table 2 provides the dates of financial liberalization, the beginning and peak of the banking crisis,

and the date of the balance of payments crisis that was nearest to the beginning of the banking crisis.5 Byselecting the nearest currency crisis, whether it predates or follows the beginning of the banking crisis, weallow the data to reveal what the temporal patterns are. The dates for the remaining crises are given in theAppendix tables.The twin crises

7

We next examine how the currency and banking crises are linked. We begin by calculating the

unconditional probability of currency crises and banking crises in our sample. For instance, the probabilitythat a currency crisis will occur in the next 24 months over the entire sample is simply 24 times 76 (thetotal number of currency crises in the sample) divided by the total number of monthly observations in thesample. These calculations yield unconditional probabilities for currency and banking crises, which aretwenty nine percent and ten percent, respectively (Table 3). The difference in the probabilities of the twokinds of crises highlights the relatively higher frequency of currency crises in the sample.

We next calculate a family of conditional probabilities. For instance, if knowing that there is a

banking crisis within the past 24 months helps predict a currency crisis then, the probability of a currencycrisis, conditioned on information that a banking crisis is underway, should be higher than the

unconditional probability of a balance of payments crisis. In other words, a banking crisis increases theprobability that a country will fall prey to a currency crisis. This is precisely what the results summarizedin Table 3 show. The probability of a currency crisis conditioned on the beginning of banking sectorproblems is 46 percent, well above the unconditional estimate 29 percent. Hence, it could be argued, asDiaz Alejandro (1985) and Velasco (1987) did for the Chilean crisis in the early 1980s, that, in animportant number of cases, the bail-out of the banking system may have contributed to the acceleration incredit creation observed prior to the currency crises (see Herminio Blanco and Peter M. Garber (1986),Sebastian Edwards (1989), and Eichengreen et. al. (1996a), and this paper). Even in the absence of alarge-scale bail-out, a frail banking system is likely to tie the hands of the central bank in defending thecurrency--witness Indonesia in August 1997.

If, instead, the peak of the banking crisis is used as the conditioning piece of information, no

valuable information is gained; indeed, the conditional probability is 22 percent and below the

unconditional. This result follows from the fact that a more common pattern (see Table 2) appears to bethat the peak of the banking crisis comes after the currency crisis. For instance, knowing that there is a

8

currency crisis does not help predict the onset of a banking crisis, this conditional probability is 8 percent;knowing that there was a currency crisis does help to predict the probability that the banking crisis willworsen, this conditonal probability is 16 percent.

Taken together, these results seem to point to the existence vicious circles. Financial sector

problems undermine the currency. Devaluations, in turn, aggravate the existing banking sector problemsand create new ones. These adverse feedback mechanisms are in line with those suggested by Mishkin(1996) and can be amplified, as we have seen in several of the recent Asian crises, by banks’ inadequatehedging of foreign exchange risk. The presence of vicious circles would imply that, a priori, the twin crisesare more severe than currency or banking crises that occur in isolation.

To measure the severity of a currency crisis, we focus on a composite measure that averages

reserve losses and the real exchange rate depreciation.6 For reserves, we use the six-month percent changeprior to the crisis month, as reserve losses typically occur prior to the devaluation (if the attack issuccessful). For the real exchange rate, we use the six-month percent change following the crisis month,because large depreciations occur after, and only if, the central bank concedes by devaluing or floating thecurrency. This measure of severity is constructed for each currency crisis in our sample and the averagesare reported in Table 4 separately for the 19 twin crises in our sample and for the others. In line with ourresults that the beginning of the banking crisis precedes the balance of payments crisis, we define the twincrises as those episodes in which a currency crisis follows the beginning of the banking crisis within thenext 48 months. For banking crises, we use the bail-out costs, as a percent of GDP, as the measure ofseverity. As Table 4 highlights, bail-out costs are significantly larger (more than double) in the twin crisesthan for banking crises which were not accompanied by a currency crisis. As to balance of paymentscrises, the results are mixed. Reserve losses sustained by the central bank are significantly bigger (Table 4)but the real depreciations are of comparable orders of magnitude.

Our results also yield an insight as to the links of crises with financial liberalization (Table 3). In

9

18 of the 26 banking crises studied here, the financial sector had been liberalized during the preceding fiveyears, usually less. Only in a few cases in our sample countries, such as the early liberalization efforts ofBrazil in 1975 and Mexico in 1974, was the liberalization not followed by financial sector stress. In the1980s and 1990s most liberalization episodes have been associated with financial crises of varying severity. Only in a handful of countries (for instance, Canada which is not in the sample) did financial sectorliberalization proceed smoothly. Indeed, the probability of a banking crisis (beginning) conditional onfinancial liberalization having taken place is higher than the unconditional probability of a banking crisis.This suggests that the twin crises may have common origins in the deregulation of the financial system andthe boom-bust cycles and asset bubbles that, all too often, accompany financial liberalization. The stylizedevidence presented in Gerald Caprio and Daniela Klingebiel (1996) suggests that inadequate regulation andlack of supervision at the time of the liberalization may play a key role in explaining why deregulation andbanking crises are so closely entwined.

II. The Macroeconomic Background of the Crises

To shed light on whether both types of crises may have common roots, we analyze the evolution of

16 macroeconomic and financial variables around the time of the crises. The variables used in the analysiswere chosen in light of theoretical considerations and subject to data availability. Monthly data was usedto get a clearer view (than would otherwise be revealed by lower frequency data) of developments as thecrisis approaches and by the desire to evaluate to what extent these indicators were giving an early signal ofimpending trouble--an issue that will be taken up in the next section.

The indicators associated with financial liberalization are the M2 multiplier, the ratio of domestic

credit to nominal GDP, the real interest rate on deposits, and the ratio of lending-to-deposit interest rates.Other financial indicators include: excess real M1 balances, real commercial bank deposits, and the ratioof M2 (converted into U.S. dollars) divided by foreign exchange reserves (in U.S. dollars).7 The indicatorslinked to the current account include the percent deviation of the real exchange rate from trend, as a

10

measure of misalignment, the value of exports and imports (in U.S. dollars), and the terms-of-trade.8 Theindicators associated with the capital account are: foreign exchange reserves (in U.S. dollars) and thedomestic-foreign real interest rate differential on deposits (monthly rates in percentage points). Theindicators of the real sector are industrial production and an index of equity prices (in U.S. dollars). 9Lastly, the fiscal variable is the overall budget deficit as a percent of GDP.

Of course, this is not an exhaustive list of potential indicators. In particular, political variables,

such as the timing of an election, can also be linked to the timing of these crises. Indeed, the evidencepresented in Deepak Mishra (1997), who examines a subset of the currency crises in this study, suggeststhat devaluations, more often than not, follow elections. Indeed, an election raises the probability of afuture devaluation, even after controlling for economic fundamentals.

Except for the interest rate variables, the deviations of the real exchange rate from trend, our proxy

for excess real M1 balances, and the lending/deposit interest rates ratio, which are in levels, we focus onthe 12-month percent changes of the remaining 10 variables. The pre- and post-crises behavior of allvariables is compared to the average behavior during tranquil periods, which are all the remainingobservations in our sample and serves as our control group.

Figures 2, 3, and 4 illustrate the behavior of the variables around the time of the balance of

payments, banking crises, and twin crises, respectively; each panel portrays a different variable. Thehorizontal axis records the number of months before and after the beginning of the crises; the vertical axisrecords the percent difference (percentage point difference for interest rates) between tranquil and crisisperiods. In all the figures the solid line represents the average for all the crises for which data wasavailable.10 Hence, if no data points are missing, the solid line represents the average behavior of thatindicator during the months around 76 currency crises and 26 banking crises. For Figures 2 and 3, thedotted lines denote plus/minus one standard error around the average. For example, the top center panel ofFigure 2 shows that, on average, the 12-month growth in the domestic credit/GDP ratio is about 15 percent

11

higher than in tranquil times. In Figure 4 the solid line shows the evolution of the indicators for the twincrises episodes while the dashed line denotes the averages for the currency crises that were notaccompanied by a banking crisis.

For currency crises we focus on the 18-month period before and after the crisis. Unlike balance of

payments crises, in which reserves are lost abruptly and currency pegs abandoned, banking crises areprotracted affairs which tend to come in waves and, hence, the depth of the crisis is seldom reached at thefirst sign of outbreak (see Table 2). For this reason, we widen the window and focus on the 18 monthsbefore the onset of the crisis, a 18-month arbitrarily chosen crisis period, and the 18 months post-crisisperiod. At any rate, because most of our analysis focuses on the causes leading up to the crises, our mainresults will not be affected whether the crises lasted less or more than a year. For the 19 episodes of thetwin crises, we focus on the 18 months prior to the balance of payments crisis. Given that banking crisesusually predate currency crises in our sample, this implies we are already looking at a period of heavyfinancial sector stress.A. The Financial Sector

Until the 1970s, most financial markets were regulated with rationed credit and, often, negative

real interest rates. The late 1970s and beginning of the 1980s, however, witnessed sweeping financialreforms both in developed and emerging markets, which led to, among other things, increases in realinterest rates. 11 Because financial liberalization often precedes banking crises--the indicators associatedwith financial liberalization presented in the first four panels of Figures 2, 3, and 4 (from left to right) meritscrutiny. The growth in the M2 multiplier rises steadily up to nine months prior to the currency crisis andthe onset of the banking crisis; indeed, for banking crises the multiplier grows at above normal rate in theentire 18 months prior to the crisis. The draconian reductions in reserve requirements that oftenaccompany financial liberalization play a role in explaining the large increases in the M2 multiplier. Yetthe rise in the multiplier prior to currency crises is entirely accounted for by its evolution ahead of the twin

12

crises, as shown in Figure 4.

The growth in domestic credit/ GDP remains above normal as the balance of payments crisis nears

(Figure 2) but particularly accelerating markedly as the twin crises approaches; throughout this period itremains well above the growth rates recorded for tranquil periods, consistent with a credit boom (and bust)story. This ratio also rises in the early phase of the banking crisis. It may be that, as the crisis unfolds, thecentral bank may be pumping money to the banks to alleviate their financial situation or the evolution of thedenominator has changed. While credit is rapidly expanding 18 to 6 months before the crisis, the economyis still in a vigorous expansion phase (see below), with healthy GDP growth. The leveraging of householdsand business becomes evident as the economy slips into recession. The real interest rate evolves verydifferently ahead of balance of payments and banking crises. For currency crises, interest rates bouncearound in the range of 0 to 2 percentage points per month below the average during periods of tranquility--this may reflect lax monetary policy ahead of the currency crisis or simply the fact that 26 of the currencycrises are in the 1970s, when interest rates were regulated and not particularly informative. By contrast,prior to banking crises and, therefore, twin crises (which are almost exclusively in the post-liberalizationpart of the sample) real interest rates are 1 to 2 percentage points higher (at a monthly rate) than in tranquiltimes in the pre-crisis period. The above normal real interest rates may have a variety of causes: Thesecould be the product of a recent financial liberalization; high real rates could also reflect increasedrisk-taking by banks;12 they could be the product of a tight monetary policy stance. Real interest rates donot return to their levels in tranquil times as the crisis deepens, perhaps reflecting that banks may respondto deposit withdrawals by keeping deposit interest rates high. The lending-deposit rate ratio hoversaround its level in tranquil times up until about six months prior to the balance of payments crises and thenbegins to climb; by the time of the crisis it is about 10 percent higher than in tranquil times, possiblyreflecting a deterioration in credit risk. For banking crises, the lending/deposit ratio remains close normallevels in the pre-crisis period. Only at around the peak of the banking crises does the lending/deposit ratio

13

increase above its level in tranquil times, as banks become increasingly unwilling to lend.

The next three panels show the evolution of the monetary indicators. The middle panel in the

second row of Figures 2 and 3 show the excess M1 balances. The periods prior to the currency andbanking crises are characterized by an excess supply of real M1 balances; the excess liquidity isparticularly pronounced for the twin crises episodes, which nearly account for all the above-normalbehavior ahead of currency crises. Without overinterpreting this result, given the shortcomings of money-demand estimation, the picture that emerges is consistent with the deficit-financing as in the Krugman(1979) framework or the excess liquidity may be created to ease conditions for troubled financialinstitutions. In any case, at some point the excess liquidity becomes incompatible with maintaining theexchange rate commitment--and a currency crisis emerges. This would suggest that the high real interestrates prior to banking crises were due to factors other than monetary policy. The next panel shows theevolution of the 12-month change in M2/ reserves of central banks. For both currency and banking crises,this ratio grows well above its norm prior to the crises. The increases are associated with both a vigorousexpansion in M2 (witness the multiplier) and a sharp decline in foreign currency reserves (discussedbelow). As Calvo and Enrique Mendoza (1996) do for Mexico 1994, we find that the M2/reserves ratioover the 76 currency crises indicates an abrupt decrease in the backing-ratio in the months preceding thecrisis. Indeed, the growth rate is 70 percent in excess of the tranquil period average, highlightingvulnerability of the system. This observation is equally descriptive of both single currency and twin crisesepisodes. The growth rate of bank deposits remains close to normal during the 18 months prior to thefinancial crises, but the loss of deposits accelerates as the crises unfold. There may be multiple reasons forthis sudden decline. Past-financial crises periods have often been characterized by massive and persistentcapital-flight. Deposits only start to recover a year and a half after the onset of the financial crises.B. The External Sector

The next four panels of Figures 2, 3, and 4 present indicators associated with the current account.

14

The middle panel of the third row in each figure chronicles the abysmal performance of the growth ofexports in the year-and-a-half preceding the currency and banking crises--exports consistentlyunderperform normal times during this period. By the time a balance of payments crisis is underway,export growth is about 20 percent below (annual rate) the average growth observed in tranquil periods. Once the appreciation is reversed, export performance improves sharply, outdoing the performanceobserved during tranquil periods about nine months after the crisis began. Export performance is

particularly poor during the twin crises episodes. The behavior of import growth is more difficult to justifyon the basis of relative price developments (see below). Import growth remains close to the norm duringtranquil periods up to about nine months before a currency crisis and then declines; for banking crises, wesee the tail end of the import boom and the subsequent slide prior to the crisis. During this pre-crisisperiod, income and relative price effects are moving in opposite directions, and the observed decline inimport growth may well be accounted for by the slowdown in economic activity (see below) during thattime. Import growth remains below that of normal periods throughout the post-crisis period.

The next panel provides evidence on the terms of trade. Crises are preceded, on average, by a

deterioration of the terms of trade, with an annual decline that is about 10 percent deeper than thoseobserved in tranquil times prior to a balance of payments crisis. This persistent adverse performance of theterms of trade erodes purchasing power and may also account for the weakness in imports in the monthspreceding the crisis. This weakness is equally evident in single and twin crises episodes. For bankingcrises, up to about a year prior to the crisis terms of trade shocks appear to have been positive--perhapshelping to explain the earlier boom (see below); as the crises nears we see some evidence of adverse termsof trade shocks. The middle panel in the fourth row shows the evolution of real exchange rates. During theyear before the balance of payments and banking crises (as stressed in Rudiger Dornbusch, et. al,. 1996),the real exchange rate shows evidence of being overvalued, relative to its average level during tranquiltimes. In periods preceding the currency crash, it is appreciating relative to its trend (an overvaluation of

15

about 20 percent relative to tranquil periods). The real exchange rate appreciation does reverse itselfrapidly with the devaluation, suggesting that productivity shocks or preference changes were unlikely toaccount for the initial appreciation. Exchange rate-based inflation stabilization plans have often given riseto large cumulative real exchange rate appreciations, as domestic inflation fails to converge to internationallevels. As noted in Reinhart and Végh (1995), many of those plans ended in a balance of payments crisis. Following the crash, the real exchange rate depreciates substantially (and is about 10 percent higher than intranquil times). Over time, higher domestic inflation erodes in part the improvement in competitiveness.

In the absence of monthly data on capital flows for most of the period and most of the countries in

our sample, we extract information about capital account developments by focusing on the indicatorsshown in the next two panels. As expected, the 12-month percentage change in foreign exchange reservesof the central banks falls substantially in the months prior to both banking and balance of payments crises. The loss of reserves is particularly steep and longer-lived following the crises for the 19 twin crisesepisodes. As early as 12 months prior to the balance of payments crisis, reserve growth is about 20 percentbelow that observed during tranquil periods; although we report 12-month changes, which introducespositive serial correlation in the data, reserves do not decrease continuously. There are modest short-livedreversals in the path followed by reserves, which suggest that the central banks may have had spells inwhich they fought the reserve loss with contractionary monetary policy (note that there are brief spellswhere real interest rates rise prior to the crisis--see third panel) before finally conceding defeat anddevaluing. Following the devaluation (or flotation), foreign exchange reserves of central banks start toincrease again.

Finally, the first panel in the bottom row shows the evolution of the domestic-foreign real interest

rate differential on deposits. Interest differentials do not reflect increasing expectations of a devaluation asthe currency crisis nears. Turning to banking crises, the picture that emerges is quite distinct from itscounterpart in Figure 2; while in balance of payments crises interest rate differentials were not appreciably

16

different from tranquil periods prior to crises, differentials in the case of banking crises remain above thoseobserved in periods of tranquility. One explanation for this difference among the two crises has to do withthe bunching of the banking crises in the post financial liberalization period.C. The Real Sector

The last two panels in the Figures show the evolution of output growth and changes in stock

prices. The deterioration of the terms of trade, the overvaluation of the currency, the weakening exportperformance is reflected in a marked slowing in economic activity and a decline in output prior to bothcrises. For balance of payments crises, the 12-month growth in output bounces in a range of 2 to 6 percentbelow the comparable growth rates during tranquil periods--with a tendency for the recession to deepen asthe crisis nears. Interestingly, and in line with the greater severity of the twin crises, the combination ofcurrency and banking problems appears to take a more devastating toll on the real economy as therecession is far deeper and longer than the recessions associated with currency crashes alone. At growthrates which are eight percent below those observed in tranquil periods, the twin crisis recession is twice assevere. As Kindelberger (1978) observes: “Financial crises are associated with the peaks in business cycles... the financial crisis is a culmination of a period of economic expansion that leads to downturn” While inthe 18 months prior to a balance of payments crisis there is no evidence of a residual economic boom, thatis not the case in the pre-banking crisis period. As Figure 3 shows, up to about 8 months before thebanking crises the economy was recording growth rates above those observed during tranquil periods. Yet,the real exchange rate appreciation that characterizes pre-crisis periods is often cited as a key factor behindthe squeeze in profit margins that eventually leads to increased bankruptcies, a rise in nonperforming loans,a deepening in the economic contraction, and banking sector problems.

The last panel shows the evolution of stock prices. During the 18 months prior to a balance of

payments crisis, the equity market steadily underperforms (relative to tranquil times); at first, not by muchbut as the crisis nears changes in stock prices, that is stock returns (in dollars) are about 40 percent below

17

those observed in non-crisis periods. The weakening in equity prices is, most likely, reflecting both thedeteriorating cyclical position of the economy, reduced foreign demand as capital inflows are reversed, andthe worsening balance sheets of firms, as the overvaluation takes its toll. The crash is particularly severewhen currency and banking crises nearly coincide (Figure 4). Unlike the onset of a banking crisis (seebelow), the equity market was already past it cyclical peak well before the crisis begins. On the eve ofbanking crises, the return on equity prices up to about nine months prior to the crises suggest a boom(relative to tranquil periods) which may (or may not ) be an asset-price bubble. During the boom phase,returns exceed those of non-crises periods by about 40 percent on an annual basis. The beginning of therecession is also reflected in the stock market, which collapses the year before the crisis; this collapse isalso apparent in other asset markets, most notably real estate.13

Finally, although not shown in the figures, the fiscal deficit/GDP ratio is higher in the two years

prior to the currency crisis and one year prior to the banking crisis. While the bigger deficit could stemfrom higher government spending, the weakness in output prior to crises could lead to a shortfall inrevenues.

III. The Anatomy of Crises

In what follows, we offer an alternative approach to examine the evolving nature of the crises,

pinpoint their origin, and gauge their probability conditioned on signals from one or more indicators. Themethodology used, while not previously applied to analyze currency and banking crises, has a long historyin the rich literature that evaluates the ability of macroeconomic and financial time series to predictbusiness cycle turning points.14 The remainder of this section is divided into two parts, the first describesthe statistical methodology used, while the second applies that methodology to the 102 currency andbanking crises that make up our sample.A. Methodology

To examine the causes of crises, gauge the vulnerability of the economy on the eve of crisis, and

18

assess whether the crisis itself could be forecasted by anomalous economic developments, we need to makefour sets of judgments: First, we must have a well-defined notion of what is classified as a crisis. Second,we must agree on a list of variables that are potential leading indicators. Third, we need to decide upon acriteria that allows us to classify the behavior of an indicator as either a signal of a crisis or normal (nosignal). Last, if an indicator is giving a signal, we have to determine if a crisis happens within a reasonableperiod of time or if the signal was a false alarm. Hence, we also need to define what is considered to be areasonable period of time. Section I dealt with the definition and dating of banking and currency crises,while the previous Section and the Data Appendix discuss the indicators. In this subsection, we describe theapproach used to define what is a signal and what is a reasonable period of time.The interval between signals and crisis: Defining a reasonable period of time

In what follows, the maximum interval of time between the signal and the crisis, was decided upon

a-priori as 24 months in the case of balance of payments crises.15 Hence, any signal given within the24-month period before the beginning of the crisis is labeled a good signal; any other signal outside that 24-month window is labeled a false alarm or noise. For banking crises, any signal given within the 12-monthperiod before the beginning of the crisis or within 12 months following the beginning of the crisis is labeleda good signal. The two different signaling windows for currency and banking crises have to do with thedifferent timing of the peaks of both crises, as previously discussed. In addition, the events that mark thebeginning of a banking crisis are often not seen as systemic at the time and are not treated by policymakersas harbingers of a crisis. Since symptoms of a crisis are sometimes evident well before the crises erupt, thenarrower windows (say, 12-months) were thought to penalize indicators (such as M2/Reserves) that tend togive an early warning.

The threshold: Defining a signal

In Section II we noted that the crises were preceded by marked declines in equity returns. Yet,

surely, not every decline in equity returns presages a crisis. Hence, we need to select an appropriate

19

threshold or cutoff that separates when a decline in equity returns is considered a signal of a crisis andwhen it is not. As is the case of selecting the size of the rejection region in hypothesis testing, choosing theoptimal threshold involves a tradeoff. Suppose that our null hypothesis is that we are in a tranquil state ofnature and (for a particular country) we were weighing whether to arbitrarily set the threshold for annualequity returns at minus 15 percent or at minus 40 percent. Suppose further, that for this country 10percent of the observations posted annual equity returns below 15 percent, but only 3 percent of theobservations showed equity returns below 40 percent. Our aim is to use the readings for this indicator totest the null hypothesis that we are in a tranquil state of nature. If we choose the minus 15 percentthreshold, the size of \" (the rejection region) is 10 percent--this is the probability of rejecting the nullhypothesis when it is true (Type I error). In this case, the threshold may be too lax--it is likely to catch allthe crises but it also likely generate a lot of false alarms. Instead, we could adopt the minus 40 percentthreshold, which cuts the size of \" to 3 percent; this reduces the probability of Type I error at the expenseof increasing the probability of Type II error (not rejecting the null hypothesis when it is false). With thistight threshold we may miss all but the most severe of the crises--the price of reducing the number of falsealarms is accurately calling a lower proportion of crises.

We select the threshold value on an indicator-by-indicator basis, by performing a fine grid search

over a broad range of critical regions up to a maximum of 30 percent. For each threshold value in our gridsearch we compute the noise-to-signal ratio.16 We then select the threshold value that minimizes thenoise-to-signal ratio. As to the location of the rejection region, whether it is the upper or lower tail of thefrequency distribution for each indicator, we rely on the theory as a guide. The threshold values for the 16indicators, as well as the location of the rejection region and its theoretical justification are given in Table5. For example, for currency and banking crises large output declines signal a crisis, so a < sign in Table 5denotes that the rejection region is located at the bottom tail of the distribution.

This criterion does have drawbacks which are worth mentioning. First, if an indicator gives an

20

early signal and policymakers heed the signal and preempt a crisis, that signal is labeled as false and theindicator is penalized with an unduly high noise-to-signal ratio. In addition, a signal within the window istreated the same irrespective of whether it was given 12 months before the crisis erupts or only the monthbefore. Naturally, from the vantage point of the policymaker the earlier signal is the more valuable one.B. The Anatomy of Crises

The methodology just described was applied to the 16 indicators and 102 crises, in the sample and

the four recent Asian crises out-of-sample.

Appendix Tables 1 and 2 show the results on a crisis-by-crisis and indicator-by-indicator basis.

An NA denotes some or all the observations were missing during the pre-crisis 24-month window; a 1denotes at least one signal was given during the 24-month window, and a zero indicates no signals wereissued. Hence, for example, column (12) in Appendix Table 1 scores the performance of foreign exchangereserves; there are four NA entries, hence we have full data for this indicator for 72 balance of paymentscrises. In 75 percent of the crises (row (1) Summary Statistics, bottom of the table) there were one ormore signals during the 24 months prior to the crisis. The last column (17) lists what proportion (inpercent) of the indicators were sending signals. Data availability permitting, the tables also show theevolution of the indicators out-of-sample for the Asian crises of 1997.

About the origins of crises

Table 6 summarizes the results in Appendix Tables 1 and 2. The indicators are shown individually

and are also grouped into sectors along the lines described in the previous section: financial liberalization,other financial, current account, capital account, real-side, and fiscal. For balance of payments crises, wealso examine subsamples before financial liberalization, which encompasses the 1970s and after financialliberalization as well as those currency crises which occured alongside a banking crises. The latter appearunder the column labelled Twin. As nearly all banking crises fall in the post-liberalization period, nosubsamples for these are reported. Table 6 presents the percentage of crises accurately called by each

21

indicator. As to the various groups we also report the simple arithmetic average of the proportion of crisesaccurately called by all the indicators in that subgroup. Capital account indicators accurately called thehighest proportion of balance of payments crises (about 81 percent). Financial liberalization indicatorswere next in line, accurately signaling 74 percent of the currency crises before they occurred; for the twincrises their performance is even better. Among the capital account and financial indicators that fared theworst are bank deposits, the lending-deposit ratio and excess M1 balances. Current account indicatorsfollowed next (68 percent accurately called) but this is largely owing to the weak performance of imports inaccurately calling crises; exports, the terms of trade, and the real exchange rate do much better. The fiscalvariable fared the worst, accurately calling only slightly over a quarter of the currency crises.

One key difference between banking and currency crises, highlighted in Table 6, is the role of the

real sector, which appears to be considerably more important for banking crises--giving early signals in 85percent of the crises.17 Indeed, output and stock prices signaled in 89 and 81 percent, respectively, of thebanking crises for which data for these indicators was available. As much of the literature on bankingcrises stresses, particularly asymmetric information models (see Charles W. Calomiris and Gary Gorton(1991)), the evidence presented here suggests that the bursting of asset price bubbles and increasedbankruptcies associated with an economic downturn appear to be closely linked to domestic financialproblems.

Yet another feature that is revealed in Table 6, is that the proportion of crises accurately called

rises for 13 out of the 16 indicators when single currency crises are compared to their twin counterparts.The improved performance of most of the indicators is not entirely surprising, in light of the greaterseverity of the twin crises episodes.Fragility on the eve of crises

Table 7 presents strong evidence that, for both banking and currency crises, multiple economic

problems were simultaneously building. We construct a measure of the fragility of the economy in the 24

22

months preceding the crisis by tallying on a crisis-by-crisis basis what proportion of the indicators weresignaling during that period.18 Hence, if 14 of the 16 indicators are sending a signal prior to the crisis, thiscrisis would be counted in the first row of Table 7, labeled 80% to 100% . It appears that crises are notsimply a story of an overvalued exchange rate or too rapid a monetary expansion. In about 30 percent ofthe currency crises, 80 percent or more of the indicators were sending signals. The economies appear to beparticularly frail on the eve of twin crises, with a higher proportion of the indicators signaling. Indeed, inabout 80 percent of the twin crises, at least 60 percent of the indicators were sending a signal. There werebasically no banking with less than 20 percent of the indicators signaling. For further evidence of thediversity of the economic problems on the eve of crises on a crisis-by-crisis basis see Appendix Tables 1and 2. The finding that when the balance of payments crises occur jointly with a banking crisis (under theheading Twin, Table 7) economies appear have more widespread problems perhaps is not entirelysurprising , given the earlier results which suggest the twin crises tend to be more severe.

These results would appear to suggest that the overwhelming majority of crises, external or

domestic, have a multitude of weak economic fundamentals at their core. While speculative attacks dooccur as market sentiment shifts and, possibly, herding behavior takes over, such self-fulfilling crisesappear to be quite rare. Indeed, in the context of the Exchange Rate Mechanism crises this issue has beenthe subject of much debate.19 Not only are the signals many, but their sources are multiple, as shown in Table 7--with the financial sector external (capital account) and domestic playing a key role.VI. Final Remarks

We have examined the empirical regularities and the sources and scope of problems in the onset of

76 currency crises and 26 banking crises. We find that banking and currency crises are closely linked in theaftermath of financial liberalization, with banking crises, in general, beginning before the currency collapse. We also find evidence of vicious cycles, in which the currency collapse further undermines an alreadyailing banking sector. When currency and banking crises occur jointly, they are far more severe than when

23

they occur in isolation. In both types of crises, a financial shock, possibly financial liberalization orincreased access to international capital markets, appear to activate a boom-bust cycle by providing easyaccess to financing. Finally, in both crises we find a multitude of weak and deteriorating economicfundamentals suggesting that it would be difficult to characterize them as self-fulfilling crises.

During much of 1997 and 1998, the financial press has frequently stressed that the crises in Asia

are a new breed, as they supposedly occured against a backdrop of immaculate fiscal and economicfundamentals. Yet our analysis of earlier episodes reveals that many of the features and antecedents of thecrises in Asia were common to a substantial number of crisis episodes in Latin America, Europe, andelsewhere. Consider an economy that had successfully stabilized inflation, enjoyed an economic boom, andwas running fiscal surpluses. However, this economy had liberalized its capital account and its domesticfinancial sector amidst an environment of weak regulation and poor banking supervision. Banking sectorproblems emerged and intensified, eventually undermining the ability of the central bank to maintain itsexchange rate committment. While this profile fits Asia rather well, this was Diaz Alejandro’s descriptionof the antecedents to the fierce Chilean crisis of 1982. At the roots of the meltdown of the Thai baht,Korean won and Indonesian rupiah lay systemic banking problems. Thus, it would appear that we can onlyconsider these crises as a new breed if we ignore the numerous lessons history offers. Thus, among thelessons that emerge from this analysis is the obvious case for strong banking regulation and supervision toallow countries to sail smoothly through the perilous waters of financial liberalization. Yet, the Asianepisodes of 1997-98, like many of their earlier Latin American counterparts, also remind us that capitalinflows can on occasion be too much of a good thing.

The results presented in this paper are a first step in evaluating the complex linkages between

currency and domestic financial crises. Analyzing how the authorities deal with the banking problems andhow the problems affect exchange rate expectations will help determine whether a banking crises will leadto a balance of payments crisis. We have only considered macroeconomic data in our list of indicators, but

24

data of the health of bank balance sheets would be a logical complement to the macro data. Future analysiscould provide a more detailed evaluation of the univariate and multivariate signaling properties of variousmacroeconomic time series and composite indices along the lines of Stock and Watson (1989) and Dieboldand Rudebusch (1989). Indeed, that would appear to be a logical first step in the design of an early warningsystem designed to help detect when a crisis is coming.

While this paper has focused on the similarities and common patterns across crises, it would also

be useful to investigate whether there is evidence of distinct regional patterns. Why is it that in somecountries currency crises and banking crises are not associated with deep and protracted recessions, whilein others, notably in Latin America, the aftermath is so severe? Lastly, events (such as a balance ofpayments crises in a neighboring country) may also help assess whether a crisis is brewing in the homefront; hence, the role of contagion effects may warrant further scrutiny.

25

References

Agénor, Pierre.; Bhandari, J.S. and Flood, Robert P. “Speculative Attacks and Models of

Balance-of-Payments Crises,” IMF Staff Papers, 1992, 39, pp. 357-94.American Banker, various issues.Blanco, Herminio and Garber, Peter M. “Recurrent Devaluations and Speculative Attacks on the Mexican

Peso,” Journal of Political Economy, 1986, 94, pp. 148-66.Calomiris, Charles W. and Gorton, Gary. “The Origins of Banking Panics: Models, Facts, and Bank

Regulation,” in R.Glenn Hubbard, ed., Financial markets and financial crises. Chicago, IL:University of Chicago Press, 1991, pp. 109-173.

Calvo, Guillermo A. “Varieties of Capital-Market Crises,” Mimeo, University of Maryland, College Park,

1995.

; Leiderman, Leonardo and Reinhart, Carmen M. “Capital Inflows and Real Exchange RateAppreciation: the Role of External Factors,” IMF Staff Papers, 1993, 40, pp. 108-51. and Mendoza, Enrique. “Petty Crime and Cruel Punishment: Lessons from the Mexican Debacle,”American Economic Review, Papers and Proceedings, 1996, 86, pp 170-75.Calvo, Sara and Reinhart, Carmen M. “Capital Flows to Latin America: is there Evidence of Contagion

Effects?”, in Guillermo A. Calvo, Morris Goldstein, and Eduard Hochreiter, eds., Private capitalflows to emerging markets. Washington, DC: Institute for International Economics, 1996, 151-71.Caprio, Gerald, Jr. and Klingebiel, Daniela. “Bank Insolvency: Bad Luck, Bad Policy, or Bad Banking?” in

Annual World Bank conference on development economics, Washington, DC: The World Bank,1996, pp. 79-104.

Diaz-Alejandro, Carlos. “Good-Bye Financial Repression, Hello Financial Crash,” Journal of DevelopmentEconomics, 1985, 19.Diebold, Francis and Rudebusch, Glen. “Scoring the Leading Indicators”, Journal of Business, 1989, 62,26

pp. 369-91.

Dornbusch, Rudiger; Goldfajn, Ilan and Valdés, Rodrigo. “Currency Crises and Collapses,” BrookingsPapers on Economic Activity, 1995, 2, pp. 219-95.Edwards, Sebastian. “Real exchange rates, devaluation, and adjustment: exchange rate policy in developing

countries,” Cambridge, MA: MIT Press, 1989.

Eichengreen, Barry, Rose, Andrew K. and Wyplosz, Charles, “Exchange Market Mayhem: the Antecedents

and Aftermath of Speculative Attacks,” Economic Policy, 1996a, 21 pp. 249-312. , “Contagious Currency Crises,” CEPR Discussion Paper 1453, August 1996b.Emerging Market Indicators, Washington, DC: International Finance Corporation, various issues.Frankel, Jeffrey and Rose, Andrew K. “Exchange Rate Crises in Emerging Markets,” Journal ofInternational Economics, 1996, 41, pp. 351-68.Galbis, Vicente. “High Real Interest Rates Under Financial Liberalization: Is There a Problem?,”

International Monetary Fund Working Paper WP/93/7, 1993.

Goldfajn, Ilan and Valdés, Rodrigo. “Balance-of-Payments Crises and Capital Flows: the Role of

Liquidity,” Mimeo, Massachussets Institute of Technology, 1995.

Gorton, Gary. “Banking Panics and Business Cycles,” Oxford Economic Papers, 1988, 40, pp. 751-81.International Financial Statistics, Washington, DC: International Monetary Fund, various issues.Kaminsky, Graciela L. and Leiderman, Leonardo. “High Real Interest Rates in the Aftermath of

Disinflation: Is it a Lack of Credibility?” Journal of Development Economics 1998, 55, 191-214.Kindelberger, Charles. Manias, panics, and crashes, New York, NY: Basic Books Inc., 1978.Krugman, Paul. “A Model of Balance-of-Payments Crises,” Journal of Money, Credit, and Banking, 1979,11, pp. 311-25.McKinnon, Ronald.I. and Pill, Huw. “Credible Liberalizations and International Capital Flows: the

Overborrowing Syndrome,” in Takatoshi Ito and Anne Krueger, eds., Financial deregulation and27

integration in East Asia, Chicago, IL: Chicago University Press, pp. 7-42.Miller, Victoria. “Central Bank Reactions to Banking Crises in Fixed Exchange Rate Regimes,” Mimeo,

Université de Québec à Montréal, 1995.

Mishkin, Frederic S. “Understanding Financial Crises: A Developing Country Perspective,” in AnnualWorld Bank conference on development economics, Washington DC:World Bank, 1996, pp. 29-62.

Mishra, Deepak. “Political Determinants of Currency Crises: Theory and Evidence,” Mimeo, University of

Maryland, College Park, 1997.

New York Times, various issues.Obstfeld, Maurice. “The Logic of Currency Crises,” Cahiers Economiques et Monetaires, Bank of France,1994, 43, pp. 189-213. . “Models of Currency Crises with Self-fulfilling Features,” NBER Working Paper 5285, 1995.Reinhart, Carmen M. and Reinhart, Vincent, R. “Forecasting Turning Points in Canada,” Mimeo,

International Monetary Fund, 1996.

, and Végh, Carlos A. “Do Exchange Rate-Based Inflation Stabilizations Sow the Seeds of their OwnDestruction?.” Mimeo, International Monetary Fund, 1996.

Stock, James H., and Watson, Mark W. “New Indices of Coincident and Leading Economic Indicators,”

NBER macroeconomics annual, 1989, pp. 351-93.Stoker, James. “Intermediation and the Business Cycle under a Specie Standard: the Role of the Gold

Standard in English Financial Crises, 1790-1850.” Mimeo, University of Chicago, Chicago, 1994.

Sundararajan, V. and Baliño, Tomas. “Issues in Recent Banking Crises,” in V. Sundararajan and Tomas

Baliño, eds., Banking crises: cases and issues. Washington, DC: International Monetary Fund,1991, pp.1-57.

Velasco, Andres. “Financial and Balance-of-Payments Crises,” Journal of Development Economics, 1987,28

27, pp.263-83.Wall Street Journal, various issues.World Development Indicators, Washington DC: The World Bank, various issues.29

Footnotes

1. See Maurice Obstfeld (1994 and 1995) and Calvo (1995).

2. See Carmen M. Reinhart and Carlos A.Végh (1996) for a review of this literature and the empiricalregularities.

3. The construction of the index is described in the data Appendix. The dates of the crises appear inAppendix Table 1, the level of the index and key events around the crises dates are reported in workingpaper version of this paper, Graciela L. Kaminsky and Reinhart (1996).

4. Bank stocks could be an indicator, but in many of the developing countries an important share of thebanks are not traded publicly.

5. If the peak month for the banking crisis is not known, we list the midpoint of that year as the date.6. The real exchange rate is used, as high inflation countries will typically have larger nominaldevaluations.

7. M2 to reserves captures to what extent the liabilities of the banking system are backed by internationalreserves. In the event of a currency crisis, individuals may rush to convert their domestic currency depositsinto foreign currency, so that this ratio captures the ability of the central bank to meet those demands(Calvo and Mendoza ,1996).

8. An increase in the real exchange rate index denotes a depreciation.

9. Detailed definitions of all the variables and their sources are provided in the data Appendix.10. See Appendix Tables 1 and 2 for a detailed indication of any missing data around crisis dates.11. See Galbis (1993).

12. See Sundararajan and Baliño (1991).

13. For example, in the boom period leading up to the 1981 Argentine banking crisis, stock returns (inU.S. dollars) were as high as 813 percent during the 12 months ending May 1979; by May 1981, the12-month capital loss was 60 percent. The crash in asset values is cited in most case studies as an

30

important factor contributing to the problems of the banks. Also, due to either mismanagement or outrightfraud, in many of the crises in our sample a substantial portion of banks and finance companies wereconsiderably overexposed to real estate.

14. See, for instance, James H. Stock and Mark W. Watson (1989), Francis Diebold and Glen Rudebusch(1989), and Reinhart and Vincent R. Reinhart (1996).

15. An 18- and 12-month window were also used; the results are available from the authors.

16. The definition of noise-to-signal ratio used throughout is best illustrated by considering the followingtwo-by-two matrix:

Crisis occcurs in the following

24 months

Indicator issues a signalIndicators does not issue a signal

AC

No crisis occurs in the following

24 months

BD

If a variable signals and a crisis occurs in the following 24 months (counted in cell A) the signal isconsidered accurate. If a variable signals and no crisis occurs in that time frame (counted in cell B), thesignal is said to be a false alarm or noise. Hence, a perfect indicator would only have entries in cells in Aand D. More generally, the noise-to-signal ratio for any indicator is given by the number of entries in[B/(B+D)]/[A/(A+C)]. Hence, it is the ratio of false signals to all possible bad signals divided by the ratioof good signals to all possible good signals. An extremely noisy indicator would have few entries in A andD, many in B.

17. For a discussion of the evolving nature of crises see the working paper, Kaminsky and Reinhart(1996).

18. These are reported for each crisis in column 17 of the Appendix tables.19. See Eichengreen, Rose, and Wyplosz (1995) and Krugman (1996).

31

Table 1

Frequency of Crises Over Time

Number of Crises

1970-1995

Type of Crisis

Total

Averageper year

1970-1979Total

Averageper year

1980-1995Total

Averageper year

Balance-of-paymentsTwinSingleBanking

76 19 57 26

2.92 0.73 2.19 1.00

26 1 25 3

2.60 0.10 2.50 0.30

50 18 32 23

3.13 1.13 2.00 1.44

Note: Episodes in which the begining of a banking crisis is followed by abalance-of-payments crisis within 48 months are classified as twincrises.

Table 2

The Timing of the Twin Crises and Financial Liberalization

Country

Financial Liberalization

1977

Banking Crisis Begining Peak

Closest BOP Crisis

February 1981September 1986February 1990September 1985November 1986October 1991August 1982March 1983August 1983November 1991September 1986October 1983July 1975December 1982December 1994May 1986October 1987October 1983July 1977November 1992November 1978November 1984March 1994December 1971October 1982May 1994

ArgentinaMarch 1980May 1985December 1994

July 1982June 1989March 1995June 1988November 1985March 1996March 1983June 1985June 1990June 1992November 1992June 1984August 1986June 1984March 1996October 1991April 1983June 1985January 1983September 1992March 1979June 1985March 1991December 1971June 1985August 1994

BoliviaBrazil

19851975

October 1987November 1985December 1994

ChileColombia DenmarkFinlandIndonesiaIsraelMalaysiaMexico

19741980Early 1980's198219831985197819741991

September 1981July 1982March 1987September 1991November 1992October 1983July 1985September 1982October 1992November 1988March 1983January 1981November 1978November 1991March 1979October 1983

NorwayPeru PhilippinesSpain SwedenThailand

198019911980197419801989

TurkeyUruguay

19801976-79

January 1991March 1971March 1981

Venezuela1981,1989October 1993

Memorandum item: Out-of-sampleIndonesiaMalaysiaPhilippinesThailand

November 1992September 1997July 1997May 1996

OngoingOngoingOngoingOngoing

August 1997August 1997July 1997July 1997

Note: Episodes in which the begining of a banking crisis is followed by a balance of payment crisis within 48 months are classified as twincrises.

Sources: American Banker, various issues; Caprio, Gerald Jr. and Daniela Klingebiel (1996); New York Times, various issues;Sundararajan, Vasundevan and Tomas Balino (1991); and Wall Street Journal, various issues.

Table 3

Probabilities of Crises

Probabilities of Balance-of-Payments Crises Type Unconditional

Conditional on the Begining of aBanking Crisis

Conditional on the Peak of aBanking Crisis

Value (in percent) 29 46 22

Probabilities of Banking Crises TypeUnconditional

Begining of a Banking CrisisConditional on a Balance-of-Payments Crisis

Begining of a Banking CrisisConditional on FinancialLiberalization

Peak of a Banking CrisisConditional on a Balance-of-Payments Crisis

Value (in percent) 10

8

14

16

Notes: The balance-of-payment crisis windows are defined as the 24 months preceding thecrisis. The banking crisis windows are defined as the 12 months before and the 12 monthsafter the begining (or peak) of the crises. The unconditional probabilities of balance-of-payment and banking crises are calculated as the total number of months in the respectivecrisis windows divided by the total number of months in the sample. The balance-of-payment probabilities conditional on a banking crisis (begining or peak) are calculated asthe number of months in the balance-of-payment crisis windows that occur within 24 monthsof the banking crises (begining or peak) divided by the total number of months in thebanking crisis windows. The probabilities of banking crises conditional on balance-of-payment crises are calculated as the number of months in the banking crisis windows thatoccur within 24 months of a balance-of-payment crisis divided by the total number of mothsin the balance-of-payment crisis windows. The probabilitity of a banking crisis conditional onfinancial liberalization is calculated as the total number of months in the banking crisiswindows that occur during times of financial liberalization divided by the total number ofmonths during which the bnaking sector was in a regime of financial liberalization. Allprobabilities were estimated using the data for the twenty countries in the 1970-mid 1995

period.

Table 4

The Severity of the Crises

Banking Crises

Severity MeasureCost of Bailout(percent of GDP)Loss of Reserves(percent)

Real Depreciation(percent)Composite Index

Twin 13.3 N.A. N.A. N.A.

Single 5.1* N.A. N.A. N.A.

Balance-of-Payment Crises Twin N.A. 25.4 25.7 25.6

Single N.A. 8.3* 26.6 17.5

Notes: Loss of reserves is the percentage change in the level of reserves in thesix months preceding the crises. Real depreciation is the precentage change inthe real exchange rate (with respect to the dollar for the countries that peg tothe dollar and with respect to the mark for the countries that peg to mark) in thesix months following the cirses. The composite index is the unweighted averageof the loss of reserves and real depreciation. Episodes in which the begining ofa banking crisis is followed by a balance-of-payments crisis within 48 monthsare classified as twin crises.

* Denotes that the measure of severity of sigle crises episodes is statisticallydifferent from the twin crises severity at the 5% level. An N.A. denotes notapplicable.

Table 5

Threshold Values for Signaling Crises

Indicators

Threshold Values and the Location of the Critical RegionBalance-of-PaymentCrises

Financial Liberalization M2 Multiplier Domestic Credit/GDP Real Interest Rate

>0.86>0.90>0.88

>0.90>0.95>0.80

Both banking and currency crises have been linked to rapid growth(boom-bust) in credit and the monetary aggregates (see McKinnonand Pill, 1996)

For banking crises, the choice is unambiguous since financialderegulation is associated with highe interest rates (which couldreflect increased risk taking (see Galkis, 1993)). A liquidity crunch(say to defend a peg) will also hurt the banks. For balance-of-payments crises it is less clear cut; higher real interest rates couldreflect higher risk premia and fears of devaluation. Yet, using thelower interest rates for signals could be justified for balance-of-payments crises on the basis of loose monetary policy. An

increase in the lending/deposit ratio can capture a decline in loanquality

BankingCrises

Comments

Lending-deposit rate ratio>0.80>0.87

Other Financial

Excess M1 balances M2/reserves Bank DepositsExternal Sector Current Account Exports Terms-of-Trade Real Exchange Rate Imports

<0.10<0.16<0.10>0.90

<0.10<0.19<0.10>0.80

Real exchange rate overvaluations and a weak external sector area part of a currency crisis. It adds vulnerability of the banking

sector, since a loss of competitiveness and external markets couldlead to a recession, business filures, and a decline in the quality ofloans. Thus large negative shocks to exports, the terms of trade,and the real exchange rate are associated with signals (seeDornbush, Goldfajn and Valdes 1995)

Theory is ambiguous as to where we should locate the rejectionregion. Rapid import growth couold be the sign of a buoyanteconomy (this would argue for a negative shock to imports); itcould also be the sign of overvaluation, hence a positive shockcould be a signal. Both possibilities are explored.

>0.94>0.87<0.10

>0.91>0.90<0.16

This is a \"loose\" monetary policy story (see Krugman, 1979)For the motivation on M2/reserves se Calvo and Mendoza,1996Capital flight and a run against the domestic banks may precedeboth currency and banking crises (see Goldfajn and Valdes, 1995)

Capital Account Reserves

Real Interest rate differentialReal Sector Output Stock Prices<0.11<0.11

<0.14<0.10

Recessions and the burst of asset price bubbles precede financialcirses (see Calomiris and Gorton, 1991)

<0.15>0.89

<0.28>0.81

See discussion under bank deposits and real interest rates.

Fiscal Sector Deficit/GDP

>0.86

>0.86

Loose fiscal policy financed by credit from the central bank (seeKrugman, 1979)

Note: The definitions and sources of the indicators are described in the Data Appendix.

Table 6

The Onset of Financial Crises: Early Signals

Percent of Crises Accurately Called Balance-of-Payments Crises

IndicatorsFinancial SectorFinancial

LiberalizationM2 MultiplierDomestic Credit/GDPReal Interest RateLending-Deposit RateRatioOther

Excess M1 BalancesM2/ReservesBank DepositsExternal SectorCurrent AccountExportsTerms of TradeReal Exchange RateImportsCapital AccountReservesReal Interest RateDifferentialReal SectorOutputStock PricesFiscal Sector

Total6774766189715737815172688575595281758669746428

Single6772755986705843795271678372575780748669736527

Twin6778786794735322894774708983673983798870776329

Before FinancialLiberalization6764745678505752744472707873587374707861685321

After FinancialLiberalization6877776591735626865672668977604083788972766831

657173501005757327567827588965860969210085898144BankingCrises

Notes: Episodes in which the begining of a banking crisis is followed by a balance-of-payments crisis within 48months are classified as twin crises. An indicator is said to have accurately called a crisis if it issues at least onesignal in the crisis window on the basis of the criterion shown in Table 5. For each indicator, each cell in thetable represents the number of times that indicator correctly calls a crisis as a percentage of the total number ofcrises. For the different sectors, each cell represents the simple average of the percentage of crises accuratelycalled by all the individual variables in that group.

Table 7

Economic Fragility on the Eve of Crises

Number ofIndicatorsSignaling aCrisis

(in percent)80-10060-7940-5920-39Less than 20

Number of Crises (in percent) Balance-of-Payments CrisesTotal

Single

Twin

BeforeFinancial

Liberalization

After

Financial

Liberalization

BankingCrises

26.745.320.06.71.3

28.6 41.1 21.4 8.9 0.0

21.1 57.9 15.8 0.0 5.3

40.0 23.3 20.0 13.3 3.3

17.8 60.0 20.0 2.2 0.0

30.8 53.8 11.5 3.9 0.0

Notes: This table captures the state of distress of the economy in different crisis episodes.

Each cell represents the proportion of crises with a given proportion of signals. For

example, 21.1% of the twin balance-of-payment crises had 80-100 percent of indicatorssignaling a crisis. Episodes in which the begining of a banking crisis is followed by abalance-of-payment crisis within 48 months are classified as twin crises.

Data Appendix

Index of Currency Market Turbulence

The index, I, is a weighted average of the rate of change of the exchange rate, )e/e, and of reserves, )R/R, withweights such that the two components of the index have equal sample volatilities I = ()e/e) - (Fe/FR)*( )R/R)

where Fe is the standard deviation of the rate of change of the exchange rate and FR is the standard deviation ofthe rate of change of reserves. Since changes in the exchange rate enter with a positive weight and changes inreserves have a negative weight attached, readings of this index that were three standard deviations or moreabove the mean were cataloged as crises. For countries in the sample that had hyperinflation, the contructionof the index was modified. While a 100 percent devaluation may be traumatic for a country with low-to-moderate inflation, a devaluation of that magnitude is commonplace during hyperiflations. A single index for thecountries that had hyperinflation episodes would miss sizable devaluations and reserve losses in the moderateinflation periods, since the historic mean is distorted by the high-inflation episode. To avoid this, we divided thesample according to whether inflation in the previous six months was higher than 150% and then constructedan index for each subsample. Our cataloging of crises for the countires coincides fairly highly with our

chronology of currency market disruptions. Eichengreen, Rose and Wyplosz (1996a) also include interest ratesin this index, however, our data on market determined interest rates on developing countries does not span theentire sample. The Indicators

Sources: International Financial Statistics (IFS), International Monetary Fund (IMF); Emerging Market Indicators,International Finance Corporation (IFC); World Development Indicators, The World Bank (WB). When data wasmissing from these sources, central bank bulletins and other country-specific sources were used assupplements. Unless otherwise noted, we used 12-month percent changes.

1. M2 Multiplier: The ratio of M2 (IFS lines 34 plus 35) to base money (IFS line 14)

2. Domestic Credit/GDP: IFS line 52 divided by IFS line 64 to obtain domstic credit in real terms, which was thendivided by IFS line 99b.p (interpolated) to obtain the domestic credit/GDP ratio. Monthly real GDP wasinterpolated from annual data.

3. Real Interest Rate: Deposit rate (IFS line 60) deflated using consumer prices (IFS line 64). Monthly ratesexpressed in percentage points. In levels.

4. Lending-Deposit rate ratio: IFS line 60p divided by IFS line 60 was used in lieu of differential to ameliorate thedistortions caused by the large percentage point spreads ovserved during high inflation. In levels.

5. \"Excess\" M1 balances: M1 (IFS line 34) deflated by consumer prices (IFS line 64) less an estimated demandfor money. The demand for real blances is determined by real GDP (interpolated IFS line 99b.p), domesticconsumer price inflation, and a time ternd. Domestic inlation was used in lieu of nominal interest rates, as

market determined interest rates were not available during the entire sample for a number of countries; the timetrend (whcih can enter log-linearly, linearly, or exponentially) is motivated by its role as a proxy for financialinnovation and/or currency substitution. In levels.

6. M2/Reserves: IFS lines 34 plus 35 converted into dollars (using IFS line ae) divided by IFS line 1L.d.7. Bank deposits: IFS line 24 plus 25 deflated by consumer prices (IFS line 64)8. Exports: IFS line 709. Imports: IFS line 71

10. Terms of Trade: The unit value of exports (IFS line 74) over the unit value of imports (IFS line 75). For thosedeveloping countries where import unit values (or import price indices) were not available, an index of prices ofmanufactured exports from industrial countries to developing countries was used.

11. The real exchange rate: The real exchange rate index is derived from a nominal exchange rate index, adjustedfor relative consumer prices (IFS line 64). The measure is defined as the relative price of foreign goods (indomestic currency) to the price of domestic goods. The nominal exchange rate index is a weighted average ofthe exchange rates of the nineteen OECD countries with weights equal to the country trade shares with theOECD countries. Since not all real appreciations reflect disequilibrium phenomena, we focus on deviations ofthe ral exchange rate from trend. The trend was specified as, altenaltivly, log, linear, and exponential; the bestfit among these was selected on a country-by-country basis. In levels.12. Reserves: IFS line 1L.d.

13. Real interest rate differential: Interest rates in the domestic economy are compared with interest rates in theUnited States (Germany) if the domestic central bank pegs the currency to the dollar (Deutsche mark). Theinterest rate differential is contructed as the difference between real rates for the domestic and foreigncountries. Real rates are deposit rates (IFS line 60) deflated using consumer prices (IFS line 64).

14. Output: For most countries, the measure of output used is disutrial production (IFS line 66). However, for

some countries, (the commodity exporters) an index of output of primary commodities is used (IFS lines 66aa),if industrial production is not available.

15. Stock returns: IFC global indices are used for all emerging markets; for industrial countries the quotes fromthe main bords are used. All stock prices are in US dollars.

16. GDP: Consolidated public sector deficit as a share of GDP. World Development Indicators.