Middle Eastern Finance and Economics ISSN: 1450-2889 Issue 5 (2009) EuroJournals Publishing, Inc. 2009 http://www.eurojournals.com/MEFE.

htm

The Determinants of Currency Crises: A Panel Data Analysis1

Serpil Kahraman Department of Economics, Yasar University, Selcuk Yasar Campus, Izmir, Turkey E-mail: [email protected] Tel: +90 232 4115216; Fax: +90 232 4115020 Ela olpan Nart Department of Economics, Yasar University, Selcuk Yasar Campus, Izmir, Turkey E-mail: [email protected] Tel: +90 232 4115217; Fax: +90 232 4115020 Gl Huyugzel Kila Department of Economics, Ege University, Ege University Campus, Izmir, Turkey E-mail: [email protected] Tel: +90 232 3732960; Fax: +90 232 3734194 Abstract This study present survey common features of currency crises in 15 emerging countries by using unbalanced panel annual data set over the period 1987-2007. We estimate ordinary least squares (OLS), least square dummy variable or fixed effect model (LSDV), random effect model (REM) and LSDV&TIME model and the models used to assess the relationship between currency crise index and seven crises indicators. The estimation results show that the leading indicators of currency crises are M2 growth, domestic credit provided by banking sector and current account balance / reserves. Keywords: Currency crises, currency crises models, leading indicators, currency crisis index, panel data analysis, emerging countries. JEL Classification Codes: F31, F33, F47.

1. Introduction
Financial crises have become world wide phenomena after the year 1980s. Through history many financial crises have exhibited a characteristic of a twin crises, banking crises with currency crises. Each currency crisis has differed in type from the preceding one but following sequences sharp depreciation of domestic currency followed by an abrupt recession, a loss of international reserves and a capital outflow always present. In the theoretical literature currency crises involve fixed Exchange rate regime or allowed currencies float within certain margins. Exchange rate crises are observed in economies which adapt fixed exchange rate regime as a result of abrupt speculative attacks of economic agents try to sell domestic currency for foreign currency until they expect that the government will not be able to defend fixed exchange rate. In response, monetary authorities are forced
1

This study is an expanded version of a paper presented at International Conference on Business, Management and Economics, eme 2009.

108

Middle Eastern Finance and Economics - Issue 5 (2009)

to defend the currency. If the authorities are no longer able to maintain the exchange rate, they decide to devalue the currency. Clearly, many reasons for expecting crises and due to a common cause, for instance policies may have similar effects on emerging countries. The paper is organized as follows. After the introduction, section 2 reviews the theoretical literature on financial crises, and section three also theoretical currency crisis models while section 4 discusses the empirical literature on the determinants on currency crises. The data, methodology and estimation results of the models analyze in the fifth section, and the last part will conclude and bring some remarks.

2. Theoretical Literature
Economists focus on the timing and causes of the crisis and try to explain different experiences of countries by using models. Theoretical studies have identified mainly two theoretical models the traditional or first generation models (Krugman, 1979) and second-generation models (Obstfeld, 1986, 1994). These models can be dubbed as Latin type crises models based on the presence of inconsistent policies generates a speculative attack. Recent literature also concentrates on the predictability of currency crises by developing statistical methods like early warning systems (Kaminsky et al. 1998). 2.1. New Keynesian Theory: Asymmetric Information New Keynessian theorists have used asymmetric information to reformulate ideas of Kindleberger. The asymmetric information framework used to understand why the financial crises occur particularly in developing countries and the importance of banking sector. Asymmetric information a situation in which one party has much accurate information than the other party and leads to adverse selection and moral hazard problems in financial system. Adverse selection occurs before the financial transaction when loans might be made bad credit risks, unlikely to pay it back who most take out a loan even though there are good credit risks. This feature based on the Market for Lemons Nobel prized article by Akerlof (1970). A lemon problem occurs when lenders have trouble determining good credit risks or bad credit risks in the equity or debt markets. Moral hazard problems occur after the transaction that the borrower has incentives to engage in immoral activities that make it less likely to loan will be repaid (Mishkin, 1996:1-3). Mishkin noted that there are four types of problems that may lead to increase in asymmetric information problems interact to produce financial crises; deterioration of financial sector and nonfinancial balance sheets due to changes in asset prices which are key factors for currency crises, increase in uncertainty and increase in interest rates (Mishkin, 2001:3). According to Diamond (1984) banks have advantages in collecting information to reducing asymmetric information because of low cost monitoring and preventing risk taking by borrowers. 2.2. Post Keynesian Theory: Debt and Financial Instability Approach According to this approach financial crisis may occur when the increased deficit will have to be financed by the money growth rate and also through debt in cyclical expansion. We identify two of the several views of debt and financial instability approach, developed by Minsky (1977, 1982) and Kindleberger (1978). Minsky states neoclassical synthesis accepted from Keynes which is that a capitalist economy with sophisticated financial practices is inherently unstable. Minsky redefined terms of financial fragility and the financial instability hypothesis (FIH) or known as Minsky Hypothesis. He first uses the term of Wall Street where current transactions are financed by debt and refers to the importance of debt-financing. Financial fragility is a micro-structural concept, each firm as having one of three types of financial structure; robust structures, ponzi structures or fragile structures (Dymski, 1997:503). The term of financial instability defined that because of financial fragility, changes in cash-flow relations transform robust structures in which the firm can meet its financing obligations even its realized assets-return outcomes are disastrous (Minsky, 1982:24). In 1989 he linked between financial

Middle Eastern Finance and Economics - Issue 5 (2009)

109

instability hypothesis and Minsky Hypothesis, described that both Ponzi and speculative financial posture increase so that borrowers and lenders risks and the ratio of debt financing to internal financing increases (Dymski, 1997:504). According to Kindleberger there are three phases; manias, panics and crashes, of financial crisis. Mania is speculative excess that occur during expansion period, economic agents shift from money to real assets. People expect that the asset price will continue to rise and no possibility to fall. Thus, they continue to buy at higher prices. Panics, the second phase at peaks they try to shift from real assets to money. Finally the last step crashes refers the collapse of real asset prices that was acquired during the mania (Kindleberger, 1978:5). However crash did not lead to a panic in money market or bank runs in financial system if there is an effective role taken by the Federal Reserve. Kindleberger suggest that the importance of the international lender of last resort. 2.3. Theory of Speculative Bubbles The term bubble was replaced to mania by Flood and Garber (1982). The word bubble implies that might easily collapse. A price bubble exists when the nominal value of assets is higher than real value. Bubbles may arise independently of the cyclical expansion (Schwartz, 281). Any reason or any doubt that the value of the asset will continue to rise, at that point some people stop buying and begin to sell their assets. So the phase mania has ended. According to monetarist economists both mania and panic would be avoided if only the money supply was stabilized at a constant growth rate. So bubbles would occur rather than the pattern of monetary growth was free of destabilizing cyclical movements (Schwartz, 282). Wood (1983) noted that market bubbles where enormous variations in price but a little variations in quantities. The recent literature under fixed Exchange rate regime on bubbles encompasses runs on a currency to other currencies. The market will shift from overvalued currency to an undervalued currency (Flood and Garber, 1982). 2.4. Random Withdrawal Risk As soon as the depositors believe that a bank had destroyed its capital base, it is optimal each depositor would try to withdraw all his deposits randomly (Benin, 2001:3). This theory was simply based on the traditional view of bank panics, one is that they are random event, unrelated to change in the macroeconomic indicators (Kindleberger, 1978). As indicated by the several experiences in the 1990s, a sudden withdrawal of banks deposits may have similar effect with bank run. Economists use loan losses or erosion of bank capital to define the banking crises. According to Demirg-Kunt and Detregiache (1997) one of these reasons had to hold; rescue operation cost more than 2 per cent of GDP, and the ratio of non performing loans to total assets more than 10 per cent (Demirg-Kunt and Detregiache, 1997:12). A systemic banking crises phase in two stages. In the first phase, banks make high risky loans to repaid or non-performing loans (NPLs) quietly destroying the banks balance sheets. The second phase begins when the economic agents take actions to remedy this situation which is publicly reported (Benin, 2001:1).

3. Theoretical Currency Crisis Models

The theoretical explanations for speculative attacks and balance of payments crises are first proposed by Krugman (1979). First-generation models mainly emphasize the importance of weak economic fundamentals and macroeconomics policies which are inconsistent with fixed exchange rate as causes of currency crises. In this canonical currency crises model has some important virtues. Firstly, many currency crises clearly do reflect a basic inconsistency between domestic policies, typically the persistence of monetization to budget deficits and the attempt to maintain a fixed exchange rate and abrupt runs on a currency need not reflect either investment irrationality or manipulations. In this speculative attack model, speculators only accelerate the process, so speculators behavior is not the source of the crisis (Burkart and Coudert, 2002:2). First-generation models are accounted well for

110

Middle Eastern Finance and Economics - Issue 5 (2009)

many of the currency crises in 1970s, 1973-1981 Mexico and 1978-1981 Argentina and also for 1982 debt crises. These models failed to explain such as in Chile (1982), Europe (1992-1993), Mexico (1994-1995) and Asia (1997-1998). Second generation models are first exemplified by Obstfeld (1986, 1994), financial crises are unexpected situation, as a result of pessimist expectations in a weak economic conditions. Therefore, second generation models consider nonlinear behavior rules by private sector and government. These interactions can lead to multiple solutions and financial crises are not always necessarily determined by the weak fundamentals. The main innovation of these models lies in identifying expectations of the economic agents has an important role for predicting the currency crises (Mariano et al.:3). As indicated in Obstfeld (1996), the expectation of a collapse leads to higher wages, higher interest rates and lower employment. For instance, possibility of banking crises will also increase. Third generation models are developed after the Asia crises (1997) and can be dubbed as a contagion models. The crises are accompanied by speculative pressures in other countries; there has been much discussion of contagion models. Third generation models attempt to identify why crises in one country might trigger crises in another and what features of countries explain such contagion effects findings are rather than similarity of macroeconomic fundamentals, trade linkages have more explanatory power. Because devaluation reduces the price competitiveness or financial intermediaries to liquidate other emerging market assets (Mason, 1998:4). These new models consider some disputed issues such as moral hazard or asymmetric information problem that lead to an underpricing of the risks in emerging markets and herding behavior of banking sector (Kibritcioglu et al., 1999:5). So both exchange rate crises and banking crises occur together in a financial system.

4. Empirical Literature
With the large numbers of financial crisis occurred in the last decade of 20th century, many researchers have paid attention to this phenomenon. The empirical literature proposes different rules for the predicting varied experiences of currency crises. Signal approach as an early warning system, logit/probit models, and regression models are carried out to prediction of crisis by using a set of economic indicators. In these models, researchers have tried to explain the main reasons behind the crises. According to leading indicators or signaling approach, potential determinants of currency crises is analyzed within tranquil (not in crisis period) and pre-crisis periods (from 12 to 24 months before crisis), abnormal behavior variables classify as a leading indicators. Signal approach is a non-parametric methodology was first used by Kaminsky, Lizondo and Reinhart (KLR-1997). In this approach, several indicators are observed which tend to exhibit an unusual behavior before potential crises. The indicators are evaluated according to their power of giving signal. When an indicator exceeds a certain threshold value, this is interpreted as a signal and currency crises may occur within 24 months. In KLRs study, 76 currency crises from a sample of 15 developing and 5 industrial countries during 19701995 are examined. The variables included are international reserves, imports, exports, the terms of trade, deviations of the real exchange rate from trend, the differential between foreign and domestic real interest rates on deposits, excess real M1 balances, the money multiplier (of M2), the ratio of domestic credit to GDP, the real interest rate on deposits, the ratio of lending to deposit interest rates, the stock of commercial banks deposits, the ratio of broad money to gross international reserves, an index of output and an index of equity prices. They followed the study of Eichengreen et al, but excluded interest rate differentials in their index. The best indicators which predict the currency crises are exports, deviations of the real exchange rate from trend, the ratio of broad money to gross international reserves, output and equity prices. Kang studied on Korean currency crises which mainly focused on signal approach. As well as signal approach, OLS estimation and logit estimation are used in order to compare the results. In this study, 22 indicators are used but export concentration ratio, change in S&P rating, foreign debt of banking sector/reserves, M2 multiplier, industry inventory index/output index are better indicators than

Middle Eastern Finance and Economics - Issue 5 (2009)

111

others in signal approach. However, among 22 indicators, seven indicators are found statistically significant in linear regression analysis. Commonly, terms of trade, capital account balance/GDP, M2/reserves and stock prices are statistically significant in both models (OLS and Logit model). Eichengreen, Rose and Wyplosz (1996) analyzed spillover effects among 20 countries during the period of 1959 and 1973, 77 crises. They built a currency crisis index constructed by changes in three variables international reserves, Exchange rate and interest rates. Findings are speculative attacks on fixed exchange rates play a significant role in currency crises. Krugman (1996) studied on 5 industrialized countries are affected by the ERM (1992-1993) crises by using annual, quarterly and also daily data over period 1988-1995. In this study four indicators, inflation, output gap, unemployment rate, public debt to GDP ratio are used and all these indicators support that ERM does not provide self-fulfilling crises. Ari and Dagtekin (2006) analyzed the 2000/2001 Turkish financial crisis with the early warning systems. They built a crisis index that is similar to the study of Eichengreen et al (1995), Sachs et al. (1996) and Kaminsky et al. (1998). So, in the first part of the index; real exchange rate, international reserves and real interest rate are used, in the second part of the index; they added the risk measures of the banking system which are the claims by the private sector, foreign exchange liabilities and total deposits. In this study, early warning indicators include current imbalance, impact of financial liberalization, debt, budgetary imbalance, monetary imbalance, degree of opening, international shocks, financial fragility. By using these indicators, they examined three models namely OLS estimation, logit estimation and of course signal approach. In these estimations, all of the indicators are found statistically significant. Corsetti, Pesenti and Roubini (1998) studied on 24 countries whose economic performances are affected by the 1997 Asia crisis. They construct a crisis index as a measure of speculative pressure on a country currency at first. Then they compute a set of indexes of financial fragility, external imbalances, official reserves adequacy and fundamental performance. The crisis index consists of exchange rate depreciation and foreign reserves. A negative sign of crisis index indicates more severe currency crisis. They found that weak cyclical performances, low foreign exchange reserves and financial deficiencies are the main reasons behind the Asian collapse. Cartapanis, Dropsy ve Mametz (2002) studied on six Asian nations (India, Indonesia, Malaysia, Philippines, South Korea, Thailand) by using their quarterly data for period 1976-1997. They construct indices of speculative pressure and use proxies for macroeconomic disequilibria and contagious effects in order to consider their impacts on a nations vulnerability to crisis. In their models, panel and logit regressions are estimated. The currency crisis index is calculated from real effective exchange rate and international reserves. Unlike the other studies, they estimate the continuous crisis index model and binary crisis index model. They use the estimates of WLS (weighted least squares) regressions for the continuous indices. In their estimations, 14 explanatory variables are used that are grouped as indicators of macroeconomic unsustainability, financial vulnerability and crisis contagion. They compare the results of LSDV model and logit model. According to estimation results, currency overvaluation indicators are found statistically significant in both models. The empirical results state that both currency overvaluation and pure contagion effects are the leading indicators of Asian crisis. Berg and Pattillo (1999) try to predict currency crises in 1997 by using and comparing three different models; KLR (1998) signals approach, Frankel and Rose (1996) probit model, Sachs, Tornell and Velasco cross-country regression model. They uses probit/logit model and obtained KLR model by adding two indicators; the current account, the ratio/M2 reserves and they are in favor of the probit models. Their findings are the KLR model was designed explicitly and the most successful. The indicators that have most explanatory power to predict the crises are (1) the deviation of the real Exchange rate from trend, (2) the current account, (3) reserve growth, (4) export growth and (5) the ratio of M2 to reserves. Cevis (2005) use panel data for 22 developing countries in order to estimate currency crisis for period 1990-2002. In this study, a currency crisis index as dependent variable is developed by using

112

Middle Eastern Finance and Economics - Issue 5 (2009)

real exchange rates, real interest rates and foreign exchange reserves. In addition, 8 explanatory variables are used for OLS, LSDV, REM1, LSDV&TIME and REM2 model. According to results, the sign of the variables show the same effect like former studies. The best model predicting the currency crisis is found LSDV&TIME model.

5. Data and Methodology

The model is estimated for a panel of annually observations for 15 emerging countries covering the period 19872007. The countries included are Argentina, Brazil, Chile, Colombia, India, Indonesia, South Korea, Malaysia, Mexico, the Philippines, Russia, Thailand, Turkey, Uruguay and Venezuela. Financial and macroeconomic variables are extracted from the World Development Indicator (World Bank) Online. A currency crisis is defined as a situation in which an attack to a currency leads to substantial reserve losses, or to a sharp depreciation of the currency, or to both. This crisis definition includes both successful and unsuccessful attacks to a currency. Crises are identified by the behaviour of a currency crises index, that is constructed for each country. In order to explain our endogenous variable, we built a series of indicators taken from the financial literature to measure the vulnerability of an economy to the speculative attacks. Following Eichengreen et al. (1995), Sachs et al. (1996), Kaminsky et al. (1997) or Cartapanis et al. (2002), the first part of our currency crisis index (CCI) is made as a weighted average of real exchange rate changes, international reserves changes and real interest rate changes. In this paper, we use CCI as a dependent variable appears in the following way; %ri %RERi %IRi CCIi = % r % RER % IR where RER = (NERP*) / P RER = Real exchange rate (an increase corresponds to a real depreciation of the domestic currency) NER = Nominal exchange rate (LCU2/USD) P* : Consumer prices index of a country (base 100 = 2000) P : Consumer prices index US (base 100 = 2000) IR : International reserves minus gold (in USD) r : real interest rate (deflated by the inflation rate to the consumption of the domestic country) : Standard deviation of the variables In this paper we use panel data methodology to assess whether some variables have explanatory power regarding the currency crises. The unbalanced panel data consists of 15 countries over 21 years which yields a sample of 313 observations. Our choice of the 7 indicators is based on the prior literature and on the availability of annualy data. Table 1 presents the definition of explanatory variables and their detailed informations used in the present study.

Local currency unit.

Middle Eastern Finance and Economics - Issue 5 (2009)

Table 1:
Notation GROWTH CABTR BANKDC BANKLQD M2GROWTH EXIM PORTFDI

113

Explanatory variables of the models and their awaited impact on the currency crisis index
Indicators GDP growth (annual %) Current account balance / Reserves Domestic credit provided by banking sector (%of GDP) Bank liquid reserves to bank assets ratio (%) M2 growth (annual %) Export (%growth) / Import (%growth) Portfolio investment / Foreign direct investment Expected Signs (-) (-) (+) (+) (-) (-) (+) Reference Obsfeld (1986), (1996) IMF (1998), Frankel and Rose (1996), Berg and Patillo (1998), Esquivel and Larrin (1998), Cartapanis, Dopsy and Mametz (2002) Corsetti, Pesenti and Roubini (1998), Kaminsky ve Reinhart (1996), Kaminsky (1998), Sachs, Tornell and Velasco (1996) IMF (1998), Rodric and Velasco (1999) Krugman (1979) Eichengreen, Rose and Wyplosz (1996), Berg ve Patillo (1998), Vlaar (2000) IMF (1998)

5.1. Estimation Results Before applying ordinary least squares (OLS), least square dummy variable or fixed effect model (LSDV), random effect model (REM) and LSDV&TIME model and the models used to assess the relationship between CCI and seven crises indicators, we conducted several pooled panel unit root tests, including the Levin, Lin and Chu-LLC (2002) test; the Im, Peseran and Shin-IPS (2003) test; the Maddala and Wu (1999) Fisher ADF test; Choi (2001) and Fisher Phillips-Peron test to ensure that the data are suitable for our analysis. The null hypothesis of the tests is that each individual time series contains a unit root against the alternative hypothesis that each time series is stationary. The results of the pooled panel unit root tests are presented in Table 2. Table 2 shows support for the hypothesis that each individual time series does not contain a unit root at level that means all of the series are stationary.
Table 2: Panel Unit Root Test Results
PP Fisher 2-stat. 51.4998* 49.2673* 95.5759* 221.608* 106.879* 82.9103* 382.822*

Variables LLC t-stat. IPS W-stat. ADF Fisher 2-stat. -2.02237** -1.72524** 45.3289** BANKDC -2.68014* -2.55347* 55.8200* BANKLQD -2.47738* -3.22815* 63.6616* CABTR -8.34702* -7.65098* 114.818* EXIM -3.77426* -3.76301* 63.6110* GROWTH -2.40963* -2.2729** 50.8903** M2GROWTH 1.27514 -3.58325* 60.6714* PORTFDI Note: * Significant at 1% level. ** Significant at 5% level. *** Significant at 10% level.

Table 3 summarizes the estimation results. In OLS model BANKDC, CABTR, EXIM and M2GROWTH indicators have expected sign respect to theory. These variables are statistically significant except EXIM. On the other hand, BANKLQD, GROWTH, PORTFDI dont have the expected sign and among these three variables only GROWTH is statistically significant. In LSDV (fixed effect in cross section) model, only GROWTH and PORTFDI dont have expected sign and they are statistically insignificant. However other indicators have expected sign but only BANKDC and CABTR are statistically significant. In random effect model (REM) BANKLQD, GROWTH and PORTFDI dont have expected signs, in contrast to BANKDC, CABTR, EXIM and M2GROWTH. Finally in LSDV&TIME (fixed effect in both cross-section and period), only GROWTH and PORTFDI dont have expected sign and also statistically insignificant, other variables have expected sign but only CABTR is statistically significant.

114
Table 3:
Variables

Middle Eastern Finance and Economics - Issue 5 (2009)

Panel Regression Analysis Results and Test Statistics
LSDV&TIME 0.00343 (0.64922) 0.01644 (0.92810) -0.76269 (-4.529007)* -0.00269 (-0.82699) 0.01006 (0.45323) -0.00035 (-0.93490) -0.00924 (-1.06771) -0.93468 (-2.35432)** 0.38 0.29 2.07 4.104 -551.69

OLS LSDV REM 0.00628 0.00789 0.00628 BANKDC (2.56212)** (1.39539)* (2.53287)** -0.00024 0.01493 -0.00024 BANKLQD (-0.01789) (0.75182) (-0.01768) -0.61965 -0.68953 -0.61965 CABTR (-3.90931)* (-3.90526)* (-3.86467)* -0.00391 -0.00396 -0.00391 EXIM (-1.11893) (-1.09675) (-1.10616) 0.03565 0.03038 0.03565 GROWTH (1.69966)*** (1.31654) (1.68023)*** -0.00071 -0.00058 -0.00071 M2GROWTH (-2.30317)** (-1.50292) (-2.27687)** -0.00277 -0.00303 -0.00277 PORTFDI (-0.29755) (-0.31375) (-0.29415) -1.03739 -1.27030 -1.03739 Constant (-4.08433)* (-2.85905)* (-4.03770)* 0.10 0.12 0.10 R2 2 0.08 0.06 0.08 R 1.75 1.79 1.75 DW-stat. 4.911 1.937 4.911 F-stat. -610.57 -606.81 --LM-stat. Note: * Significant at 1% level. ** Significant at 5% level. *** Significant at 10% level.

According to the estimation results, three indicators have important effect in common on crisis index namely BANKDC, CABTR AND M2GROWTH. Actually, BANKDC variable can be evaluated as an indicator of the fragility of banking system and the emergence of banking crises is associated with weak macroeconomic conditions. So, the sign of BANKDC is expected positive. There is evidence that outcome of a lending boom substantial reserve losses and precedes financial crises. Furthermore, the current account deficit is the most important indicator of vulnerability against currency crisis in a country. By looking at the estimation results, it can be said that the CABTR variable has important power to explain the currency crisis. Feature of recent currency crises is low amounts of international reserves in selected countries seemed to be more vulnerable to crises. In the light of first-generation models, current account balance, domestic interest rates and the real Exchange rates may used the leading indicators for currency crises and the relationship between the money supply and the CCI is negative.

6. Conclusion
In this paper we try to explain the causes of currency crises by using some crucial indicators. In order to analyze the currency crises we include the GDP growth, bank liquid reserves to bank assets ratio, domestic credit provided by banking sector, current account balance to reserves ratio, M2 growth, export growth to import growth ratio and portfolio investment to foreign direct investment ratio variables. These indicators are commonly used in signal approach, linear regression model and logit/ probit models. Here, the unbalanced panel data estimation results show that the leading indicators of currency crises are M2 growth, domestic credit provided by banking sector and current account balance to reserves ratio. The results of our empirical analysis provide evidence in support of the thesis that currency crises are systematically related to the fundamental weaknesses in the real and financial sectors of the economy. For example, the current account deficit is one of the main reasons of the weaknesses toward a possible currency crisis. Thus crises do not have self-fulfilling components. At the same time, our empirical results are parallel with the other empirical studies as well.

Middle Eastern Finance and Economics - Issue 5 (2009)

115

References
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] Ari A. and Dagtekin R. (2006), Early Warning Signals of the 2000/2001 Turkish Financial Crisis, Available at: http://crem.univrennes1.fr/manifestations/GDREmonnaie07/papiers/ARIDAGTEKIN.pdf. Benin D. O., What triggers a systemic banking crises?, Columbia University, June 19, 2001, p.3 Berg A. and Pattilo C. (1999), Are Currency Crises Predictable A Test, IMF Staff Papers, Vol.46, No: 2, pp.107-138. Brggemann A. And Linne T. (1999), How Good are Leading Indicators for Currency and Banking Crises in Central and Eastern Europe? An Empirical Test, Halle Institute for Economic Research Discussion Paper, No. 95. Burkart O. and Coudert V. (2002), Leading Indicators of Currency Crises for Emerging Countries, Emerging Market Review, Vol.3, No:2, pp.107-133. Cartapanis A., Dropsy V. and Mametz S. (2002), The Asian Currency Crises: Vulnerability, Contagion, or Unsustainability, Review of International Economics, 10(1), pp.79-91. Cevis I. (2005), Para Krizlerine Ampirik Bir Yaklam, Ankara, Sermaye Piyasas Kurulu. Corsetti G., Pesenti P. and Roubini N. (1998), Fundamental Determinants of the Asian Crisis: A Preliminary Empirical Assessment. Available at: http://pages.stern.nyu.edu/~nroubini/papers/jimf06.pdf. Choi I. (2001), Unit root tests for panel data, Journal of International Money and Finance, Vol. 20, pp. 249272. Demirg-Kunt A. and Detregiache E., The Determinants of Banking Crises: Evidence From Developed and Developing Countries, 1997, pp.1-27. Diamond, D., Financial Intermediation and Delegated Monitoring, Review of Economic Studies, Vol.53, pp.393-414, 1984. Dymski G. A., Deciphering Minskys Wall Street Paradigm, Journal of Economic Issues, Vol.XXXI, No.2, June 1997, pp.501-508. Eichengreen B., Rose A. and Wyplosz C. (1996), Contagious Currency Crises. NBER Working Papers, No. 5681. Esquivel G. and Larrain F. (1998), Explaining Currency Crisis, HIID, No:666. Frankel J. A. and Rose A.K. (1996), Currency Crushes in Emerging Markets: An Emprical Treatment, Journal of International Economics, No.41, pp.351-366. Im, K. S., Pesaran M.H. and Shin Y. (2003), Testing for unit roots in heterogeneous panels, Journal of Econometrics, Vol. 115, pp. 5374. IMF (1998), Financial Crises: Characteristics and Indicators of Vulnerability, in: World Economic Outlook, Washington, DC, May, pp.110-142. Kaminsky G. (1998), Currency and Banking Crises: The Early warning of Distress, IFDP Working Paper, No.629. Kaminsky G., Lizondo S. and Reinhart M.C., (1998), Leading Indicators of Currency Crises, IMF Staff Papers, Vol.45, No: 1. Kang I. S., A Study on the Early Warning Indicators of Currency Crisis: A Regional Perspective, available at: http://wwwsoc.nii.ac.jp/jsie/Tohoku_Univ/211_2.html. Kibritcioglu B. (2001), Early Warning Indicators of Currency Crises in Turkey, University of Illinois at Urbana Campaign, pp. 1-44. Kibritcioglu B., Kose B. and Ugur G. (1999), A Leading Indicators Approach to the Predictability of Currency Crises: The Case of Turkey, General Directorate of Economic Research the Undersecretariat of Turkey, Ankara. Kindleberger C. P., Manias, Panics and Crashes: A History of Financial Crises. Krugman P. (1979), A Model of Balance of Payment Crises Ohio State University Press, Journal of Money, Credit and Banking, Vol.11, No:3, pp.311-325, available at:http://www.columbia.edu/~rhc2/Spring2006/G6904/Papers/Krugman_79.pdf.

116 [25] [26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40] [41]

Middle Eastern Finance and Economics - Issue 5 (2009) Krugman P., Are Currency Crises Self-Fulfilling?, NBER Working Paper, 1996. Krugman P. Currency Crises, available at: http://web.mit.edu/krugman/www/crises.html. Krugman P. (1999), Balance Sheets, the Transfer Problems and Financial Crises, available at: http://web.mit.edu/krugman/www/FLOOD.pdf. Levin A., Chien-Fu L. and Chia-Shang C. (2002), Unit root test in panel data: Asymptotic and finite sample properties, Journal of Econometrics, Vol. 108, pp. 125. Maddala G.S. and Shaowen W. (1999), A comparative study of unit root tests with panel data and a new simple test, Oxford Bulletin of Economics and Statistics, Vol. 61, pp.631652. Mariano R. S., Gultekin B., Ozmucur S. and Shabbir T., Models of Economic and Financial Crises, available at: http://www.luc.edu/orgs/meea/volume2/ozmucur1.pdf. Mason P., (1998), Contagion Monsoonal Effects, Spillovers and Jumps Between Multiple Equilibria, IMF Working Paper, pp.1-32. Minsky H. P. (1982), The Financial Instability Hypothesis: Capitalist Process and The Behavior of the Economy, Cambridge University Pres, pp.13-39. Mishkin F. S. (1996), Understanding Financial Crises: A Developing Country Perspective, NBER, Working Paper 5600. Mishkin F. S. (2001), Financial Policies and the Prevention of Financial Crises in Emerging Market Economies, NBER, Working Paper 8087. Obstfeld M. (1986), Rational and Self-Fulfilling Balance of Payments Crises, American Economic Review, Vol. 40, pp.72-80. Obstfeld M. (1994), The Logic of Currency Crises, NBER Working Papers, No: 4640. Obstfeld M. (1996), Models of Currency Crises with Self-fulfilling Features, European Economic Review, Vol.40, pp.1037-1047. Rodric D. and Velasco A. (1999), Short-Term Capital Flows, NBER Working Papers, No. 7364, September. Sachs J., Tornell A. and Velasco A. (1996), Financial Crisis in Emerging Markets: The Lessons from 1995, Brookings Papers on Economic Activity, Vol. 27, No: 1, pp. 147-199. Schwartz A. J., Real and Pseudo Financial Crises, pp. 271.288. Vlaar P. J.G., Early Warning System for Currency Crises, available at: http://www.bis.org/publ/confer08m.pdf.