Rapid credit growth and current account deficit as the leading determinants of financial crises

Ganioğlu, Aytül Working Paper Rapid credit growth and current account deficit as the leading determinants of financial crises Economics Discussion Papers, No. 2013-35 Provided in Cooperation with: Kiel Institute for the World Economy – Leibniz Center for Research on Global Economic Challenges Suggested Citation: Ganioğlu, Aytül (2013) : Rapid credit growth and current account deficit as the leading determinants of financial crises, Economics Discussion Papers, No. 2013-35, Kiel Institute for the World Economy (IfW), Kiel This Version is available at: https://hdl.handle.net/10419/76807 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by/3.0/ Received May 24, 2013 Accepted as Economics Discussion Paper June 25, 2013 Published July 8, 2013 © Author(s) 2013. Licensed under the Creative Commons License - Attribution 3.0 Discussion Paper No. 2013-35 | July 08, 2013 | http://www.economics-ejournal.org/economics/discussionpapers/2013-35 Rapid Credit Growth and Current Account Deficit as the Leading Determinants of Financial Crises Aytül Ganioğlu Abstract In this study, the main purpose is to analyze the factors that stimulate the probability of financial crises. The period of analysis covers the years of 1970-2008, thereby including the impact of recent global financial crisis. The analysis aims to make a comparison for the developed and developing country financial crises separately. Panel logit estimation technique is used for the analysis which includes 24 developed and 26 developing countries, amounting to 50 countries as total. According to estimation results, current account deficit and credit expansion carry the risk of raising the probability of financial crises significantly both in advanced countries and developing countries. More specifically, credit expansions in developed countries and current account deficits in developing countries raise the probability of financial crises more strongly. JEL C33 E51 F41 G01 Keywords Financial crisis; predictors of financial crisis; rapid credit expansion; current account deficit Authors Aytül Ganioğlu, Central Bank of Turkey, Monetary Policy and Research Department, [email protected].tr The views expressed in this paper are those of the author and do not reflect the official views of the Central Bank of Turkey. All errors are our own. Citation Aytül Ganioğlu (2013). Rapid Credit Growth and Current Account Deficit as the Leading Determinants of Financial Crises . Economics Discussion Papers, No 2013-35, Kiel Institute for the World Economy. http:// www.economics-ejournal.org/economics/discussionpapers/2013-35 1 1. INTRODUCTION There has been frequent financial instability in both developed 3 and developing 4 countries accompanied by the increased global capital mobility after the collapse of the Bretton Woods system in 1971 5 . In some cases, a severe enough financial instability even led to almost complete breakdown in the functioning of the financial markets, which is called as a financial crisis. As one type of crisis may develop into another, they might also take place together 6 . There are different explanations about how crises occur. In the literature, every model has been developed in the aftermath of a new crisis in order to explain the dynamics of the crisis and desire to generalize main aspects. However, both theoretical and empirical analyses of the crises in the literature give direction to different conclusions. Since there is no single way of measuring explanatory variables in the empirical analysis, besides no agreement on which explanatory variables to include, different results are obtained as regards to the impact of explanatory variables included in the empirical analyses. Hence, these models have not been successful in generating a consensus, as apparent from controversial views in the literature. Even, there is no consensus in the literature as regards to the definition of crisis. As a matter of fact, one cannot make generalization as to the macroeconomic conditions under which financial and currency crises have occurred. For example, in some crises such as the ones in Mexico, Thailand and Turkey, current account deficits were large 3 The episodes in developed countries include the banking and real estate crises in the United States lasting more than a decade from the late 1970s, the major slumps in the global stock market in 1987 and 1989, the currency crisis of the European Monetary System (EMS) in 1992 and the ongoing instability in Japanese financial markets that started with the bursting of the bubble in the early 1990s (Akyüz and Cornford, 1999:15). See also Fourçans and Franck (2003) 4 The episodes of crises in developing countries include the Southern Cone crisis of the late 1970s and early 1980s, the Mexican crisis of 1994-1995, the East Asian crisis beginning in 1997, the Russian crisis of 1998, Argentina crisis in 2001 and Turkish crises in 2000 and 2001 (Akyüz and Cornford, 1999:15). 5 This classification of the crises as post-Bretton Woods crises belongs to Akyüz and Cornford (1999:15). 6 During the 1970s, there was no apparent link between currency and banking crises, when financial markets were highly regulated. In the 1980s, banking and currency crises become more interlinked, as many of the countries have both currency crises and banking crises around the same time called twin crises by Kaminsky and Reinhart (1999). Then, the link between banking and currency crises began to take attention. 2 and unsustainable, while it was small in the crises of Indonesia and Russia. Although there were significant overvaluation of the domestic currency in the crises of Mexico, Russia, Brazil and Turkey which used exchange rate as a nominal anchor to bring inflation down, this has not always been the case, as the appreciation of currency was moderate or negligible in most East Asian countries. In addition, while large budget deficits were associated with the crises in Russia, Brazil and Turkey, the budget was balanced or in surplus in Mexican and East Asian crises. Finally, in Brazilian and Russian crises, external debt was owed primarily by the public, while primarily it was by the private sector in East Asian crises 7 . The global financial crisis has deeply influenced the views related to the interaction between macroeconomic outcomes and financial system. Among various theoretical approaches in the literature, the two diverse approaches need to be mentioned here. On the one side, it is suggested by the monetarist view of Friedman and Schwartz (1963), as well as recently dominant Neo-Keynesian synthesis of Woodford (2003) that macroeconomic outcomes are broadly independent of the performance of financial system. On the other side, it is argued by Fisher (1933), Minsky (1978), Bernanke (1983, 1993) and Gertler (1988) to varying degrees that financial system can have a strong and dominant impact on macroeconomic outcomes 8 . In the aftermath of the global financial crisis, a new interest sparkled about the fluctuations in monetary aggregates and credit as well as their roles in the amplification, propagation and generation of shocks especially during financial distress 9 . The view that has been influential especially after the global crisis is that expansion in credit aggregates as well as increased risk involves important information for policy makers monitoring financial and economic stability especially about the likelihood of future financial crises. Furthermore, it is 7 Akyüz and Cornford (1999:17) 8 For more discussion, see Schularick and Taylor (2010) 9 Schularick and Taylor (2010) 3 argued that excessive credit growth generates risks such as “imbalances” and “financial instability” 10 . Adherence to the money view has been seriously called into question by the crisis. Analysis of Schularick and Taylor (2010) clearly suggests that “the credit system matters above and beyond its role as propagator of shocks as in the financial accelerator model. The credit system seems all too capable of creating its very own shocks, judged by how successful past credit growth performs as a predictor of financial crises”. Even though the association between excessive credit expansion and financial crises 11 is not new, the empirical evidences regarding this relationship are very few. Although financial crises of developing countries are examined more often in the literature 12 , studies related to financial crises of developed countries are very few, since financial crises in developed countries are rather rare events. In two recent studies 13 , credit booms appear as a strong predictor of financial crisis. In Schularick and Taylor (2010) study, credit booms are stronger predictor of financial crisis than monetary aggregates. In the study of Jorda and others (2010), credit boom over the previous 5 years is indicative of a heightened risk of financial crisis, and is a superior predictor of financial crisis than current account imbalances. The main purpose of this study is to contribute to “few” empirical studies in the literature examining the financial crises of developed countries as well as introducing the impact of global financial crisis into the analysis of financial crises of developing countries. An almost standard set of macroeconomic variables are involved in the panel data estimations in this study. Additionally, we have been inspired by the study of Jorda and others (2010) to introduce “credit boom” as an explanatory variable that propagates financial crises. Our 10 Borio and Lowe (2002, 2003); White (2004); Goodhart (2007) 11 Kindleberger (1978); Hume and Sentence (2009); Reinhart and Rogoff (2009); Eichengreen and Mitchener (2003); Caprio and Honohan (2008) 12 McKinnon and Pill (1997); Kaminsky and Reinhart (1999) 13 Schularick and Taylor (2009); Jorda et. al. (2010). In both studies, the analysis covers 14 developed countries for the period of 1870-2008. 4 analysis differs from that of Jorda and others (2010) 14 in that number of developed countries involved in the analysis has been extended to 24 countries while the period of analysis has been restricted to the period of 1970-2008, considering the fact that dynamics of crises change substantially when we extend the period. Furthermore, we incorporate developing countries into the analysis so that we can make a comparison in terms of financial crises of developed and developing countries separately. Last but not least, sources of database are different, especially as regards the banking crises database. Hence, findings of this study have been different from that of Jorda and others (2010), especially related to the impact of credit boom on financial crises confirming a non-consensus on which explanatory variables to include in the analysis of financial crises as mentioned above. Empirical results in this study point to the robust significance of current account deficit in leading to crises in developing countries, carrying a stronger risk of increasing the probability of financial crises than that of credit expansion. Our test results indicate that in developed countries, both current account deficit and credit expansion together with monetary expansion raises the risk of financial crises, while the credit expansion appears as having a more robust impact. This paper is organized as follows: Section 2 discusses data and descriptive statistics. Section 3 is devoted to our empirical analysis and findings. Section 4 concludes. 14 Their analysis covers 14 developed countries for the period of 1870-2008. 5 2. DATA AND DESCRIPTIVE STATISTICS The data set has annual data for 50 countries, (24 developed 15 and 26 developing 16 ) and covers the period of 1970-2008. As regards the dataset, Appendix 1 provides information about all variables by name, definition, sources and the time period that the data covered. All regressions included a standard list of macroeconomic variables. These variables are inflation rate calculated by formula ln (1+ Inflation rate) (INF); current account balance (% of GDP) (CABGDP); GDP per capita growth rate (GDPPCGR); GDP growth rate (GDPGR); real interest rate (REALINTR); domestic credit provided by banking sector to private sector (% of GDP) (DOMCREDPR); money and quasi money (% of GDP) (M2GDP); broad money (% of GDP) (BROADMONEY); percentage change in 5-year moving average of domestic credit provided by banking sector to private sector (% of GDP) (CREDBOOM) as the macroeconomic factors likely to lead to a crisis; as well as interaction term which is the product of domestic credit provided by banking sector to private sector (% of GDP) and current account balance (% of GDP) (INTDOMCRCABGDP). Source of explanatory variables is World Bank World Development Indicators database. The dependent variable, banking crisis series are mainly based on Laeven and Valencia (2010, 2012) 17 . Detailed information about the definition of banking crisis is provided in the studies of Laeven ve Valencia (2010, 2012). Table 2 presents descriptive statistics of independent variables. Comparison of developed and developing countries as regards the averages of variables reveals the following 15 Developed countries are Austria, Australia, Belgium, Canada, Denmark, Finland, France, Germany, Greece, Iceland, Ireland, Israel, Italy, Japan, Korea, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden, Switzerland, United Kingdom and United States. 16 Developing countries are Argentina, Brazil, Chile, China, Colombia, Czech Republic, Egypt, Hungary, India, Indonesia, Kazakhstan, Latvia, Malaysia, Mexico, Morocco, Peru, Philippines, Poland, Russia, Slovenia, Sri Lanka, Thailand, Turkey, Ukraine, Venezuela and Zimbabwe. 17 See Appendix 2 for a detailed list of banking crises. 6 information: while average value of current account balance (% of GDP) for developed countries is -0,77 percent, it is -1,45 percent for developing countries; average value of domestic credit provided to private sector (% of GDP) is 83,6 percent for developing countries, as it falls to 46,8 percent for developed countries; average value of broad money supply (% of GDP) is 66,8 percent for developed countries, while it is 40,7 percent for developing countries; average of real interest rate is 4,69 percent for developed countries and 11 percent for developing countries. Furthermore, maximum value of DOMCREDPR and M2GDPare very high in developing countries as compared to developed countries. Table 1: Descriptive Statistics of Independent Variables Developed Countries Variable Name Number of Observation (N) Average Standard Deviation Min. Max. CABGDP 834 -0.77 4.71 -26.89 17.76 DOMCREDPR 926 83.6 48 0.03 319 M2GDP 548 79.7 47.6 15.3 242.2 BROADMONEY 616 66.8 39.24 15.3 242.2 INF 913 1.77 0.85 -2.25 5.92 GDPPCGR 934 2.37 2.5 -7.9 13.27 GDPGR 934 3.12 2.6 -7.28 13.6 REALINTR 724 4.69 6 -19.48 88 Developing Countries Variable Name Number of Observation (N) Average Standard Deviation Min. Max. CABGDP 729 -1.45 5.19 -22.68 19.8 DOMCREDPR 823 46.8 293 0 8404.02 M2GDP 827 70.49 423.37 6.21 7015.56 BROADMONEY 804 40.7 27.44 6.2 145.3 INF 799 2.6 1.35 -2.74 10.1 GDPPCGR 892 2.56 5.04 -31.34 18.56 GDPGR 892 4.08 5.18 -32 22.5 REALINTR 553 11 41.5 -91.7 578 7 Figures 1a and 1b provide a graphical representation of the relationship between banking crisis and current account balance/GDP ratio in year preceding the crisis. Developing country group form a cluster on this scatter plot. It is observed in the Figure 1a that in the year preceding the crisis, most of the countries run current account deficits and even large ones. On the scatter for developed countries (Figure 1b), countries running current account deficits and surpluses in the year preceding the crisis are more evenly distributed. Figures 1a: Current Account Balance/GDP Ratio (Developing Countries*) -12 -10 -8 -6 -4 -2 0 2 4 6 0 5 10 15 20 25 30 35 Current Account Balance/GDP Ratio (percent) *Argentina, Brazil, Chile, China, Colombia, Czech Republic, Egypt, Hungary, India, Indonesia, Japan, Malaysia, Mexico, Morocco, Peru, Philippines, Poland, Russia, Sri Lanka, Thailand, Turkey, Venezuela, Zimbabwe. Figure 1b: Current Account Balance/GDP Ratio (Developed Countries*) -12 -10 -8 -6 -4 -2 0 2 4 6 8 10 0 5 10 15 Current Account … *Finland, Germany, Ireland, Israel, Japan, Korea, Netherlands, Norway, Spain, Sweden, Switzerland, United Kingdom, United States Note: For each country, current account balance/GDP ratio is the one in the year preceding the banking crisis. 14 Table 4b: Panel Regressions (Developed Countries) Logit Fixed-Effect Model Regression Dependent Variable: Financial Crisis (7) (8) (9) (10) (11) CABGDP t - 1 -0.24 ** -0.4731 * -0.22 * -0.44 * -0.13 ** -(2.07) -(1.87) -(1.82) -(1.88) -(2.10) CREDBOOM t - 1 - - - - 0.04 - - - - (1.61) DOMCREDPR t - 1 0.05 *** 0.06 *** 0.05 *** 0.06 *** - (4.57) (3.30) (4.46) (3.30) - BROADMONEY t - 1 - - - - - - - - - - M2GDP t - 1 - - - - - - - - - - INTDOMCRCAB t - 1 0.002 ** 0.003 ** 0.002 ** 0.0025 ** - (2.08) (2.03) (2.01) (2.12) - INF t - 1 - 0.38 0.34 0.32 -0.46 - (0.44) (0.83) (0.36) -(1.38) GDPGR t - 1 - - 0.11 - - - - (0.86) - - GDPPCGR t - 1 - 0.16 - - - - (0.94) - - - REALINTR t - 1 - 0.03 - 0.01 - - (0.18) - (0.09) - LR chi2(df) 43.15 26.65 42.99 25.69 9.73 Prob>chi2 (p value) 0.00 0.00 0.00 0.00 0.02 Notes : *** 1 percent significance level, ** 5 percent significance level, *10 percent significance level. t-tests are provided in the parenthesis. Predictive ability of the model is tested for the regressions (2), (4) and (7), using ROC curve analysis, which is shown in Figure 3. xb1 ROC area represents predictive ability of regression (2), xb2 ROC area represents that of regression (4) and xb3 ROC area shows predictive ability of regression (7). The area under the ROC curve for the regressions (2) and (7), where domestic credit extension is involved, is greater than that of regression (4), where credit boom is placed into the regression. This means that predictive ability of the model involving domestic credit extension is better than that of including credit boom. Furthermore, 15 integration of current account balance and interaction term to the regression (2) that involves domestic credit raised predictive ability of the model only a little. In other words, predictive ability test results confirm the conclusion from the regression analysis that higher domestic credits are better indicative of an increasing risk of financial crisis than current account imbalances in developed countries. Figure 3: Predictive Ability Testing (Developed Countries) ROC curve comparison Panel logit fixed effect estimation results for developing countries are summarized in Tables 5a and 5b. Estimation results underline mainly the impact of two variables, namely current account imbalances (CABGDP) and domestic credit trends (DOMCREDPR), for developed countries as well, as in the case of developed countries. Similar to the results obtained for developed countries, domestic credit extension is again indicative of an increasing risk of financial crisis. On the other side, for developing countries, current account imbalances are more robust in raising the probability of financial crisis than credit trends. Interaction term of INTDOMCRCABGDP is not found to be statistically significant. Again in contrast to 0.00 0.25 0.50 0.75 1.00 Sensitivity 0.00 0.25 0.50 0.75 1.00 Specificity xb1 ROC area: 0.7585 xb2 ROC area: 0.6126 xb3 ROC area: 0.7633 Reference 16 estimation results related to developed countries, credit booms are not statistically significant in raising the probability of financial crisis. Credit booms, together with current account imbalances and control variables, are found to be not statistically significant. In other words, current account imbalances seem to be the most robust indicative of an increasing risk of financial crisis. Table 5a: Panel Regressions (Developing Countries) Logit Fixed-Effect Model Regression Dependent Variable: Financial Crisis (1) (2) (3) (4) (5) CABGDP t - 1 -0.16 *** - -0.15 *** -0.16 *** -0.16 *** -(3.67) - -(3.41) -(3.58) -(3.64) CREDBOOM t - 1 - - - - 0.0042 - - - - (0.45) DOMCREDPR t - 1 - 0.03 *** 0.03 *** - - - (3.31) (2.85) - - BROADMONEY t - 1 - - - 0.01 - - - - (1.04) - M2GDP t - 1 - - - - - - - - - - INTDOMCRCAB t - 1 - - - - - - - - - - INF t - 1 - - - - - - - - - - GDPGR t - 1 - - - - - - - - - - GDPPCGR t - 1 - - - - - - - - - - REALINTR t - 1 - - - - - - - - - - LR chi2(df) 15.34 15.53 23.20 14.21 16.34 Prob>chi2 (p value) 0.00 0.00 0.00 0.00 0.00 Notes : *** 1 percent significance level, ** 5 percent significance level, *10 percent significance level. t-tests are provided in the parenthesis. 17 Table 5b: Panel Regressions (Developing Countries) Logit Fixed-Effect Model Regression Dependent Variable: Financial Crisis (6) (7) (8) CABGDP t - 1 -0.22 *** -0.21 *** -0.16 *** -(3.15) -(2.47) -(3.47) CREDBOOM t - 1 0.004 0.41 DOMCREDPR t - 1 0.03 *** 0.02 * (3.23) (1.89) BROADMONEY t - 1 M2GDP t - 1 INTDOMCRCAB t - 1 0.001 0.001 (1.40) (0.81) INF t - 1 0.10 0.14 - (0.36) (0.86) GDPGR t - 1 - - GDPPCGR t - 1 - -0.01 0.017 - -(0.25) 0.42 REALINTR t - 1 - 0.01 - (0.64) LR chi2(df) 24.89 18.14 16.15 Prob>chi2 (p value) 0.00 0.00 0.00 Notes : *** 1 percent significance level, ** 5 percent significance level, *10 percent significance level. t-tests are provided in the parenthesis. Predictive ability of regressions (1), (2) and (3) can be viewed in Figure 4. Predictive ability of regressions (1) (2) and (3) is represented by the ROC areas xb1, xb2 and xb3 respectively. Predictive ability of the model (1) having current account balance as the sole explanatory variable (xb1) is much higher than the predictive ability of the model (2) that involves only domestic credit extension as the explanatory variable (xb2). When we let current account enter into the model in regression (3), predictive ability of the model has increased 18 significantly, as represented by the area (xb3). Therefore, those findings support the results obtained regarding robust significance of current account imbalances in the panel logit model estimations. Figure 4: Predictive Ability Testing (Developing Countries) ROC curve comparison To summarize the key results of estimation results, it is viewed that widening current account imbalances for developing countries and credit expansions for developed countries are better indicative of an increasing risk of financial crisis. As a matter of fact, both current account imbalances and high domestic credit extensions raises the probability of financial crisis in both country groups. Credit booms are found to be statistically significant in raising the probability of financial crisis only for developed countries. 0.00 0.25 0.50 0.75 1.00 Sensitivity 0.00 0.25 0.50 0.75 1.00 Specificity xb1 ROC area: 0.6583 xb2 ROC area: 0.5266 xb3 ROC area: 0.6299 Reference 19 2. CONCLUSION There are different explanations about how crises occur. These models have been developed in response to changing characteristics of the crises over time, especially in the 1990s. Every model has been developed in the aftermath of a new crisis in order to explain the dynamics of the crisis and desire to generalize main aspects. However, both theoretical and empirical analysis of the crises in this period in the literature point to different conclusions. One of the main reasons of this non-consensus is that one cannot make generalization as to the macroeconomic conditions under which financial crises have occurred. Empirical analysis obtain different results as regards to the impact of explanatory variables, since there is no single way of measuring the explanatory variables, besides no agreement on which explanatory variables to include. Even, there is no consensus in the literature as regards to the definition of crisis. In this study, factors affecting the probability of financial crisis are evaluated separately for developed and developing countries over the period 1970-2008. Together with integrating the recent global crisis into the analysis, it has been possible to question the impact of credit growth on financial crisis. Furthermore, it is aimed to question the role of current account imbalances in leading financial crisis in terms of developed countries as well in addition to the findings related to developing countries in the literature. To summarize the estimation results, current account imbalances and credit trends have been found to be robust indicators of an increasing risk of financial crisis both in developed and developing countries. On the other side, credit trends in developed countries, current account imbalances in developing countries have been more statistically significant indicators of raising probability of financial crisis. While these are mostly confirmations of previous studies, our estimation results differentiating the role of these macro variables 20 separately in developed and developing countries as well as incorporating the recent financial crisis into the analysis provide insights into the determinants of banking crises in these two different country groups. 21 APPENDIX 1: Name and Definition of Variables and Sources Variable Name Definition Source BANKCRISIS Systemic Banking Crisis Dummy Variable, where 1 indicates a crisis. Laeven ve Valencia (2010) http://www.luclaeven.com/D ata.htm INF Inflation Natural log of (1+ CPI Growth Rate) World Development Indicators (WDI) online database, World Bank CABGDP Current Account Balance (percent of GDP) World Development Indicators (WDI) online database, World Bank GDPPCGR Per capita growth rate (percent) World Development Indicators (WDI) online database, World Bank M2GDP Money Supply (percent of GDP) World Development Indicators (WDI) online database, World Bank BROADMONEY Broad Money Supply (percent of GDP) World Development Indicators (WDI) online database, World Bank GDPGR GDP Growth Rate (percent) World Development Indicators (WDI) online database, World Bank DOMCREDPR Domestic credit given to the private sector (percent of GDP) World Development Indicators (WDI) online database, World Bank REALINTR Real Interest Rate (percent) World Development Indicators (WDI) online database, World Bank 22 APPENDIX - 2: Banking Crises Country Banking Crisis (year of start) Argentina 1980, 1989, 1995, 2001 Austria 2008 Belgium 2008 Brazil 1990, 1994 Czech Republic 1996 China 1998 Denmark 2008 Chile 1976, 1981 Colombia 1982, 1998 Egypt 1980 Finland 1991 France 2008 Germany 2008 Greece 2008 Hungary 1991, 2008 Iceland 2008 India 1993 Indonesia 1997 Ireland 2008 Israel 1977 Japan 1992, 1997 Kazakhstan 2008 Korea 1997 Letonia 1995, 2008 Malaysia 1997 Mexico 1981, 1994 Fas 1980 Netherlands 2008 Norway 1991 Peru 1983 Philippines 1983, 1997 Poland 1992 Portugal 2008 Russia 1998, 2008 Slovenia 1992, 2008 Spain 1977, 2008 Sri Lanka 1989 Sweden 1991, 2008 Switzerland 2008 Thailand 1983, 1997 Turkey 1982, 2000 Ukraine 1998, 2008 United Kingdom 2007 United States 1988, 2007 Venezuela 1994 Zimbabwe 1995 Source: Laeven and Valencia (2010:11, 2012:24-26) 23 REFERENCES Akyüz Y. and A. Cornford. 1999. “Capital Flows to Developing Countries and the Reform of the International Financial System”. UNCTAD Discussion Papers, UNCTAD/OSG/DP/143. Bernanke, B. S., 1983. “Nonmonetary Effects of the Financial Crisis in Propagation of the Great Depression”. American Economic Review, 73(3): 257–76. Bernanke, B. S. 1993. “Credit in the Macroeconomy”. Quarterly Review, Federal Reserve Bank of New York, Spring, p. 50–70. Borio, C. and P. Lowe. 2002. “Asset Prices, Financial and Monetary Stability: Exploring the Nexus”. BIS Working Papers 114, Bank for International Settlements. Borio, C., and P. Lowe. 2003. “Imbalance or Bubbles? Implications for Monetary and Financial Stability”, in Asset Price Bubbles: The Implications for Monetary, Regulatory, and International Policies, edited by W. C. Hunter, G. C. Kaufman, and M. Pomerleano. Cambridge, Mass.: MIT Press, p. 247–70. Calvo, G.A.. 2005. “Crises in Emerging Market Economies-A Global Perspective”. NBER Working Paper, No.11305. Calvo, G.A.. 2007. “Crises in Emerging Market Economies: A Global Perspective”. NBER Working Paper, No:11305, Cambridge, MA: National Bureau of Economic Research Caprio G. and P. Honohan. 2008. “Banking Crises”. IIIS Discussion Paper, No. 242 Institute for International Integration Studies Chang, R. and A. Velasco. 2001. “A model of Financial Crises in Emerging Market. Quarterly Journal of Economics, 116(2): 489-517. Claessens, S., G. Dell’Ariccia, D. Igan and L. Leaven. 2010a. “Cross-Country experiences and policy implications from the global financial crisis”. Economic Policy, p.267-293. Claessens, S., G. Dell’Ariccia, D. Igan and L. Leaven. 2010b. “Lessons and Policy Implications from the Global Financial Crisis”. IMF Working Paper, WP/10/44.