(Luận văn) macroeconomic, financial and institutional determinants of banking crisis, the money market pressure index approach

UNIVERSITY OF ECONOMICS HO CHI MINH CITY VIETNAM INSTITUTE OF SOCIAL STUDIES THE HAGUE THE NETHERLANDS t to ng hi VIETNAM - NETHERLANDS PROGRAMME FOR M.A IN DEVELOPMENT ECONOMICS ep w n lo ad ju y th yi pl MACROECONOMIC, FINANCIAL AND INSTITUTIONAL DETERMINANTS OF BANKING CRISIS: THE MONEY MARKET PRESSURE INDEX APPROACH n ua al n va fu ll A thesis submitted in partial fulfilment of the requirements for the degree of m MASTER OF ARTS IN DEVELOPMENT ECONOMICS oi at nh z z k jm CHAU THE VINH ht vb By om l.c gm n va Assoc Prof NGUYEN TRONG HOAI an Lu Academic Supervisor: ey t re HO CHI MINH CITY, December2014 CERTIFICATION t to ng “I certify that the substance of this thesis has not already been submitted for any hi ep degree and has not been currently submitted for any other degree w I certify that to the best of my knowledge and help received in preparing this thesis n lo and all used sources have acknowledged in this dissertation” ad y th CHAU THE VINH ju yi Date: 31st December 2014 pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re i ACKNOWLEDGEMENT t to Upon completing this thesis, I have received a great deal of encouragement and ng hi support from many people ep First of all, I would like to express my deepest gratitude towards Assoc Prof w n Nguyen Trong Hoai, my esteemed academic supervisor, for his patient guidance, lo ad encouragement and valuable critiques for my research work y th Also, I would like to thank Dr Truong Dang Thuy for his guidance and advice in ju yi econometric techniques, Dr Pham Khanh Nam for his encouragement and valuable pl advice in the starting phase of my thesis research design ua al n My gratefulness is also extended to all of my lecturers and staffs of the Vietnam- va n Netherlands Program for their assistance during my first days in this programme ll fu oi m Besides, I would love to thank my parents and my families for their ceaseless encouragement and support during my study period Moreover, my special thanks nh at to my C.E.O – Mr Nguyen Huu Tram, who understands and gives me approval for z my long personal leave to finalize my thesis on time Without them, I would not z jm ht vb have opportunities and incentives to have my thesis finished k Finally, I would like to thank all my friends and other people who have had any om l.c gm help and support for my thesis but are not above-mentioned an Lu n va ey t re ii ABSTRACT t to The thesis estimates a logit regression model by fixed effect with a combination of ng hi some macroeconomic and financial indicators from the work of Hagen and Ho ep (2007) and Worldwide Governance Indicators (WGI) from the updated database of Kaufmann (2013) as explanatory variables for binary dependent variable banking w n lo crises generated from the approach of money market pressure index (Hagen and Ho, ad 2007) The monthly panel dataset, which is available in full range and easy of y th approach from International Financial Statistics CD-ROM (2011), of 18 countries ju yi from Latin America and Asian over the scope of 2001 – 2010is applied Some pl specific lag lengths of indicators are also applied according to the suggestion of al n ua “flexibility in forecast horizon” of Drehmann et al (2011) va The crisis phenomenon of banking system seems to be well-described in light of the n present of depreciation, former year crisis, high real interest rate in prior of 36 fu ll months, growth of credit to GDP in prior 12 months Moreover, impact of inflation m oi seems to support the school of thought that it is negative effect to crisis nh at Simultaneously, growth rate of bank deposits to GDP is likely useful to prevent z banking systems from profitability risks exposure that leads to banking crisis z ht vb probability However, unfortunately, the indicators of growth of monetary base and jm growth of M2 to reserves give incorrect expected sign and negligible effect on k banking crisis Furthermore, the included institutional variables from WGI give gm insignificant statistic meaning Hence, another set of institutional indicators such as l.c that from International Country Risk Guide (ICRG) should be considered in future om analysis to test for the relationship between Government health and banking crisis an Lu probability likely no proper explanation on the tranquil periods of banking system Hence, it is iii ey the would-be banking crisis determinants Nevertheless, on the other hand, there is t re the future, this thesis attempts to contribute so-called new updates information on n va Despite, on one hand, there should be a more adequate research to be examined in suggested that thereshould be some assessment ofsuch time of banking system, which over a long time has beenneglected (Kauko, 2014) t to ng Key words: banking crisis, tranquiltime, determinants, institutional indicators, fixed hi ep effect logitregression w n lo ad ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re iv TABLE OF CONTENTS t to CHAPTER 1: INTRODUCTION ng hi ep Problem statement 1.2 Research objective 1.3 Research question 1.4 Structure of the thesis w 1.1 n lo CHAPTER 2: LITERATURE REVIEW ad Defining banking crisis 2.2 Trends of banking crises researchtogether with crises mechanism ju y th 2.1 2.2.1 The first trend yi pl 2.2.2 The second trend 10 ua al 2.2.3 The third trend 14 Money Market Pressure (MMP) Index (Hagen and Ho, 2007) 19 2.4 Chapter summary 21 n 2.3 va n CHAPTER 3: METHODOLOGY, MODEL SPECIFICATION AND DATA 28 fu Model selection 28 3.2 Model specification 31 ll 3.1 oi m nh 3.2.1 Macroeconomic indicators 33 at 3.2.2 Financial indicators 34 z 3.2.3 Institutional indicators 36 z Estimation strategies and relevant model diagnostics 40 jm 3.3 ht vb 3.2.4 Use of lagged terms 37 k 3.3.1 Calculation of MMP for banking crisis assessment 40 gm 3.3.2 Model estimation steps and diagnostics 41 Data scope and sources 43 3.5 Conceptual framework 46 3.6 Research Process 47 om l.c 3.4 an Lu CHAPTER 4: RESUTLS AND FINDINGS 48 4.2 Statistical tests for model 51 4.1.3 Test for multicollinearity 51 v ey 4.1.2 Goodness of fit test 51 t re 4.2.1 Model specification test 51 n Descriptive statistics of explanatory indicators 48 va 4.1 4.3 Coefficients interpretation 53 4.3.1 Macroeconomic indicators 53 t to 4.3.2 Financial indicators 55 ng 4.3.3 Institutional indicators 57 hi CHAPTER 5: CONCLUSION, POLICY RECOMMENDATION AND LIMITATION 58 ep 5.1 Conclusion 58 5.2 Policy recommendation 58 w n 5.3 Limitation of the research 60 lo ad REFERENCES 61 y th APPENDICES 65 ju Table 2.1 Summary of literature reviewed 22 yi Figure 2.1 Mechanisms of banking crisis 27 pl Table 3.1 Data for MMP index calculation 44 al ua Table 3.2 Data and sources of explanatory variables 45 n Table 4.1 Banking crisis dates retrieved from MMP index 65 va n Table 4.2 Summary statistics of variables used in the regression 49 ll fu Table 4.3a The correlation on the sample observations 50 oi m Table 4.3b The correlation on the sample observations 50 Table 4.4Linktest for specification error of logit model 66 nh at Table 4.5 Goodness of fit test of model 67 z Tabel 4.6 Full model multicollinearity test result 67 z vb Table 4.7 Dropping significantly high correlated variables GE, RL: 68 jm ht Table 4.8 Dropping high correlated variables GE, RL and CC 68 Table 4.9 Using interactive term of GE and RL 69 k gm Table 4.10 Full model 69 l.c Table 4.11 Restricted model without GE, RL, CC 70 Table 4.12 Fixed effect model with lags 70 om Table 4.13 Random effect model with lags 71 an Lu Table 4.14 Simple logit model with lags 72 Table 4.15Comparison of lagged terms of indicators in simple logit, FEM and REM 73 n va ey t re vi ABBREVIATION t to MMP: Money Market Pressure ng WGI: World Governance Indicator hi ep WB: World Bank IMF: International Monetary Fund w n IFS: International Financial Statistics lo ad ICRG: International Country Risks Guide ju y th FEM: Fixed Effect Model REM: Random Effect Model yi pl BC: Banking Crisis n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re vii CHAPTER 1: INTRODUCTION 1.1 Problem statement t to Banking crisis in nowadays economies is not a new issue or even an old one that ng hi has been given awareness to, discussed and researched from many angles and ep perspectives by applying many approaches from simple to complicate There have w been three trends of banking system crisis researches from its first trend of n lo qualitative description by Friedman and Schwartz (1963) about US crisis over its ad past decades to the second trend in which econometric analysis with panel data were y th ju employed according to relatively enough banking crises observations and to the yi third trend since the 2007 “global financial turmoil” The trends of banking crisis pl ua al research contribute most of important indicators related to macroeconomics and banking sectors such as reserves, current account, real exchange rate (Kaminsky et n n va al, 1998) Despite the fact that the logistic regression approach focused more on ll fu quantitative economics model, it has seemed to be an important tool for anticipating oi m the crisis signals and timing as well as significant indicators However, there was also some noise that could affect the effectiveness of this model Hence, it led to the nh at rise of further studies in terms of developing new method and other new critical z z variables vb As suggested, there have been many criteria to help researchers with banking crisis ht k jm identification Amongst, money market pressure index from the work of Hagen and gm Ho (2007), who expanded the literature of Eichengreen, Rose, and Wyplosz (1995, l.c 1996a, 1996b) for currency crisis, stands out to be convenient for understanding and om data collecting but still provide good judgment value for banking crisis symptom an Lu Such index observed the periods that banking systems experience its liquidity problem by considering simultaneously the phenomenon of both high central bank ey scope analyzed t re index provides the criterion to indicate whether there is a crisis or not under the n va reserves demand and fluctuations of short-term real interest rate Originally, the Banks relevant data, to some extent, seems to be difficult to obtain precisely due to i their sensitiveness Given those difficulties, the research will make use of macroeconomic indicators as suggested in a survey that emphasized “the analysis of t to macroeconomic variables is of some help for banking supervisors in order to fully ng hi assess banks’ health” (Quagliariello, 2008) In accordance with both suggestion ep from Quagliariello (2008) and Hagen and Ho (2007), some available macroeconomic and financial variables such as inflation, growth of monetary base, w n lo depreciation, real interest rate, growth of private credit over GDP, growth of ad deposits over GDP and growth of M2 over reserves are examined In recent years, y th there has been the use of institutional signals (Kaufmann et al, 2008) to predict for ju yi the probability of vulnerability and crisis occurrence besides quantitative economic pl indicators to enhance the limitation of the model by Kaminsky et al (1998) al n ua Moreover, being motivated by the work of Breuer et al (2006) on institutional va variables and currency crisis, this research will take this idea together with the n combination with six updated world governance indicators (Kaufmann, 2013) fu ll namely voice and accountability, government effectiveness, political stability, rule m oi of law, regulatory quality, control of corruption to assess the role of “health” of nh at Government in the relationship with crisis time of the banking systems Last but not z least, the 12-month lagged term of banking crisis included into the regression model z ht vb (Falcetti and Tudela, 2006) also give significant assessment jm Nevertheless, it seems that most of relevant researches tend to try to explain the k reasons for a banking crisis occurrence but not that why banking crisis does not take gm place in some situation over some period in some country The attempt to l.c understand or even forecast the crisis is important on one hand But, on the other om hand, future researches should be carried out with the tranquil time of the banking an Lu system, i.e the “non-crisis” situation, still has its important role which seems to be macroeconomic and financial background in a same model Thus, the contribution ey there are likely few works considering simultaneously the health of Government, t re Although there have been researches and studies on banking crisis, it seems that n va belittled or even no need to be explained (Kauko, 2014) background likely push banking system to the position of exposure to risks and vulnerabilities, which may trigger banking crises t to Firstly, in terms of macroeconomic indicators, the regression outcome indicates that ng hi prior 12-month crisis may be a worrying sign for present banking crisis to occur ep and/or keep sustained on the basic of prior banking crisis Thus, authorities should pay more attention once confronting a banking crisis in the present time to deploy w n lo proper responses from either the standpoint of crisis alleviation or crisis elimination ad to reduce or even avoid the costly consequences of banking crisis may give bad y th impacts to output of the economy that may make banking crisis even worse ju yi Moreover, in terms of financial indicators, high real interest rate, growth of credit pl over GDP, growth of deposit over GDP with their lagged terms and depreciation al n ua seem to be helpful in banking crisis probability predicting These indicators, more va or less, reflect the healthiness of banking system as a whole, thus authorities should n take into account their sudden fluctuations to deploy some appropriate fu ll interventions One important figure to be considered thoroughly of these indicators m oi is that, their impact may not explicit in real time, yet its accumulations over time, nh at which sometimes are likely to be ignored, may lead to high probability of banking z crisis Furthermore, unlike some studies, high inflation in this time of thesis seems z ht vb to be a good sign, however, it may be due to country scope of the thesis analysis jm that emerging countries in which inflation may go along with economic growth k Inflation is still under debate, hence, cautious observations of inflation indicator are gm highly recommended In addition, as a matter of fact, there is likelihood that l.c inflation stability program may cause volatility to real interest rate which, in turn, om contributes to the gain or loss of bank’s profitability Bad profit of banks tends to an Lu put them in a position of highly exposure to banking crisis Hence, inflation control 59 ey (Dermirguc-Kunt and Detragiache, 1998) t re between program effectiveness and the cost of banking crisis once occurs n va program should be designed carefully and cautiously, particular in harmonizing 5.3 Limitation of the research Although there are attempts to have banking crisis determinants indicated, outcomes t to are still not yet perfect The thesis follows the so-called “old-fashioned” analyses ng hi that tend to study the determinants of banking crisis but not that of non-banking- ep crisis which is likely important to explain why some countries are exposure to risks while some are not Some criteria for the non-crisis time, or tranquil time, have to w n lo be argued and introduced for further studies The combination of factors causing ad crisis and those can help maintain non-crisis has to be put to consideration to have a y th more overall view which is critical for authorities in managing the banking system ju yi as well as the economy The thesis uses the banking crisis definition from the pl judging index of money market pressure, the fixed effect model regression from the al n ua work of Hagen and Ho (2007) under the assumption that this usage of the index is va widely known and accepted, however, the availability of relevant variables as well n as the ability of the author to access similar updated data may lead to data collection fu ll bias Moreover, some indicators only shows their significance with some lags, thus, m oi more data seems needed to prevent the model from missing observation which nh at could lead to sample size reduction problem Furthermore on lagged terms usage, z there should be the trial and error processes for all possible lags of all relevant z ht vb indicators, but it seems to be impractical when there are many variables; thus, the jm thesis has to compromise to use lags suggested from previous studies with some k slight subjective changes Hence, there leaves a question on how one could test for gm suitable lags to be applied Unfortunately, all the institutional variables applied in l.c the estimation model give insignificant results As mentioned, this pessimistic om outcome maybe because of inappropriate dataset used, thus, the testing should be an Lu employed with more proper database according to newly empirical studies or at 60 ey related dates of crises have been discussed in this scope of analysis t re studies, there is no evidence found for the end of a banking crisis, only starts and n va least the same database of Breuer et al (2006) from ICRG After all, like almost REFERENCES Aizenman, J., Pasricha, G.K., 2012 Dereminants of financial stress and recovery t to during the great recession.Int J Finance Econ 17, 347–372 ng hi Artha, I.K.D.S., de Haan, J., 2011 Labor market flexibility and the impact of the ep financial crisis.Kyklos 64, 213–230 w Berkmen, S.P., Gelos, G., Rennhack, R., Walsh, J.P., 2012 The global financial n lo crisis: explaining cross-country differences in the output impact J Int ad Money Finance 31, 42–59 y th ju Bordo, M D and Meissner, C.M (2012) Does inequality lead to a financial crisis? yi Journal of International Money and Finance 31: 2147-2161 pl ua al Breuer, J.B, Shimpalee, P.L 2006.Currency crises and institutions Journal of International Money and Finance 25: 125-145 n n va Corsetti, G., Pesenti, P and Roubini, N (1999) Paper tigers? ll fu A model of the Asiancrisis European Economic Review 43: 1211-1236 oi m Chang, R., & Velasco, A (2000) Liquidity crises in emerging markets: theory and policy InNBER Macroeconomics Annual 1999, Volume 14 (pp 11-78) nh at MIT z z Demirgtic-Kunt, A., &Detragiache, E (1998a) The determinants of banking crises vb in developing and developed countries IMF Staff Papers, 45, 81-109 ht gm fragility.IMF Working Paper,No.WP/83/98 k jm Demirgtic-Kunt, A., &Detragiache, E.(1998b) Financial liberalisation and financial l.c Eichengreen, B., & Rose, A K (1998) Staying afloat when the wind shifts: om External factors and emerging-market banking crises (No w6370).National an Lu Bureau of Economic Research Eichengreen, B., Rose, A K., Wyplosz, C., Dumas, B., & Weber, A (1995) banking and currency crises Economica, 75(298), 199-221 61 ey Falcetti, E., &Tudela, M (2008) What twins share? A joint probit estimation of t re attacks Economic policy, 249-312 n va Exchange market mayhem: the antecedents and aftermath of speculative Frankel, J.A., Saravelos, G., 2010 Are Leading Indicators of Financial Crises Useful for Assessing Country Vulnerability? Evidence from the 2008–2009 t to Global Crisis NBER Working Paper Series 16047 ng hi Glick, R., & Hutchison, M (2000) Banking and currency crises: how common are ep the twins? Financial crises in emerging markets.Cambridge University Press w n lo González-Hermosillo, B (1996) Banking sector fragility and systemic sources of ad fragility, IMF Staff Papers, No.WP/96/12 (February) ju y th Gujarati, D., N., 2004, Basic Econometrics, 4th edition, United States: Gary Burke yi Kaminsky, G., Reinhart, C.M., 1998 Financial crisis in Asia and Latin America: pl then and now In: The American Economic Review 88: Papers and al n ua Proceedings of the Hundred and Tenth Annual Meeting of the American va Economic Association pp 444–448 n Kaminsky, G., Lizondo, S., Reinhart, C.M., 1998.Leading indicators of currency fu ll crises.IMF Staff Pap 45, 1–48 m oi Kaminsky, G L., & Reinhart, C M (1999) The twin crises: the causes of banking nh at and balance-of-payments problems American economic review, 473-500 z Kaufmann, D., Kraay, A., &Mastruzzi, M (2007).Governance Matters VII: z ht vb Aggregate and Individual Governance Indicators, 1996-2007 World Bank, k Group, Macroeconomics and Growth Team jm World Bank Institute, Global Programs Division, and Development Research gm Klomp, J and de Haan, J (2009) Central bank independence and financial om l.c instability.Journal of Financial Stability 5: 321-338 Kokko, A (1999) The Asian crisis, many similarities with the Swedish an Lu crisis.EkonomiskDebatt 27: 81-92 literature Economic Systems 38: 289–308 62 ey Kauko, K (2014) How to foresee banking crises? A survey of theempirical t re financialcrisis Economics Letters 115: 196-199 n va Kauko, K (2012) External deficits and non-performing loans in the recent McKinnon, R I., & Pill, H (1996) Credible liberalizations and international capital flows: the “over-borrowing syndrome” Financial Deregulation and t to Integration in East Asia, NBER-EASE Volume (pp 7-50) University of ng hi Chicago Press ep McKinnon, R I., & Pill, H (1998) The overborrowing syndrome: are East Asian economies different? Managing Capital Flows and Exchange Rates: w n lo Perspectives from the Pacific Basin, 322-55 ad Miller, V (1999).The timing and size of bank-financed speculative attacks.Journal y th of International Money and Finance, 18(3), 459-470 ju yi Mishkin, F S (1997) Understanding financial crises: a developing country pl perspective (No w5600) National Bureau of Economic Research al va banking n ua Noy, I (2004).Financial liberalization, prudential supervision and the onset of n crises Emerging Markets Review 5: 341-359 fu ll Obstfeld, M (2012).Does the current account still matter? American Economic oi m Review102: 1-23 nh at Obstfeld, M (1995) The logic of currency crises (pp 62-90).Springer Berlin z Heidelberg z ht vb Reinhart, C M., &Végh, C A (1995) Nominal interest rates, consumption booms, jm and lack of credibility: A quantitative examination Journal of Development k Economics, 46(2), 357-378 gm Rojas-Suárez, L., &Weisbrod, S R (1995).Financial fragilities in Latin America: om l.c the 1980s and 1990s (Vol 132) IMF Rose, A.K and Spiegel, M.M (2011) Cross-country causes and consequences of an Lu thecrisis: an update European Economic Review 55: 309-324 ey t re the2008 crisis: early warning Japan and the World Economy 24: 1-16 n va Rose, A.K and Spiegel, M.M (2012) Cross-country causes and consequences of 63 Rossi, M (1999) Financial Fragility and Economic Performance in Developing Economies-Do Capital Controls Prudential Regulation and Supervision t to Matter?IMFWorking Paper, No WP/99/66 (May) ng hi Velasco, A (1987) Financial crises and balance of payments crises: a simple model ep of the Southern Cone experience Journal of development Economics,27(1), 263-283 w n lo Von Hagen, J., & HO, T K (2007).Money market pressure and the determinants of ad banking crises Journal of Money, Credit and Banking, 39(5), 1037-1066 ju y th yi pl n ua al n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 64 APPENDICES Table 4.1 Banking crisis dates retrieved from MMP index t to Argentina 2001M3 Costa Rica 2005M1 ng hi ep w n lo 2008M8 2006M1 2001M8 2008M12 2007M1 2001M11 2009M7 2008M1 2002M4 2010M3 2009M1 2004M5 2010M5 2009M11 2004M10 2010M1 ad Indonesia yi 2006M2 2006M4 2004M2 2006M10 2007M3 2007M8 2007M12 2008M9 2008M5 2008M4 2008M11 2008M8 2008M7 2009M2 2010M9 pl 2005M4 al ju y th 2002M12 n ua va 2008M12 ll 2002M11 m 2002M9 Thailand 2002M8 2003M1 2003M7 2003M11 2004M3 2005M11 2004M6 2007M7 2004M9 2008M11 2009M5 oi 2003M3 Philippines 2001M11 fu 2001M11 Japan n Brunei Paraguay 2001M7 2005M4 Brazil 2008M6 nh 2003M7 at 2004M1 z 2006M9 2010M8 Singapore 2001M3 z 2002M9 2009M2 2006M12 2009M4 2007M5 2010M10 2002M6 2006M12 2008M12 2009M6 Uruguay om Malaysia 2002M12 l.c 2008M11 gm 2002M6 2002M3 k 2005M10 2001M12 jm 2001M4 2008M10 2005M10 2008M1 2007M10 2003M8 2009M1 2008M7 2004M3 2010M4 2008M10 2007M6 2010M9 2009M5 2002M2 2009M11 2007M12 2008M2 Mexico 2003M4 65 Vietnam 2002M3 ey 2002M1 t re 2006M3 n 2006M4 va 2010M6 an Lu China 2003M4 ht Chile Korea vb 2008M12 Colombia t to ng hi 2010M4 2004M12 2003M3 2003M5 2005M12 2004M3 2005M4 2006M4 2005M3 2005M12 2008M10 2006M12 2007M5 2010M12 2008M4 ep 2007M6 2008M6 2007M8 w 2010M3 n lo ad Table 4.4Linktest for specification error of logit model y th logit BC_MMP MB_grDepreR_iCrdGDP_GR DEPOGDP_GR M2RES_GR VA PS ju GE RQ RL CC INF l12 BC_MMP yi Logistic regression pl = 1943 LR chi2(14) = 33.30 Pseudo R2 = 0.0389 = 0.0026 ua al Prob> chi2 Number of obs Log likelihood = -411.64834 n -Coef va BC_MMP | Std Err z P>|z| [95% Conf Interval] n -+ fu INF | -2.58 0.010 -.5749169 -.0783745 2891568 4.28 0.000 6702991 1.803773 1.52 0.130 -.0074561 0583348 2.48 0.013 0143107 123205 0.58 -.0214654 0394781 0004054 0448159 1266713 1.237036 MB_gr | 0254394 0167837 Depre| 0687579 0277797 R_i| 0090064 0155471 ll -.3266457 | oi m BC_MMP(12) CrdGDP_GR | 0226107 0113294 2.00 0.046 DEPOGDP_GR | -.0241729 0133022 -1.82 M2RES_GR | -.0053557 0051703 -1.04 0.300 VA | -.0727918 1784815 -0.41 0.683 PS | -.0436996 1796811 -0.24 0.808 GE | -.3891117 4795028 -0.81 0.417 RQ | 009629 3931909 0.02 0.980 -.761011 780269 RL | 1470301 5874459 0.25 0.802 -1.004343 at nh 0.562 z 0.069 z 001899 -.0154893 0047778 -.422609 2770254 -.3958681 308469 -1.32892 5506965 k jm ht vb -.0502448 2175405 369696 0.59 0.556 -.5070503 9421313 _cons | -2.842534 2100018 -13.54 0.000 -3.25413 -2.430938 om l.c CC | gm 1.298403 -linktest = = 1943 LR chi2(2) = 34.43 Pseudo R2 = 0.0402 0.0000 Log likelihood = -411.08473 Coef Std Err z P>|z| [95% Conf Interval] _hat | 1.951366 891996 2.19 0.029 2030855 3.699646 _hatsq | 1986192 1823874 1.09 0.276 -.1588537 556092 66 ey BC_MMP | -+ t re n Number of obs va Prob> chi2 an Lu Logistic regression _cons | 1.050287 1.066478 0.98 0.325 -1.039971 3.140545 t to Table 4.5 Goodness of fit test of model ng lfit, group(10) table hi ep Logistic model for BC_MMP, GOODNESS-OF-FIT TEST (Table collapsed on quantiles of estimated probabilities) + + Prob | Obs_1 | Exp_1 | Obs_0 | Exp_0 | Total | w | Group | n | -+ + -+ -+ -+ -+ -| lo | | 0.0324 | | 5.0 | 190 | 190.0 | 195 | | 0.0372 | | 6.8 | 186 | 187.2 | 194 | ad | y th | 0.0414 | | 7.7 | 189 | 186.3 | 194 | | | 0.0450 | 10 | 8.4 | 185 | 186.6 | 195 | | | 0.0484 | | 9.0 | 186 | 185.0 | 194 | ju | yi pl | -+ + -+ -+ -+ -+ -| | 0.0921 | | 10 | 0.3826 | 188 | 184.2 | 194 | 180 | 184.2 | 195 | 11 | 12.1 | 183 | 181.9 | 194 | 15 | 14.7 | 179 | 179.3 | 194 | 29 | 27.7 | 165 | 166.3 | 194 | n | 0.0670 | | 9.8 | 10.8 | va | | 15 | n | 0.0583 | ua | 0.0527 | | al | number of observations = 10 5.05 0.7524 at nh Prob> chi2 = oi Hosmer-Lemeshowchi2(8) = m number of groups = 1943 ll fu + + z z Tabel 4.6 Full model multicollinearity test result VIF SQRT VIF Tolerance vb Variable R-Squared 0.1500 0.9964 0.0036 MB_gr 1.01 1.00 0.9922 0.0078 Depre 1.12 1.06 0.8911 0.1089 R_i 1.70 1.30 0.5897 0.4103 1.55 1.25 0.6433 0.3567 VA 1.92 1.39 0.5201 0.4799 PS 3.11 1.76 0.3217 0.6783 GE 12.34 3.51 0.0811 0.9189 RL 21.13 4.60 0.0473 0.9527 CC 10.68 3.27 0.0936 0.9064 RQ 8.44 2.91 0.1185 0.8815 -Mean VIF 4.91 67 ey 0.4013 M2RES_GR t re 0.4556 0.5987 n 0.5444 1.29 va 1.36 1.67 an Lu 1.84 om CrdGDP_GR DEPOGDP_GR l.c 0.8500 1.00 gm 1.08 1.00 k 1.18 jm INF BC_MMP12 ht Table 4.7 Dropping significantly high correlated variables GE, RL: Variable VIF SQRT VIF Tolerance R-Squared t to ng 1.17 1.08 0.8521 0.1479 1.00 1.00 0.9970 0.0030 MB_gr 1.01 1.00 0.9924 0.0076 Depre 1.12 1.06 0.8925 0.1075 R_i 1.60 1.26 0.6258 0.3742 CrdGDP_GR 1.81 1.34 0.5540 0.4460 1.64 1.28 0.6090 0.3910 hi INF BC_MMP12 ep w 1.53 1.24 0.6556 0.3444 VA 1.53 1.24 0.6535 0.3465 PS 2.20 1.48 0.4548 0.5452 CC 8.26 2.87 0.1210 0.8790 RQ 5.93 2.44 0.1685 0.8315 lo M2RES_GR y th n DEPOGDP_GR ad ju yi -2.40 pl Mean VIF ua al Table 4.8 Dropping high correlated variables GE, RL and CC SQRT VIF Tolerance R-Squared va VIF n Variable 1.17 1.08 0.8573 0.1427 1.00 1.00 0.9972 0.0028 MB_gr 1.01 1.00 0.9932 0.0068 Depre 1.12 1.06 0.8933 fu INF BC_MMP12 ll n DEPOGDP_GR 1.58 1.26 0.6319 0.3718 0.4281 0.3681 z 0.6282 0.5719 at 1.26 1.32 nh 1.59 1.75 oi m R_i CrdGDP_GR 0.1067 1.52 1.23 0.6588 0.3412 VA 1.31 1.14 0.7657 0.2343 PS 1.73 1.31 0.5785 0.4215 RQ 2.05 1.43 0.4883 0.5117 z M2RES_GR k om l.c gm 1.44 jm Mean VIF ht vb an Lu n va ey t re 68 Table 4.9 Using interactive term of GE and RL Variable VIF SQRT VIF Tolerance R-Squared t to ng 1.18 1.08 0.8506 0.1494 1.00 1.00 0.9970 0.0030 MB_gr 1.01 1.00 0.9924 0.0076 Depre 1.12 1.06 0.8919 0.1081 R_i 1.63 1.28 0.6139 0.3861 CrdGDP_GR 1.81 1.34 0.5537 0.4463 1.64 1.28 0.6087 0.3913 hi INF BC_MMP12 ep w 1.53 1.24 0.6536 0.3464 VA 1.61 1.27 0.6211 0.3789 PS 2.23 1.49 0.4482 0.5518 GERL 2.51 1.58 0.3984 0.6016 CC 9.97 3.16 0.1003 0.8997 RQ 5.95 2.44 0.1680 0.8320 lo M2RES_GR y th n DEPOGDP_GR ad ju yi pl -2.55 n ua al Mean VIF n logit va Table 4.10 Full model BC_MMP INF l12 BC_MMP MB_grDepreR_iCrdGDP_GR DEPOGDP_GR M2RES_GR VA PS fu GE RQ RL CC ll 0.0026 nh = oi Prob> chi2 m Logistic regression at Log likelihood = -411.64834 Number of obs = 1943 LR chi2(14) = 33.30 Pseudo R2 = 0.0389 z Coef Std Err z P>|z| vb BC_MMP | z -[95% Conf Interval] -.3266457 1266713 -2.58 0.010 | 1.237036 2891568 4.28 0.000 MB_gr | 0254394 0167837 1.52 0.130 Depre | 0687579 0277797 2.48 0.013 0143107 123205 R_i | 0090064 0155471 0.58 0.562 -.0214654 0394781 CrdGDP_GR | 0226107 0113294 2.00 0.046 0004054 0448159 DEPOGDP_GR | -.0241729 0133022 -1.82 0.069 -.0502448 M2RES_GR | -.0053557 0051703 -1.04 0.300 -.0154893 VA | -.0727918 1784815 -0.41 0.683 -.422609 PS | -.0436996 1796811 -0.24 0.808 -.3958681 308469 GE | -.3891117 4795028 -0.81 0.417 -1.32892 5506965 RQ | 009629 3931909 0.02 0.980 -.761011 780269 RL | 1470301 5874459 0.25 0.802 -1.004343 1.298403 CC | 2175405 369696 0.59 0.556 -.5070503 9421313 _cons | -2.842534 2100018 -13.54 0.000 -3.25413 -2.430938 BC_MMPL12 -.5749169 jm INF | ht -+ -.6702991 k 1.803773 0583348 om l.c gm -.0074561 -.0783745 001899 an Lu 0047778 2770254 n ey t re 69 va Table 4.11 Restricted model without GE, RL, CC logit BC_MMP INF BC_MMP12 MB_grDepreR_iCrdGDP_GR DEPOGDP_GR M2RES_GR VA PS RQ t to Logistic regression ng hi Prob> chi2 = ep Number of obs = 1943 LR chi2(11) = 32.45 Pseudo R2 = 0.0379 0.0006 Log likelihood = -412.07303 w n BC_MMP | Coef Std Err z P>|z| [95% Conf Interval] lo -+ .3219181 1268047 -2.54 0.011 -.5704507 -.0733856 1.246891 2886808 4.32 0.000 6810872 1.812695 MB_gr | 0257815 0167912 1.54 0.125 -.0071286 0586917 Depre | 0679522 027654 2.46 0.014 0137514 122153 R_i | 0100531 0149234 0.67 0.501 -.0191962 0393024 ad INF | BC_MMP12 | ju y th yi 0233479 0109173 2.14 0.032 0019503 0447454 -.0228573 -1.78 0.075 -.0480517 002337 M2RES_GR | al 0128545 -.0054279 0051016 -1.06 0.287 -.0154269 0045711 VA | -.0018056 1476724 -0.01 0.990 -.2912381 2876269 PS | -.0071222 1360878 -0.05 0.958 -.2738494 2596051 RQ | -.0206309 _cons | -2.948038 pl CrdGDP_GR | DEPOGDP_GR | n ua va -0.10 0.918 -.412779 3715171 1731015 -17.03 0.000 -3.287311 -2.608766 n 2000792 fu ll oi m BC_MMP INF l12.BC_MMP at xtlogit nh Table 4.12 Fixed effect model with lags MB_grDepre l36.R_i l12.CrdGDP_GR z z l6.DEPOGDP_GR M2RES_GR VA PS RQ, fe vb Number of obs = 1511 Group variable: country1 Number of groups = 18 = 48.88 max = 84 =-279.62085Prob> chi2 = 0.0000 om l.c LR chi2(11) Log likelihood gm 83.9 83 k avg = jm Obs per group: = ht Conditional fixed-effects logistic regression -| Coef Std Err z P>|z| [95% Conf Interval] an Lu BC_MMP -+ 2.52 0.012 -.7695856 -.0957662 3351207 3.42 0.001 4880346 1.801683 MB_gr | 0448394 0189506 2.37 0.018 007697 0819818 Depre | 0983399 0376113 2.61 0.009 0246231 1720567 R_i(36) | 0279047 0110093 2.53 0.011 0063268 0494825 CrdGDP_G(12) | 0370456 012508 2.96 0.003 0125303 061561 DEPOGDP_GR(L6)| -.0354308 0153012 -2.32 0.021 -.0654207 -.005441 0004024 0073081 0.06 0.956 -.0139212 0147261 M2RES_GR | 70 ey 1718959 1.144859 t re -.4326759 | n | va INF BC_MMP12 VA | 1.95909 PS | -.9584184 RQ | 1.370052 -.8472151 1.43 0.153 -.7261627 8581543 -1.12 0.264 -2.64037 4.644342 7235331 1.309447 -0.65 0.518 -3.413683 1.719253 t to ng hi Table 4.13 Random effect model with lags ep xtlogit BC_MMP INF l12.BC_MMP MB_grDepre l36.R_i l12.CrdGDP_GR l6.DEPOGDP_GR M2RES_GR VA PS RQ, re w Random-effects logistic regression n Group variable: country1 lo ad Random effects u_i ~ Gaussian avg = = 1511 Number of groups = 18 Obs per group: = 83 83.9 y th max = Number of obs 84 Wald chi2(11) ju =-313.92006Prob> chi2 = yi Log likelihood = 48.62 0.0000 BC_MMP | pl -Coef Std Err z P>|z| [95% Conf Interval] al ua -+ .413111 1.159605 1467951 -2.81 0.005 -.7008241 -.125398 3339441 3.47 0.001 5050869 1.814124 MB_gr | 0441357 0189147 2.33 0.020 0070636 0812078 Depre | 0970619 0374037 2.59 0.009 0237521 1703717 R_i(36)| 0262877 0097476 2.70 0.007 0071827 0453927 CrdGDP_GR(12)| 0395713 0103906 3.81 DEPOGDP_GR(6)| -.0328716 014082 M2RES_GR | 0033222 0067812 n INF | BC_MMP12| n va ll fu 0192061 0599365 0.020 -.0604719 -.0052713 0.49 0.624 -.0099686 016613 oi m 0.000 -2.33 1821686 0.63 0.526 -.2414846 472603 -.0908372 1507151 -0.60 0.547 -.3862334 204559 RQ | -.0532946 2346824 -0.23 0.820 -.5132638 4066745 _cons | -3.156307 2033296 -15.52 0.000 -3.554826 -2.757788 at 1155592 PS | z nh VA | z vb -14.30817 43.4502 -99.46899 jm /lnsig2u | ht -+ -70.85265 k -+ -.0007817 0169817 2.52e-22 2.43e+15 rho | 1.86e-07 8.07e-06 1.92e-44 l.c gm sigma_u | -0.00 Prob>= chibar2 = 1.000 om Likelihood-ratio test of rho=0: chibar2(01) = an Lu n va ey t re 71 Table 4.14 Simple logit model with lags logit BC_MMP INF l12.BC_MMP MB_grDepre l36.R_i l12.CrdGDP_GR l6.DEPOGDP_GR M2RES_GR VA PS RQ t to Logistic regression ng hi Prob> chi2 = ep Number of obs = 1511 LR chi2(11) = 48.89 Pseudo R2 = 0.0722 0.0000 Log likelihood = -313.92003 -BC_MMP | Coef Std Err z P>|z| [95% Conf Interval] w n -+ .413118 1468117 -2.81 0.005 -.7008636 -.1253723 1.1596 333979 3.47 0.001 505013 1.814187 MB_gr | 0441349 0189171 2.33 0.020 007058 0812118 lo INF | BC_MMP12 | ad y th 0970666 0374072 2.59 0.009 02375 1703833 0262878 0097487 2.70 0.007 0071807 0453949 CrdGDP_GR(12)| 0395718 0103918 3.81 0.000 0192042 0599394 DEPOGDP_GR(6)| -.0328699 0140837 -2.33 0.020 -.0604735 -.0052662 M2RES_GR | 0033208 0067821 0.49 0.624 -.0099718 0166134 ju Depre | R_i(36)| yi pl al 1155499 1821923 0.63 0.526 -.2415405 4726403 PS | -.0908037 1507346 -0.60 0.547 -.3862381 2046307 RQ | -.0533487 2347126 -0.23 0.820 -.5133769 4066795 _cons | -3.1566 203355 -15.52 0.000 -3.555168 -2.758031 n ua VA | n va ll fu oi m at nh z z k jm ht vb om l.c gm an Lu n va ey t re 72 Table 4.15 Comparison of lagged terms of indicators in simple logit, FEM and REM t to ng Variables Simple logit model hi INF Fixed effect model ep -0.413118 0.005*** -0.4326759 1.1596 0.001*** BC_MMP12 Random effect model 0.012** -0.413111 0.005*** 1.144859 0.001*** 1.159605 0.001*** 0.0448394 0.018** 0.0441357 0.020** Depre 0.0970666 0.009*** 0.0983399 0.009*** 0.0970619 0.009*** 0.0262878 0.007*** 0.0279047 0.011** 0.0262877 0.007*** 0.0395718 0.000*** 0.0370456 0.003*** 0.0395713 0.000*** DEPOGDP_GR(6) -0.0328699 0.020** -0.0354308 0.021** -.0328716 0.020** 0.0033208 0.624 0.0004024 0.956 0.0033222 0.624 0.1155499 0.526 1.95909 0.153 0.1155592 0.526 -0.0908037 0.547 -0.9584184 0.264 -0.0908372 0.547 0.820 -0.8472151 0.518 -0.0532946 0.820 n 0.020** ju w 0.0441349 lo MB_gr R_i(36) ad CrdGDP_GR(12) y th M2RES_GR yi 0.000*** -3.156307 n 1511 0.000*** 1511 1511 121 121 -279.62085 -313.92006 ll fu observations -3.1566 va Number of -0.0533487 n _constant ua RQ al PS pl VA 48.88 at - - - 48.62 0.000 z Wald chi2(11) 48.89 nh LR chi2(11) -313.92003 oi Log likelihood 121 m Number of crises 0.000 0.000 Pseudo R2 0.072 - z Prob> chi2 vb - ht jm Note: k - One, two, and three asterisks show significance levels of less than 10%, less than 5%, and less than gm 1% respectively om l.c an Lu n va ey t re 73