{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T10:49:58Z","timestamp":1792061398315,"version":"4.4.0"},"posted":{"date-parts":[[2023]]},"group-title":"SSRN","reference-count":107,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2023,1,1]],"date-time":"2023-01-01T00:00:00Z","timestamp":1672531200000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/www.uspto.gov\/ip-policy\/copyright-policy\/copyright-basics"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>&lt;p&gt;&lt;span&gt;To examine if one machine learning model can consistently elucidate financial vulnerabilities, both over time and across levels of development, this paper applies 13 machine learning algorithms to evaluate comparative forecasting performance across several banking crises&lt;\/span&gt;&lt;span&gt;.&amp;nbsp;&lt;\/span&gt;&lt;span&gt;The study concurrently contributes to the literature on early warning signals in general, and modeling frameworks in particular.&amp;nbsp;&lt;\/span&gt;&lt;span&gt;F&lt;\/span&gt;&lt;span&gt;our decades of banking crises are appraised through a vector of 12 leading indicators&lt;\/span&gt;&lt;span&gt;&amp;nbsp;&lt;\/span&gt;&lt;span&gt;encompassing real, banking and external sectors,&lt;\/span&gt;&lt;span&gt;&amp;nbsp;&lt;\/span&gt;&lt;span&gt;and accounting for a representative sample of 19 emerging markets and developed economies. Through quarterly time series supporting expedient policy responses, findings suggest that bank deposits, capital output ratio,&amp;nbsp;&lt;\/span&gt;&lt;span&gt;exchange rates, gross domestic product, consumption expenditure and interest rates represent the most prominent leading indicators. Evaluated by applying root mean squared error, mean absolute error, and receiver operating characteristics with area under curve estimates, random forests exhibit highest predictive strength in minimising the prediction error across a panel format, individual country format and out-of-time format, and outperform all models in recursively predicting banking crises out-of-sample and out-of-time. On aggregate, one algorithm consistently fits most crises, which is advantageous for a global policymaker, however results from the individual country dimension highlight a more nuanced approach to algorithmic selection.&amp;nbsp;&lt;\/span&gt;&lt;\/p&gt;<\/jats:p>","DOI":"10.2139\/ssrn.4463667","type":"posted-content","created":{"date-parts":[[2023,5,30]],"date-time":"2023-05-30T07:20:05Z","timestamp":1685431205000},"source":"Crossref","is-referenced-by-count":0,"title":["A Horse Race of Machine Learning Methods to Predict Banking Crises&lt;b&gt;&lt;\/b&gt;"],"prefix":"10.2139","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-9519-5829","authenticated-orcid":true,"given":"Emile","family":"du Plessis","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","reference":[{"key":"ref1","year":"1984","journal-title":"IFS Code Country Crisis Dates 111 United States"},{"key":"ref2","author":"Denmark","year":"1987"},{"key":"ref3","year":"1994","journal-title":"132 France"},{"key":"ref4","year":"1990","journal-title":"136 Italy"},{"key":"ref5","year":"1991","journal-title":"144 Sweden"},{"key":"ref6","year":"1977","journal-title":"199 South Africa"},{"key":"ref7","year":"1987","journal-title":"238 Costa Rica"},{"key":"ref8","year":"1981","journal-title":"273 Mexico"},{"key":"ref9","year":"1985","journal-title":"542 Korea Republic"},{"key":"ref10","year":"1981","journal-title":"566 Philippines"},{"key":"ref11","author":"Poland","year":"1991"},{"key":"ref12","author":"A Abiad","year":"2003","journal-title":"Early-Warning Systems: A Survey and a Regime-Switching Approach"},{"key":"ref13","author":"D Aikman","year":"2009","journal-title":"Funding Liquidity Risk in a Quantitative Model of Systemic Stability"},{"key":"ref14","doi-asserted-by":"crossref","first-page":"520","DOI":"10.1016\/j.ejpoleco.2011.01.003","article-title":"Quasi Real Time Early Warning Indicators for Costly Asset Price Boom\/Bust Cycles: A Role for Global Liquidity","volume":"27","author":"L Alessi","year":"2011","journal-title":"European Journal of Political Economy"},{"key":"ref15","author":"L Alessi","year":"1723","journal-title":"Identifying Excessive Credit Growth and Leverage"},{"key":"ref16","author":"L Alessi","year":"2015","journal-title":"Comparing Different Early Warning Systems: Results from a"},{"key":"ref17","author":"M Andersson","year":"1999","journal-title":"Managing and preventing financial crises -Lessons from the Swedish experience. Documents 10117. Bank for Internation Settlements. Bank of England"},{"key":"ref18","doi-asserted-by":"crossref","first-page":"2255","DOI":"10.1016\/j.jbankfin.2010.02.015","article-title":"Bank Regulation, Property Prices and Early Warning Systems for Banking Crises in OECD Countries","volume":"34","author":"R Barrell","year":"2010","journal-title":"Journal of Banking and Finance"},{"key":"ref19","doi-asserted-by":"crossref","DOI":"10.3386\/w14760","author":"R J Barro","year":"2009","journal-title":"Stock-Market Crashes and Depressions"},{"key":"ref20","doi-asserted-by":"crossref","DOI":"10.2307\/30035972","article-title":"Assessing Early Warning Systems: How Have They Worked in Practice?","volume":"52","author":"A Berg","year":"2005","journal-title":"IMF Staff Papers"},{"issue":"98","key":"ref21","article-title":"Are Currency Crises Predictable? A Test","author":"A Berg","year":"1998","journal-title":"IMF Working Paper"},{"key":"ref22","article-title":"An evaluation of early warning models for systemic banking crises: Does machine learning improve predictions","volume":"48","author":"J Beutel","year":"2018"},{"key":"ref23","article-title":"The Method of probits","volume":"79","author":"C I Bliss","year":"1934","journal-title":"Science"},{"key":"ref24","doi-asserted-by":"crossref","DOI":"10.1111\/j.1744-7348.1935.tb07713.x","article-title":"The calculation of the dosage-mortality curve","volume":"22","author":"C I Bliss","year":"1935","journal-title":"Annals of Applied Biology"},{"key":"ref25","article-title":"Credit growth, the yield curve and financial crisis prediction: evidence from a machine learning approach","volume":"848","author":"K Bluwstein","year":"2020"},{"issue":"114","key":"ref26","article-title":"Asset prices, financial and monetary stability: Exploring the nexus","author":"C Borio","year":"2002"},{"key":"ref27","doi-asserted-by":"crossref","DOI":"10.1145\/130385.130401","article-title":"A training algorithm for optimal margin classifiers","author":"B E Boser","year":"1992","journal-title":"Proceedings of the Fifth Annual Workshop on Computational Learning Theory, COLT"},{"key":"ref28","author":"L Breiman","year":"1984","journal-title":"Classification and Regression Trees"},{"key":"ref29","volume":"45","author":"L Breiman","year":"2001","journal-title":"Random forests, Machine Learning"},{"key":"ref30","article-title":"The global financial crisis. Why didn't anybody notice?","author":"British Academy","year":"2009","journal-title":"British Academy Forum. British Academy"},{"key":"ref31","article-title":"Master machine learning algorithms: Discover how they work and implement them from scratch","author":"J Brownlee","year":"2016","journal-title":"Machine Learning Mastery"},{"key":"ref32","doi-asserted-by":"crossref","first-page":"953","DOI":"10.1016\/j.jimonfin.2006.07.007","article-title":"Towards a new early warning system of financial crises","volume":"25","author":"M Bussi\u00e8re","year":"2006","journal-title":"Journal of International Money and Finance"},{"key":"ref33","article-title":"An early warning system for banking crises: From regression-based analysis to machine learning techniques","volume":"235","author":"E J Casabianca","year":"2019","journal-title":"Dipartimento di Scienze Economiche \"Marco Fanno"},{"key":"ref34","volume":"54","author":"M Chamon","year":"2007","journal-title":"Can we predict the next capital account crisis? IMF Staff Papers"},{"key":"ref35","first-page":"22","article-title":"MUC-4 evaluation metrics","author":"N Chinchor","year":"1992","journal-title":"Proc. of the Fourth Message Understanding Conference"},{"key":"ref36","doi-asserted-by":"crossref","first-page":"249","DOI":"10.1016\/S0261-5606(00)00048-6","article-title":"Unit root tests for panel data","volume":"20","author":"I Choi","year":"2001","journal-title":"Journal of International Money and Finance"},{"issue":"1","key":"ref37","doi-asserted-by":"crossref","DOI":"10.1109\/TIT.1967.1053964","article-title":"Nearest neighbor pattern classification","volume":"13","author":"T Cover","year":"1967","journal-title":"IEEE Transactions on Information Theory"},{"key":"ref38","author":"J S Cramer","year":"2002","journal-title":"The origins of logistic regression"},{"issue":"30","key":"ref39","article-title":"Anatomy of banking crises in developing and emerging market countries","author":"R Dattagupta","year":"2011","journal-title":"Journal of International Money and Finance"},{"key":"ref40","doi-asserted-by":"crossref","DOI":"10.1016\/j.jfs.2007.12.004","article-title":"Comparing early warning systems for banking crises","volume":"4","author":"E P Davis","year":"2008","journal-title":"Journal of Financial Stability"},{"key":"ref41","article-title":"Could early warning systems have helped to predict the sub-prime crisis?","author":"E P Davis","year":"2008","journal-title":"National Institute Economic Review"},{"key":"ref42","article-title":"Should multivariate early warning systems for banking crises pool across regions?","author":"E P Davis","year":"2011","journal-title":"Review of World Economics"},{"key":"ref43","first-page":"10","author":"G Dell'ariccia","year":"2010","journal-title":"Monetary policy, leverage and bank risk taking"},{"key":"ref44","doi-asserted-by":"crossref","first-page":"837","DOI":"10.2307\/2531595","article-title":"Comparing the areas under two or more correlated receiver operating characteristic curves: A nonparametric approach","volume":"44","author":"E R Delong","year":"1988","journal-title":"Biometrics"},{"key":"ref45","article-title":"Monitoring banking sector fragility: A multivariate logit approach","volume":"14","author":"` G De Nicolo","year":"2000","journal-title":"The World Bank Economic Review"},{"key":"ref46","author":"A Demirg\u00fc\u00e7-Kunt","year":"2005","journal-title":"Cross-country empirical studies of systemic bank distress: A Survey"},{"key":"ref47","doi-asserted-by":"crossref","first-page":"745","DOI":"10.1016\/j.ijforecast.2017.02.003","article-title":"Predicting recessions with boosted regression trees","volume":"33","author":"J D\u00f6pke","year":"2017","journal-title":"International Journal of Forecasting"},{"key":"ref48","author":"M J Dueker","year":"2001","journal-title":"Forecasting qualitative variables with vector autoregressions: A qual VAR model of U.S. recessions. Federal Reserve Bank of St. Louis, Working Paper"},{"key":"ref49","volume":"46","author":"E Du Plessis","year":"2022","journal-title":"Multinomial Modeling Methods: Predicting Four Decades of International Banking Crises"},{"issue":"1","key":"ref50","doi-asserted-by":"crossref","first-page":"477","DOI":"10.1080\/15140326.2020.1816132","article-title":"Dynamic Forecasting of Banking Crises with a Qual VAR","volume":"25","author":"E Du Plessis","year":"2022","journal-title":"Journal of Applied Economics"},{"key":"ref51","author":"E Du Plessis","year":"2023","journal-title":"Replication data for: A Horse Race of Machine Learning Methods to Predict Banking Crises"},{"issue":"1","key":"ref52","doi-asserted-by":"crossref","DOI":"10.1002\/for.3184","article-title":"New forecasting methods for an old problem: Predicting 147 years of systemic financial crises","volume":"44","author":"E Du Plessis","year":"2025","journal-title":"Journal of Forecasting"},{"key":"ref53","article-title":"Using market information for banking system risk assessment","volume":"2","author":"H Elsinger","year":"2006","journal-title":"International Journal of Central Banking"},{"issue":"2","key":"ref54","doi-asserted-by":"crossref","first-page":"293","DOI":"10.1016\/S1514-0326(16)30012-5","article-title":"Qual VAR revisited: Good forecast, bad story","volume":"19","author":"M El-Shagi","year":"2016","journal-title":"Journal of Applied Economics"},{"issue":"8","key":"ref55","doi-asserted-by":"crossref","first-page":"861","DOI":"10.1016\/j.patrec.2005.10.010","article-title":"An introduction to ROC analysis","volume":"27","author":"T Fawcett","year":"2006","journal-title":"Pattern Recognition Letters"},{"key":"ref56","volume":"11","author":"Fdic","year":"1998","journal-title":"The Banking Crises of the 1980s and Early 1990s. FDIC Banking Review"},{"key":"ref57","author":"G T Fechner","year":"1860","journal-title":"Elemente der Psychophysik"},{"issue":"3","key":"ref58","doi-asserted-by":"crossref","DOI":"10.1137\/0109031","article-title":"An Application of the Wiener-Kolmogorov Smoothing Theory to Matrix Inversion","volume":"9","author":"M Foster","year":"1961","journal-title":"Journal of the Society for Industrial and Applied Mathematics"},{"key":"ref59","article-title":"Answering the queen: Machine learning and financial crises","author":"J Fouliard","year":"2019","journal-title":"BIS"},{"key":"ref60","article-title":"Currency crashes in emerging markets: An empirical treatment","author":"J A Frankel","year":"1996","journal-title":"Journal of International Economics"},{"key":"ref61","doi-asserted-by":"crossref","first-page":"216","DOI":"10.1016\/j.jinteco.2011.12.009","article-title":"Can leading indicators assess country vulnerability? Evidence from the 2008-09 Global Financial Crisis","volume":"87","author":"J Frankel","year":"2012","journal-title":"Journal of International Economics"},{"key":"ref62","doi-asserted-by":"crossref","first-page":"109","DOI":"10.1002\/ijfe.203","article-title":"On currency crises and contagion","volume":"8","author":"M Fratzscher","year":"2003","journal-title":"International Journal of Finance and Economics"},{"issue":"1","key":"ref63","doi-asserted-by":"crossref","DOI":"10.1006\/jcss.1997.1504","article-title":"A Decision-theoretic generalization of on-line learning and an application to boosting","volume":"55","author":"Y Freund","year":"1997","journal-title":"Journal of Computer and System Sciences"},{"issue":"5","key":"ref64","doi-asserted-by":"crossref","DOI":"10.1214\/aos\/1013203451","article-title":"Greedy function approximation: A gradient boosting machine","volume":"29","author":"J H Friedman","year":"2001","journal-title":"The Annals of Statistics"},{"key":"ref65","first-page":"3133","article-title":"Do we need hundreds of classifiers to solve real world classification problems?","volume":"15","author":"M Fernandez-Delgado","year":"2014","journal-title":"Journal of Machine Learning Research"},{"key":"ref66","author":"Fsb","year":"2022","journal-title":"Covid-19 financial stability impacts"},{"issue":"183","key":"ref67","article-title":"Report on biological standards iii: Methods of biological assay depending on quantal response","author":"J H Gaddum","year":"1933","journal-title":"Medical Research Council"},{"key":"ref68","author":"A Gayt\u00e1n","year":"2002","journal-title":"A review of the literature on early warning systems for banking Crises. Central Bank of Chile Working Paper"},{"issue":"3","key":"ref69","doi-asserted-by":"crossref","DOI":"10.2307\/4149942","article-title":"Structural vulnerabilities and currency crises","volume":"50","author":"S R Ghosh","year":"2003","journal-title":"IMF Staff Papers"},{"key":"ref70","author":"W E Greene","year":"2008","journal-title":"Econometric Analysis. 6 th Edition"},{"issue":"136","key":"ref71","first-page":"1","article-title":"Machine learning applied to banking supervision a literature review","volume":"9","author":"P Guerra","year":"2021","journal-title":"Risks"},{"key":"ref72","author":"J D Hamilton","year":"1994"},{"key":"ref73","author":"D C Hardy","year":"1998","journal-title":"Leading indicators of banking crises: Was asia different? IMF Working Paper"},{"key":"ref74","author":"P Hartmann","year":"2012","journal-title":"Widespread financial instabilities and the macroeconomy-regime switching in the Euro area?"},{"key":"ref75","doi-asserted-by":"crossref","DOI":"10.1007\/978-0-387-84858-7","author":"T Hastie","year":"2009","journal-title":"The Elements of Statistical Learning: Data Mining, Inference and Prediction"},{"key":"ref76","author":"A Hjerpe","year":"2016","journal-title":"Computing random forest variable importance measures (VIM) on mixed continuous and categorical data"},{"key":"ref77","author":"J Hoeting","year":"1993","journal-title":"Model selection and accounting for model uncertainty in linear regression models"},{"issue":"3","key":"ref78","article-title":"Application of ridge analysis to regression problems","volume":"58","author":"A E Hoerl","year":"1962","journal-title":"Chemical Engineering Progress"},{"key":"ref79","article-title":"Financial crises: Causes and indicators. World Economic Outlook","author":"Imf","year":"1998","journal-title":"International Financial Statistics"},{"key":"ref80","author":"G James","year":"2013","journal-title":"An introduction to statistical learning -with applications in R"},{"key":"ref81","article-title":"Banking and currency crises: Differential diagnostics for developed countries. european central bank","author":"M Joy","year":"2015","journal-title":"Working Paper Series"},{"issue":"1","key":"ref82","article-title":"Leading indicators of currency crises","volume":"45","author":"G Kaminsky","year":"1998","journal-title":"IMF Staff Papers"},{"key":"ref83","doi-asserted-by":"crossref","DOI":"10.1257\/aer.89.3.473","article-title":"The twin crises: The causes of banking and balance-of-payments problems","volume":"89","author":"G L Kaminsky","year":"1999","journal-title":"American Economic Review"},{"key":"ref84","doi-asserted-by":"crossref","first-page":"726","DOI":"10.1111\/j.1468-5965.2012.02264.x","article-title":"Macroeconomic imbalances as indicators for debt Crises in Europe","volume":"50","author":"T Knedlik","year":"2012","journal-title":"Journal of Common Market Studies"},{"key":"ref85","author":"C S Lin","year":"2006","journal-title":"A New approach to modelling early warning systems for currency crises: Can a machine-learning fuzzy expert system predict the currency crises effectively?"},{"key":"ref86","author":"A L\u00f3pez-Claros","year":"1989","journal-title":"Growth-oriented adjustment: Spain in the 1980s. Finance and Development"},{"key":"ref87","article-title":"Rules of thumb' for sovereign debt crises","author":"P Manasse","year":"2009","journal-title":"Journal of International Economics"},{"issue":"1","key":"ref88","article-title":"Two and two make one: The collapse of the State Bank of South Australia","volume":"11","author":"G Mccarthy","year":"1996","journal-title":"Policy, Organization and Society"},{"key":"ref89","doi-asserted-by":"crossref","first-page":"626","DOI":"10.1111\/lamp.12322","article-title":"Economic crises in Mexico: Right and wrong policy measures and a long period of structural failures","volume":"14","author":"J C Moreno-Brid","year":"2023","journal-title":"Latin American Policy"},{"issue":"2","key":"ref90","doi-asserted-by":"crossref","DOI":"10.1257\/jep.31.2.87","article-title":"Machine learning: An applied econometric approach","volume":"31","author":"S Mullainathan","year":"2017","journal-title":"Journal of Economic Perspectives"},{"issue":"6","key":"ref91","article-title":"The financial crisis in Japan during the 1990s: How the Bank of Japan responded and the lessons learnt","author":"H Nakaso","year":"2001","journal-title":"BIS Papers"},{"key":"ref92","author":"R Nyman","year":"2016","journal-title":"Predicting economic recessions using machine learning algorithms"},{"issue":"2","key":"ref93","doi-asserted-by":"crossref","first-page":"335","DOI":"10.1093\/biomet\/75.2.335","article-title":"Testing for a unit root in time series regression","volume":"75","author":"C B Phillips","year":"1988","journal-title":"Biometrika"},{"key":"ref94","doi-asserted-by":"crossref","DOI":"10.1145\/321105.321114","article-title":"A technique for the numerical solution of certain integral equations of the first kind","volume":"9","author":"D L Phillips","year":"1962","journal-title":"Journal of the ACM"},{"key":"ref95","article-title":"Cross-country causes and consequences of the 2008 crisis: Early warning","author":"A K Rose","year":"2009"},{"issue":"4","key":"ref96","doi-asserted-by":"crossref","DOI":"10.1137\/0907087","article-title":"Linear inversion of band-limited reflection seismograms","volume":"7","author":"F Santosa","year":"1986","journal-title":"SIAM Journal on Scientific and Statistical Computing"},{"issue":"19","key":"ref97","article-title":"Multidimensional distance-to-collapse point and sovereign default prediction. Intelligent Systems in Accounting","author":"R Savona","year":"2012","journal-title":"Finance and Management"},{"key":"ref98","volume":"8","author":"C Strobl","year":"2007","journal-title":"Bias in random forest variable importance measures: Illustrations, sources and a solution"},{"key":"ref99","doi-asserted-by":"crossref","first-page":"1389","DOI":"10.1109\/LSP.2014.2337313","article-title":"Fast implementation of DeLongs algorithm for comparing the areas under correlated receiver operating characteristic curves","volume":"21","author":"X Sun","year":"2014","journal-title":"IEEE Signal Processing Letters"},{"issue":"1","key":"ref100","article-title":"Regression shrinkage and selection via the lasso","volume":"58","author":"R Tibshirani","year":"1996","journal-title":"Journal of the Royal Statistical Society"},{"issue":"5","key":"ref101","article-title":"On the Stability of Inverse Problems","volume":"39","author":"A N Tikhonov","year":"1943","journal-title":"Doklady Akademii Nauk SSSR"},{"key":"ref102","article-title":"Solution of incorrectly formulated problems and the regularization method","volume":"4","author":"A N Tikhonov","year":"1963","journal-title":"Soviet Mathematics"},{"key":"ref103","author":"P J G Vlaar","year":"2000","journal-title":"Currency Crises Models for Emerging Markets"},{"key":"ref104","author":"C J Van Rijsbergen","year":"1979","journal-title":"Information Retrieval"},{"issue":"2","key":"ref105","doi-asserted-by":"crossref","first-page":"359","DOI":"10.1002\/jae.2525","article-title":"Spotting the danger zone: Forecasting financial crises with classification tree ensembles and many predictors","volume":"32","author":"F Ward","year":"2017","journal-title":"Journal of Applied Econometrics"},{"key":"ref106","author":"H Wold","year":"1985","journal-title":"Partial Least Squares. Encyclopaedia of Statistical Sciences"},{"key":"ref107","year":"2021","journal-title":"Covid-19 and financial fragility"}],"container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,10,15]],"date-time":"2026-10-15T10:20:16Z","timestamp":1792059616000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=4463667"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023]]},"references-count":107,"URL":"https:\/\/doi.org\/10.2139\/ssrn.4463667","relation":{},"subject":[],"published":{"date-parts":[[2023]]},"subtype":"preprint"}}