A Machine Learning Approach: Financial Crisis Forecasting in G7 Countries

Öz

This study investigates the effectiveness of machine learning methods in predicting systemic banking crises, particularly using the XGBoost algorithm, which stands out for its ability to capture complex and non-linear patterns. A XGBoost-based model was developed to predict systemic banking crises using financial and macroeconomic data from G7 countries for the period 1870-2020. Additionally, SHAP (SHapley Additive exPlanations) methods were applied to analyze the causal relationships between model results, thereby overcoming the model's “black box” nature and providing a deeper understanding of decision-making processes. This allowed for a transparent analysis of the causal relationships between predictive variables and crisis risk. The findings demonstrate that XGBoost exhibits high predictive performance, offering new opportunities for practitioners and policymakers in assessing crisis risk. Additionally, SHAP values significantly enhance the transparency and accountability of machine learning models by revealing the complex relationships between predictive variables and crisis risk. This approach provides a robust and reliable analytical framework for identifying the fundamental economic drivers of financial crises, highlighting the potential of machine learning methods in financial crisis prediction.

Anahtar Kelimeler

• Financial Crises
• Machine Learning
• XGBoost
• SHAP Value
• Forecasting

Bir Makine Öğrenimi Uygulaması: G7 Ülkelerinde Finansal Kriz Tahminleme

Öz

Bu çalışma, sistemik bankacılık krizlerini tahmin etmede makine öğrenmesi yöntemlerinin etkinliğini, özellikle karmaşık ve doğrusal olmayan örüntüleri yakalama yeteneğiyle öne çıkan XGBoost algoritmasını kullanarak araştırmaktadır. 1870-2020 dönemi için G7 ülkelerine ait finansal ve makroekonomik veriler kullanılarak sistemik bankacılık kriz tahmininde XGBoost tabanlı bir model geliştirilmiştir. Ayrıca, modelin 'kara kutu' doğasını aşarak karar alma süreçlerini derinlemesine anlamlandırmak amacıyla SHAP (SHapley Additive exPlanations) yöntemleri uygulanarak model sonuçları arasındaki nedensel ilişkiler analiz edilmiş, böylece tahmin edici değişkenler ile kriz riski arasındaki nedensel ilişkiler şeffaf bir şekilde analiz edilmiştir. Bulgular, XGBoost'un yüksek tahmin performansı sergileyerek uygulayıcılar ve politika yapıcılar için kriz riskini değerlendirmede yeni olanaklar sunduğunu göstermektedir. Ek olarak SHAP değerleri, tahmin edici değişkenler ile kriz riski arasındaki karmaşık ilişkileri ortaya çıkararak makine öğrenimi modellerinin şeffaflığını ve hesap verebilirliğini önemli ölçüde artırmaktadır. Bu yaklaşım, finansal krizlerin temel ekonomik itici güçlerini belirleme konusunda sağlam ve güvenilir bir analitik altyapı sunarak finansal kriz tahmininde makine öğrenmesi yöntemlerinin potansiyelini vurgulamaktadır.

Anahtar Kelimeler

• Finansal Krizler
• Makine Öğrenimi
• XGBoost
• SHAP Değeri
• Tahminleme

Kaynakça

1. Aikman, D., Galesic, M., Gigerenzer, G., Kapadia, S., Katsikopoulos, K., Kothiyal, A., Murphy, E. and Neumann, T. (2021). Taking uncertainty seriously: Simplicity versus complexity in financial regulation. Industrial and Corporate Change, 30(2), 317-345. https://doi.org/10.1093/icc/dtaa024
2. Akinci, O. and Olmstead-Rumsey, J. (2018). How effective are macroprudential policies? An empirical investigation. Journal of Financial Intermediation, 33, 33-57. https://doi.org/10.1016/j.jfi.2017.04.001
3. Aksoy, N. and Genc, I. (2023). Predictive models development using gradient boosting based methods for solar power plants. Journal of Computational Science, 67, 101958. https://doi.org/10.1016/j.jocs.2023.101958
4. Alessi, L. and Detken, C. (2018). Identifying excessive credit growth and leverage. Journal of Financial Stability, 35, 215-225. https://doi.org/10.1016/j.jfs.2017.06.005
5. Babecký, J., Havránek, T., Matějů, J., Rusnák, M., Šmídková, K. and Vašíček, B. (2014). Banking, debt, and currency crises in developed countries: Stylized facts and early warning indicators. Journal of Financial Stability, 15, 1-17. https://doi.org/10.1016/j.jfs.2014.07.001
6. Berg, A. and Pattillo, C. (1999). Are currency crises predictable? A test. IMF Staff Papers, 46(2), 107-138. https://doi.org/10.2307/3867664
7. Bluwstein, K., Buckmann, M., Joseph, A., Kapadia, S. and Şimşek, Ö. (2023). Credit growth, the yield curve and financial crisis prediction: Evidence from a machine learning approach. Journal of International Economics, 145, 103773. https://doi.org/10.1016/j.jinteco.2023.103773
8. Brunnermeier, M.K. and Oehmke, M. (2013). Bubbles, Financial crises, and systemic risk. In G.M. Constantinides, M. Harris and R.M. Stulz (Eds.), Handbook of the economics of finance (pp. 1221-1288). https://doi.org/10.1016/B978-0-44-459406-8.00018-4

9. Cardarelli, R., Elekdag, S. and Lall, S. (2011). Financial stress and economic contractions. Journal of Financial Stability, 7(2), 78-97. https://doi.org/10.1016/j.jfs.2010.01.005
10. Chatzis, S.P., Siakoulis, V., Petropoulos, A., Stavroulakis, E. and Vlachogiannakis, N. (2018). Forecasting stock market crisis events using deep and statistical machine learning techniques. Expert Systems with Applications, 112, 353-371. https://doi.org/10.1016/j.eswa.2018.06.032
11. Chawla, N.V., Bowyer, K.W., Hall, L.O. and Kegelmeyer, W.P. (2002). SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16, 321-357. https://doi.org/10.1613/jair.953
12. Chen, S., Huang, Y. and Ge, L. (2024). An early warning system for financial crises: A temporal convolutional network approach. Technological and Economic Development of Economy, 30(3), 688–711. https://doi.org/10.3846/tede.2024.20555
13. Dabrowski, J.J., Beyers, C. and de Villiers, J.P. (2016). Systemic banking crisis early warning systems using dynamic Bayesian networks. Expert Systems with Applications, 62, 225-242. https://doi.org/10.1016/j.eswa.2016.06.024
14. Dhaliwal, S.S., Nahid, A.-A. and Abbas, R. (2018). Effective ıntrusion detection system using XGBoost. Information, 9(7), 149. https://doi.org/10.3390/info9070149
15. Giese, J., Nelson, B., Tanaka, M. and Tarashev, N.A. (2013). How could macroprudential policy affect financial system resilience and credit? Lessons from the literature (Bank of England Financial Stability Paper No. 21). https://doi.org/10.2139/ssrn.2266366
16. He, Y., Lu, X., Fournier-Viger, P. and Huang, J.Z. (2024). A novel overlapping minimization SMOTE algorithm for imbalanced classification. Frontiers of Information Technology & Electronic Engineering, 25(9), 1266-1281. https://doi.org/10.1631/FITEE.2300278
17. Hoggarth, G., Reis, R. and Saporta, V. (2002). Costs of banking system instability: Some empirical evidence. Journal of Banking & Finance, 26(5), 825-855. https://doi.org/10.1016/S0378-4266(01)00268-0
18. Jordà, Ò., Schularick, M. and Taylor, A.M. (2011). Financial crises, credit booms, and external imbalances: 140 years of lessons. IMF Economic Review, 59(2), 340-378. https://doi.org/10.1057/imfer.2011.8
19. Kaminsky, G.L. and Reinhart, C.M. (1999). The twin crises: The causes of banking and balance-of-payments problems. American Economic Review, 89(3), 473-500. doi:10.1257/aer.89.3.473
20. Kauko, K. (2014). How to foresee banking crises? A survey of the empirical literature. Economic Systems, 38(3), 289-308. https://doi.org/10.1016/j.ecosys.2014.01.001
21. Kindleberger, C.P. (2000). Manias, panics, and crashes: A history of financial crises. The Scriblerian and the Kit-Cats, 32(2), 379. Retrieved from https://www.proquest.com/
22. Laeven, M.L. and Valencia, M.F. (2012). Systemic banking crises database: An update (IMF Working Paper No. 12/163). Retrieved from https://www.imf.org/external/pubs/ft/wp/2012/wp12163.pdf
23. Laeven, M.L. and Valencia, M.F. (2018). Systemic banking crises revisited (IMF Working Paper No. 2018/206). Retrieved from https://www.imf.org/en/Publications/WP/Issues/2018/09/14/Systemic-Banking-Crises-Revisited-46232
24. Levine, R. (2005). Finance and growth: Theory and evidence. In P. Aghion and S.N. Durlauf (Eds.), Handbook of economic growth (pp. 865-934). https://doi.org/10.1016/S1574-0684(05)01012-9
25. Liu, L., Chen, C. and Wang, B. (2022), Predicting financial crises with machine learning methods, Journal of Forecasting, 41(5), 871-910. https://doi.org/10.1002/for.2840
26. Lundberg, S. and Lee, S.-I. (2017). A unified approach to interpreting model predictions. In I. Guyon, U. Von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan and R. Garnett (Eds.), Advances in neural information processing systems. Papers presented at the NIPS2017, California: Curran Associates.
27. Meng, Y., Yang, N., Qian, Z. and Zhang, G. (2020). What makes an online review more helpful: An interpretation framework using XGBoost and SHAP values. Journal of Theoretical and Applied Electronic Commerce Research, 16(3), 466-490. https://doi.org/10.3390/jtaer16030029
28. Minsky, H.P. (1995). Sources of financial fragility: Financial factors in the economics of capitalism (Levy Economics Institute of Bard College No. 69). Retrieved from https://digitalcommons.bard.edu/cgi/viewcontent.cgi?article=1068&context=hm_archive
29. Mitchell, R. and Frank, E. (2017). Accelerating the XGBoost algorithm using GPU computing. PeerJ Computer Science, 3, e127. https://doi.org/10.7717/peerj-cs.127
30. Nazareth, N. and Ramana Reddy, Y.V. (2023). Financial applications of machine learning: A literature review. Expert Systems with Applications, 219, 119640. https://doi.org/10.1016/j.eswa.2023.119640
31. Obstfeld, M. (2012). Does the current account still matter? American Economic Review, 102(3), 1-23. doi: 10.1257/aer.102.3.1
32. Jordà, Ò., Schularick, M. and Taylor, A.M. (2017). Macrofinancial history and the new business cycle facts. NBER Macroeconomics Annual, 31(1), 213-263. https://doi.org/10.1086/690241
33. Osman, A.I.A., Latif, S.D., Boo, K.B.W., Ahmed, A.N., Huang, Y.F. and El-Shafie, A. (2024). Advanced machine learning algorithm to predict the implication of climate change on groundwater level for protecting aquifer from depletion. Groundwater for Sustainable Development, 25, 101152. https://doi.org/10.1016/j.gsd.2024.101152
34. Qin, C., Zhang, Y., Bao, F., Zhang, C., Liu, P. and Liu, P. (2021). XGBoost optimized by adaptive particle swarm optimization for credit scoring. Mathematical Problems in Engineering, 2021(1), 6655510. https://doi.org/10.1155/2021/6655510
35. Reinhart, C.M. and Rogoff, K.S. (2011). This time is different: Eight centuries of financial folly. New Jersey: Princeton University Press. https://doi.org/10.1515/9781400831722
36. Schularick, M. and Taylor, A.M. (2012). Credit booms gone bust: monetary policy, leverage cycles, and financial crises, 1870–2008. American Economic Review, 102(2), 1029-1061. doi:10.1257/aer.102.2.1029
37. Shankarpandala, S. (2024). LazyPredict: A library for quick and easy model selection in machine learning. Retrieved from https://pypi.org/project/lazypredict/
38. Shapley, L.S. (1953). A value for n‐person games. In H.W. Kuhn (Ed.), Classics in game theory (pp. 307-317). New Jersey: Princeton University Press.
39. Tölö, E. (2020). Predicting systemic financial crises with recurrent neural networks. Journal of Financial Stability, 49, 100746. https://doi.org/10.1016/j.jfs.2020.100746
40. Tran, K.L., Le, H.A., Nguyen, T.H. and Nguyen, D.T. (2022). Explainable machine learning for financial distress prediction: Evidence from Vietnam. Data, 7(11), 160. https://doi.org/10.3390/data7110160
41. Vašíček, B., Žigraiová, D., Hoeberichts, M., Vermeulen, R., Šmídková, K. and de Haan, J. (2017). Leading indicators of financial stress: New evidence. Journal of Financial Stability, 28, 240-257. https://doi.org/10.1016/j.jfs.2016.05.005
42. World Bank. (2023). World development indicators. Retrieved from https://databank.worldbank.org/source/world-development-indicators