BoE published a working paper that details the early warning models developed for financial crisis prediction using machine learning techniques on macro-financial data for 17 countries over 1870–2016. The paper shows that machine learning models mostly outperform logistic regression in out‑of‑sample predictions and forecasting. Due to their greater flexibility, machine learning models have the advantage that they may uncover important non-linear relationships and variable interactions, which may be difficult to identify using classical techniques. The results help policy makers to identify the risk of financial crises in advance and potentially act on these signals.
The paper describes the dataset and reviews literature on the variables that the authors choose as predictors. It then presents the benchmark logistic regression, outlines the methodology, and provides a brief description of the different machine learning models applied and the Shapley value framework. Next, the paper compares the predictive performance of all models and investigates the importance of the predictors using Shapley values. Finally, the paper looks at the robustness of indicators in a macro-prudential policy context and investigates in detail the role of the yield curve.
This paper shows that machine learning models outperform logistic regression in predicting financial crises on a macroeconomic data set covering 17 countries between 1870 and 2016 in both out-of-sample cross-validation and recursive forecasting. The gains in predictive accuracy justify the use of initially more opaque machine learning models. All models consistently identify similar predictors for financial crises, although there are some variations across time reflecting changes in the nature of the global monetary and financial system. While the crucial role of credit is an established result in the literature, the predictive power of the yield curve has obtained far less attention as an early warning indicator. The authors also inspect non-linearities and interactions identified by the machine learning models. Global credit shows a particularly strong non-linearity—only very high global credit growth beyond a certain point influences the prediction of the models. Interactions are particularly strong between global and domestic indicators.
Overall, the findings suggest a combination of low risk perception, search-for-yield behavior, and strong credit growth in the years preceding a crisis. The results highlight the potential value of machine learning models for broader economic policy making in two key dimensions. First, the approach illustrates how machine learning techniques can uncover important non-linearities and interactions that facilitate superior out-of-sample prediction and forecasting even in situations characterized by relatively small data sets with limited observations of the event of interest. Second, the novel Shapley value approach demonstrates how the black box concern linked to the practical policy application of machine learning models may be overcome. By providing a mechanism to identify the key economic drivers of the predictions generated by such models, the approach allows insights from machine learning models to be integrated into a broader decision making framework while preserving the transparency and accountability of any resulting public policy decision.
Related Link: Working Paper
Keywords: Europe, UK, Banking, Machine Learning, Yield Curve, Financial Stability, Dataset, Research, Early Warning Model, Fintech, BoE
Previous ArticleBCB on Innovation Projects Selected for LIFT 2019 Edition
The European Commission (EC) published a public consultation on the review of revised payment services directive (PSD2) and open finance.
The European Commission (EC) has issued two letters mandating the European Supervisory Authorities (ESAs) to jointly propose amendments to the regulatory technical standards under Sustainable Finance Disclosure Regulation or SFDR.
The European Banking Authority (EBA) published its annual report on convergence of supervisory practices for 2021. Additionally, following a request from the European Commission (EC),
The Farm Credit Administration published, in the Federal Register, the final rule on implementation of the Current Expected Credit Losses (CECL) methodology for allowances
The U.S. Securities and Exchange Commission (SEC) looks set to intensify focus on crypto-assets and cyber risk and extended the comment period on the proposed rules to enhance and standardize climate-related disclosures for investors.
The Australian Prudential Regulation Authority (APRA) announced reduction in the aggregate Committed Liquidity Facility and issued an update on the operational preparedness for zero and negative market interest rates.
The Commission for the Financial Market (CMF) in Chile published capital adequacy ratios (as of February 2022, January 2022, and December 2021) for 17 banks and for the banking system.
The Prudential Regulation Authority (PRA) issued a statement on the European Banking Authority (EBA) guidelines on management of non-performing exposures (NPEs) and forborne exposures.
The European Banking Authority (EBA) updated the implementing technical standards that specify the data collection for the 2023 supervisory benchmarking exercise in relation to the internal approaches used in market risk, credit risk, and IFRS 9 accounting.
The European Insurance and Occupational Pensions Authority (EIOPA) published a feedback statement on the responses received to the consultation on blockchain and smart contracts in insurance.