An Evaluation of Early Warning Models for Systemic Banking Crises: Does Machine Learning Improve Predictions?
Johannes Beutel, Sophia List, Gregor von Schweinitz
Abstract
This paper compares the out-of-sample predictive performance of different early warning models for systemic banking crises using a sample of advanced economies covering the past 45 years. We compare a benchmark logit approach to several machine learning approaches recently proposed in the literature. We find that while machine learning methods often attain a very high in-sample fit, they are outperformed by the logit approach in recursive out-of-sample evaluations. This result is robust to the choice of performance measure, crisis definition, preference parameter, and sample length, as well as to using different sets of variables and data transformations. Thus, our paper suggests that further enhancements to machine learning early warning models are needed before they are able to offer a substantial value-added for predicting systemic banking crises. Conventional logit models appear to use the available information already fairly effciently, and would for instance have been able to predict the 2007/2008 financial crisis out-of-sample for many countries. In line with economic intuition, these models identify credit expansions, asset price booms and external imbalances as key predictors of systemic banking crises.
Artikel Lesen
Banks Fearing the Drought? Liquidity Hoarding as a Response to Idiosyncratic Interbank Funding Dry-ups
Helge Littke, Matias Ossandon Busch
IWH Discussion Papers,
Nr. 12,
2018
Abstract
Since the global financial crisis, economic literature has highlighted banks’ inclination to bolster up their liquid asset positions once the aggregate interbank funding market experiences a dry-up. To this regard, we show that liquidity hoarding and its detrimental effects on credit can also be triggered by idiosyncratic, i.e. bankspecific, interbank funding shocks with implications for monetary policy. Combining a unique data set of the Brazilian banking sector with a novel identification strategy enables us to overcome previous limitations for studying this phenomenon as a bankspecific event. This strategy further helps us to analyse how disruptions in the bank headquarters’ interbank market can lead to liquidity and lending adjustments at the regional bank branch level. From the perspective of the policy maker, understanding this market-to-market spillover effect is important as local bank branch markets are characterised by market concentration and relationship lending.
Artikel Lesen
On the Empirics of Reserve Requirements and Economic Growth
Jesús Crespo Cuaresma, Gregor von Schweinitz, Katharina Wendt
Abstract
Reserve requirements, as a tool of macroprudential policy, have been increasingly employed since the outbreak of the great financial crisis. We conduct an analysis of the effect of reserve requirements in tranquil and crisis times on credit and GDP growth making use of Bayesian model averaging methods. In terms of credit growth, we can show that initial negative effects of higher reserve requirements (which are often reported in the literature) tend to be short-lived and turn positive in the longer run. In terms of GDP per capita growth, we find on average a negative but not robust effect of regulation in tranquil times, which is only partly offset by a positive but also not robust effect in crisis times.
Artikel Lesen
Banking Globalization, Local Lending, and Labor Market Effects: Micro-level Evidence from Brazil
Felix Noth, Matias Ossandon Busch
Abstract
This paper estimates the effect of a foreign funding shock to banks in Brazil after the collapse of Lehman Brothers in September 2008. Our robust results show that bank-specific shocks to Brazilian parent banks negatively affected lending by their individual branches and trigger real economic consequences in Brazilian municipalities: More affected regions face restrictions in aggregated credit and show weaker labor market performance in the aftermath which documents the transmission mechanism of the global financial crisis to local labor markets in emerging countries. The results represent relevant information for regulators concerned with the real effects of cross-border liquidity shocks.
Artikel Lesen
Sovereign Credit Risk Co-movements in the Eurozone: Simple Interdependence or Contagion?
Manuel Buchholz, Lena Tonzer
International Finance,
Nr. 3,
2016
Abstract
We investigate credit risk co-movements and contagion in the sovereign debt markets of 17 industrialized countries during the period 2008–2012. We use dynamic conditional correlations of sovereign credit default swap spreads to detect contagion. This approach allows us to separate contagion channels from the determinants of simple interdependence. The results show that, first, sovereign credit risk co-moves considerably, particularly among eurozone countries and during the sovereign debt crisis. Second, contagion varies across time and countries. Third, similarities in economic fundamentals, cross-country linkages in banking and common market sentiment constitute the main channels of contagion.
Artikel Lesen
To Invest or Not to Invest, That Is the Question: Analysis of Firm Behavior under Anticipated Shocks
Dejan Kovač, Nikola Kleut, Boris Podobnik, Vuk Vukovic
Plos One,
Nr. 8,
2016
Abstract
When companies are faced with an upcoming and expected economic shock some of them tend to react better than others. They adapt by initiating investments thus successfully weathering the storm, while others, even though they possess the same information set, fail to adopt the same business strategy and eventually succumb to the crisis. We use a unique setting of the recent financial crisis in Croatia as an exogenous shock that hit the country with a time lag, allowing the domestic firms to adapt. We perform a survival analysis on the entire population of 144,000 firms in Croatia during the period from 2003 to 2015, and test whether investment prior to the anticipated shock makes firms more likely to survive the recession. We find that small and micro firms, which decided to invest, had between 60 and 70% higher survival rates than similar firms that chose not to invest. This claim is supported by both non-parametric and parametric tests in the survival analysis. From a normative perspective this finding could be important in mitigating the negative effects on aggregate demand during strong recessionary periods.
Artikel Lesen