We develop early warning models for financial crisis prediction using machine learning techniques on macrofinancial data for 17 countries over 1870-2016. Machine learning models mostly outperform logistic regression in out-of-sample predictions and forecasting. We identify economic drivers of our machine learning models using a novel framework based on Shapley values, uncovering non-linear relationships between the predictors and crisis risk. Throughout, the most important predictors are credit growth and the slope of the yield curve, both domestically and globally. A flat or inverted yield curve is of most concern when nominal interest rates are low and credit growth is high.
The 2008 financial crisis has shown that financial busts can influence the real economy. However, there is less evidence to suggest that the same holds for financial booms. Using a Markov-Switching vector autoregressive model and euro area data, I show that financial booms tend to be less procyclical than financial busts. To identify the sources of asymmetry, I estimate a non-linear DSGE model with a heterogeneous banking sector and an occasionally binding borrowing constraint. The model matches the key features of the data and shows that the borrowers' balance sheet channel accounts for the asymmetry in the macro-financial linkages. The muted macro-financial transmission during financial booms can be exploited for macroprudential policies. By comparing capital buffer rules with monetary policy 'leaningagainst-the-wind' rules, I find that countercyclical capital buffers improve welfare.
We use monthly US utility patent applications to construct an external instrument for identification of technology news shocks in a rich-information VAR. Technology diffuses slowly, and affects total factor productivity in an S-shaped pattern. Responsible for about a tenth of economic fluctuations at business cycle frequencies, the shock elicits a slow, but large and positive response of quantities, and a sluggish contraction in prices, followed by an endogenous easing in the monetary stance. The ensuing economic expansion substantially anticipates any material increase in TFP. Technology news are strongly priced-in in the stock market on impact, but measures of consumers' expectations take sensibly longer to adjust, consistent with a New-Keynesian framework with nominal rigidities, and featuring informationally constrained agents.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.