Forecasting Stock Market Crashes via Machine Learning

56 Pages Posted: 11 May 2021 Last revised: 17 Aug 2021

See all articles by Hubert Dichtl

Hubert Dichtl

dichtl research & consulting GmbH; University of Hamburg

Wolfgang Drobetz

University of Hamburg

Tizian Otto

University of Hamburg

Date Written: May 10, 2021

Abstract

This paper uses a comprehensive set of variables from the five largest Eurozone countries to compare the performance of simple univariate and machine learning-based multivariate models in predicting stock market crashes. The statistical predictive performance of a support vector machine-based stock market crash prediction model is significantly different from zero and among the best-performing univariate benchmarks, while still being truly out-of-sample. The ability to forecast subsequent stock market crashes out-of-sample translates into value-added to investors under realistic trading assumptions (net of transaction costs). Incorporating nonlinear and interactive effects is both imperative and foundation for the predictive performance of support vector machines. This adds an economic component to the advantageousness of machine learning-based multivariate crash prediction models over their univariate counterparts. It helps identify and explain the complex relationships in the underlying economic conditions (key economic drivers) that precede substantial stock market downturns.

Keywords: extreme event prediction, stock market crashes, machine learning, active trading strategy

JEL Classification: G11, G12, G14, G17

Suggested Citation

Dichtl, Hubert and Drobetz, Wolfgang and Otto, Tizian, Forecasting Stock Market Crashes via Machine Learning (May 10, 2021). Available at SSRN: https://ssrn.com/abstract=3843319 or http://dx.doi.org/10.2139/ssrn.3843319

Hubert Dichtl

dichtl research & consulting GmbH ( email )

Am Bahnhof 7
65812 Bad Soden am Taunus
Germany

HOME PAGE: http://www.dichtl-research-consulting.de

University of Hamburg ( email )

Moorweidenstr. 18
Hamburg, 20148
Germany

Wolfgang Drobetz

University of Hamburg ( email )

Moorweidenstrasse 18
Hamburg, 20148
Germany

Tizian Otto (Contact Author)

University of Hamburg ( email )

MoorweidenstraƟe 18
Hamburg, 20148
Germany

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Downloads
290
Abstract Views
937
rank
134,447
PlumX Metrics