This paper uses a comprehensive set of predictor variables from the five largest Eurozone countries to compare the performance of simple univariate and machine learning-based multivariate models in forecasting stock market crashes. In terms of statistical predictive performance, a support vector machine-based crash prediction model outperforms a random classifier and is superior to the average univariate benchmark as well as a multivariate logistic regression model. Incorporating nonlinear and interactive effects is both imperative and foundation for the outperformance of support vector machines. Their ability to forecast stock market crashes out-of-sample translates into substantial value-added to active investors. From a policy perspective, the use of machine learning-based crash prediction models can help activate macroprudential tools in time.

Portfolio insurance strategies are used on both the institutional and the retail side of the asset management industry. While standard utility theory struggles to provide an explanation, this study justifies the popularity of portfolio insurance strategies in a behavioral finance context. We run Monte Carlo simulations as well as historical simulations for popular portfolio insurance strategies and benchmark strategies in order to evaluate the outcomes using cumulative prospect theory. Our simulation results indicate that most portfolio insurance strategies are the preferred investment strategy for a prospect theory investor. Moreover, the analysis provides insights into how portfolio insurance products should be designed and structured to meet the preferences of prospect theory investors as accurately as possible.