This manuscript proposes a market crash forecasting methodology based on the analysis of the shape parameter of the alpha-stable distribution. Moreover, the parameters of Pareto type II distribution are also employed for comparison purposes. The model performance is assessed by Quadratic Probability Score, Logarithmic Probability Score, and Directional Accuracy measures. Our applications for developed and emerging markets show a good accuracy for alpha-stable distribution and replicability of the results. Thus, our model is an adequate early warning system for future crashes.
