Crisis Risk Prediction with Concavity from Polymodel

Financial crises is an important research topic because of their impact on the economy, the businesses and the populations. However, prior research tend to show systemic risk measures which are reactive, in the sense that risk surges after the crisis starts. Few of them succeed in predicting financial crises in advance. In this paper, we first introduce a toy model based on a dynamic regime switching process producing normal mixture distributions. We observe that the relative concavity of various indices increases before a crisis. We use this stylized fact to introduce a measure of concavity from nonlinear Polymodel, as a crisis risk indicator, and test it against known crises. We validate the indicator by using it for a trading strategy that holds long or short positions on S&P 500, depending on the indicator value.