CMStatistics 2020: Start Registration
View Submission - CFE
Title: Proper measures of connectedness for systemic risk detection: An application on real financial data Authors:  Mishel Qyrana - University of Pavia (Italy) [presenting]
Pierpaolo Uberti - University of Genoa (Italy)
Silvia Figini - University of Pavia (Italy)
Abstract: In the context of systemic risk analysis and forecasting, some particular instances of the so-called proper measures of connectedness are implemented. Proper measures of connectedness can be read as a generalization of the condition number of a matrix, allowing to consider the multiple collinearity between assets in a market. This aspect represents the loss in terms of diversification opportunities typical during systemic events. First, the choice of the parameter $k$, which represents the number of diversification opportunities at risk is discussed. Then, an accurate in-sample and out-of-sample application on real financial data is provided, discussing both the descriptive and predictive power of the measure. A final focus is dedicated to the out-of-sample analysis, in order to highlight when the proposed approach shows interesting forecasting power, making the measures good tools to be used as early warning signals of systemic risk.