CMStatistics 2016: Start Registration
View Submission - CFE
Title: Risk measure estimates in quiet and turbulent times: An empirical study Authors:  Marie Kratz - ESSEC Business School, CREAR (France) [presenting]
Abstract: The aim is to test the relevance in practice, especially in time of financial crisis, of the Sample Quantile Process as a risk measure compared with the use of the historical VaR, or when using a GARCH model. We also specifically analyse the VaR following a period of crisis characterized by a high volatity. To do so, we develop an empirical study inspired by a previous one performed on S\& P 500 data and show that in time of high volatility the risk measure calculated on the historical data overestimate the risk measure computed on the future data. This expected feature is evidenced by our empirical results. Moreover, this fact introduces a pro-cyclical behavior of the risk measure estimate, tending to overestimate the risk in the future, whereas, in quiet time (low volatility), the risk measure estimate tends to underestimate the future risk by a large amount.