CMStatistics 2017: Start Registration
View Submission - CFE
Title: Daily vs intraday risk assessment using asynchronous tick-by-tick data Authors:  Simona Boffelli - Bergamo University (Italy) [presenting]
Giovanni Urga - Cass Business School (UK)
Abstract: When tick-by-tick data are available, risk evaluation conducted at high-frequency is shown to outperform their counterparts executed with the traditional daily frequency. To this purpose, it is crucial to adopt an appropriate synchronization scheme and a correct estimator of the integrated covariance matrix when working in a multivariate framework. We evaluate a portfolio of European government bonds over the period 1st June 2007 31st May 2012 focusing on a trading strategy based on risk control measure and a risk management backtesting exercise using unilevel, multilevel Value-at-Risk tests and Tail Risk measures. We provide clear cut evidence of the benefit of exploiting the additional information in the intraday data with respect to daily frequency. Both trading strategy and risk management backtesting tests more accurately forecast VaR when high-frequency data are used.