The identification of the best model in terms of volatility forecast accuracy is a troublesome task and many evaluation methods have been proposed on the basis of a statistical or economic approach. The aim of this work is to investigate the opportunity to use a statistical approach in a VaR framework, i.e. evaluating the VaR measures by means of a loss function. By using high-frequency data it is possible to achieve a consistent estimate of the VaR bootstrapping the intraday increments of an asset. Hence, the performances of the volatility models are compared with that employing the VaR consistent estimate. In particular, the ‘true’ VaR is used to find a threshold discriminating low from high loss function values for each volatility model. The proposed procedure is assessed by means of a Monte Carlo simulation.
The use of loss functions in assessing the VaR measures
AMENDOLA, Alessandra;CANDILA, VINCENZO
2014-01-01
Abstract
The identification of the best model in terms of volatility forecast accuracy is a troublesome task and many evaluation methods have been proposed on the basis of a statistical or economic approach. The aim of this work is to investigate the opportunity to use a statistical approach in a VaR framework, i.e. evaluating the VaR measures by means of a loss function. By using high-frequency data it is possible to achieve a consistent estimate of the VaR bootstrapping the intraday increments of an asset. Hence, the performances of the volatility models are compared with that employing the VaR consistent estimate. In particular, the ‘true’ VaR is used to find a threshold discriminating low from high loss function values for each volatility model. The proposed procedure is assessed by means of a Monte Carlo simulation.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.