Title: The impact of higher order moments added to the multiscale Carhart model for different investment periods
Authors: Anyssa Trimech
Addresses: Faculty of Economic and Management Sciences, Computational Mathematics Laboratory, Tunisia
Abstract: Nonlinear market risk is considered to have a significant effect on the return variation. This risk as well as those proposed by Carhart (1997) determines the financial changes. This paper seeks to explore the importance of taking into account this effect through the investigation conducted on the model of Carhart augmented by the higher order moments, using the multiresolution analysis (MRA) based on wavelet decomposition over different time scales. This tool improves the accuracy and the performance of the period and portfolio investment selection. The empirical results show the importance of nonlinear market systematic risk (as measured by skewness and kurtosis) in addition to the risks captured by Carhart risks (linear systematic market risk, unsystematic risk connected to size, value and momentum). Moreover, they emphasise the high sensitivity and dependence of the model explanatory power and risks to the time scale variation.
Keywords: risk measures; nonlinear market risk; size; value; momentum; multiresolution analysis; timescale; higher order moments; multiscale Carhart models; investment periods; wavelet decomposition; portfolio investment selection; skewness; kurtosis.
International Journal of Decision Sciences, Risk and Management, 2014 Vol.5 No.3, pp.263 - 276
Received: 14 Apr 2014
Accepted: 01 Jul 2014
Published online: 18 Feb 2015 *