Content area
Full Text
1. Introduction and literature review
The emergence of risk management has attracted attention from financial institutions and contributed to the financial market’s integration. Consequently, Value-at-Risk (VaR) has become a ubiquitous measure of risk due to its conceptual simplicity. It has also become a crucial tool for portfolio management (Jorion, 2002). From a theoretical point of view, several studies have evaluated the relevance of VaR models for various financial markets (Danielsson and De Vries, 2000; Angelidis et al., 2004; Ané, 2006). Mabrouk and Saadi (2012) assess the performance of several VaR models based on long memory conditional volatility models, with many distributions, for seven stock market indices. Their results reveal that conditional volatility with the skewed Student distribution provide accurate estimations for VaR models. Orhan and Köksal (2012), meanwhile, compare a list of Generalized Autoregressive Conditional Heteroskedastic (GARCH) models when quantifying VaR for developed and emerging stock market during the subprime financial crisis. Their findings show that Autoregressive Conditional Heteroskedasticity (ARCH) and GARCH model with Student distribution ate the best estimator.
Dimitrakopoulos et al. (2010) conduct study on several emerging markets in Latin America, Europe and Asia. They conclude that the performance of the parametric models outperforms non-parametric ones during post-crisis periods because extreme events were included in the estimation sample.
Mokni and Mansouri (2011), s asymmetric show that and long memory GARCH model provide good performance for VaR estimation for both short and long positions, and hence an important tool for risk managers.
Gençay and Selçuk (2004), Ho et al. (2000) and McNeil and Frey (2000) investigate the performance of VaR models with daily data for emerging stock markets. They all demonstrate that the traditional parametric VaR model with normal density does not successfully estimate loss during financial crises.
These findings have encouraged many other researchers to assess equity risk quantification for Middle East and North Africa (MENA) (Maghyereh and Al-Zoubi, 2006; Assaf, 2009). More specifically, Assaf (2015) examines the out-of-sample performance of VaR models for four MENA equity markets using the APARCH model. His findings show that APARCH model with student distribution show best performance for VaR estimation than those with normal distribution.
The huge investment opportunities in some emerging markets have fascinated the courtesy of investors. Indeed, the MENA region is one...