Khartoum University Journal of
Management Studies

OFFICIAL PUBLICATION OF THE SCHOOL OF MANAGEMENT STUDIES, UNIVERSITY OF KHARTOUM
  • Abbreviation: Khartoum Univ. J. Manage. Stud.
  • Language: English
  • ISSN: 1585-8069
  • DOI: 10.5897/KUJMS
  • Start Year: 1994
  • Published Articles: 35

Full Length Research Paper

Enhancing the accuracy of the historical simulation value-at-risk using exponentially weighted moving average (EWA) technique: A case of Khartoum stock exchange

Abdaljbbar B. A. Dawod
  • Abdaljbbar B. A. Dawod
  • Department of Forest Management, Faculty of Forestry, University of Khartoum, Khartoum, Sudan.
  • Google Scholar
Zakaryia M. S. Mohammad
  • Zakaryia M. S. Mohammad
  • Department of Statistics, Faculty of mathematical sciences, University of Khartoum, Khartoum, Sudan.
  • Google Scholar


  •  Received: 22 February 2018
  •  Accepted: 16 July 2018
  •  Published: 31 August 2018

Abstract

Value-at-Risk (VaR) is a quantitative risk measure estimated either parametrically or non-parametrically. Historical simulation is a non-parametric approach that relies on historical data and this data span for a long period, which might encounter changes over time in the market circumstances (that is, volatility). The accuracy of the historical simulation VaR (HSVaR) increases as the volatility adjusted to the current regime of the market. Various techniques are used to adjust the volatility such as Generalized Autoregressive Heteroscedastic (GARCH), Exponential GARCH (EGARCH), Exponentially Weighted Moving Average (EWMA), etc. Adjusting the volatility on the Sudanese markets using EWMA did not investigated so far; hence, this paper investigated the performance of HSVaR in Khartoum Stock Exchange (KSE) under EWMA technique. The findings showed that, under the current circumstances of KSE, either the conventional or EWMA approach with λ = 0.99 or 0.997 at 5 or 10% significance levels can produce an accurate HSVaR. For the best performance, the researchers recommend to use a decay factor λ= 0.99 or 0.997.

 

Key words: Value-at-Risk (VaR), generalized autoregressive heteroscedastic (GARCH), volatility, unconditional coverage test, independence test, backtesting.