Value-at-risk (VaR) is a risk measure of potential loss on a specific portfolio. The main uses of VaR are in risk management and financial reporting. Researchers are continuously looking for new and efficient ways to evaluate VaR, and the 2008 financial crisis has given further impetus to finding new and reliable ways of evaluating and using VaR. In this study, the authors use genetic algorithm (GA) to evaluate VaR and compare the results with conventional VaR techniques.
In essence, the authors propose two modifications to the standard GA: normalized population selection and strict population selection. For a typical set of simulation, eight chromosomes were used each with eight stored values, and the authors get eight values for VaR.
The experiments using data from four different market indices show that by adjusting the volatility, the VaR computed using GA is more conservative as compared to those computed using Monte Carlo simulation.
The proposed methodology is designed for VaR computation only. This could be generalized for other applications.
This is achieved with much less cost of computation, and hence, the proposed methodology could be a viable practical approach for computing VaR.
The proposed methodology is simple and, at the same time, novel that could have far-reaching impact on practitioners.
The first author acknowledges financial support from the University of Manitoba through University of Manitoba Graduate Fellowship. The second and third authors acknowledge financial support from the Natural Sciences and Engineering Research Council Canada through the Discovery Grant program.
Sharma, B., Thulasiram, R.K. and Thulasiraman, P. (2015), "Computing value-at-risk using genetic algorithm", Journal of Risk Finance, Vol. 16 No. 2, pp. 170-189. https://doi.org/10.1108/JRF-09-2014-0132
Emerald Group Publishing Limited
Copyright © 2015, Emerald Group Publishing Limited