Journal of Organisational Studies and Innovation
Vol. 2, no.2, Summer, 2015
Efficient Portfolio Optimisation Using the Conditional Value at Risk Approach
Olu Abraham Ayodeji, Lucie Ingram
University of West London

Abstract: Fifty years after Harry Markowitz’s ground breaking work on mean variance portfolios, the problem of what asset(s) to invest in is still central to finance today, as is how to minimise risk for a given level of expected return or maximise the return for a given level of risk. Initially, standard deviation and variance were the prominent risk measures but for inherent flaws, their usage and application yielded inconsistent and unreliable results; while The Value-at-Risk (VaR) which was introduced was not subadditive. This lack of coherence led to its modification, the result being the Conditional-Value-at-Risk (CVaR) which is the expected loss exceeding the Value-at Risk (VaR), which proved to satisfy the coherence demand in addition to other advantages.

The goal of this work is to apply this CVaR method to resolve optimisation problems by minimizing risk under CVaR constraints. This research considers the proble of finding the optimal weights and the efficient frontier associated with the risk-return portfolio optimization model of a sample portfolio of 20 FTSE100 stocks using the historical prices (data) for scenario generation and makes use of the R software for the optimization process. This research has discovered that the CVaR optimization process was much more robust than the MVO model. In addition, the 1998 financial crisis had a significant impact on the result of the optimized portfolio. It also discovered that stability and validity of results increased with the volume of historical data used. Moreover, findings suggest isolating 2008 data will allow further investigation into the impact and significance of the crisis amongst others.

Keywords: optimisation, CVar, MVO, portfolio, risk.