The appeal of the conditional value-at-risk measure is mainly due to its theoretical properties. In this issue of The Journal of Risk we explore its practicality as a criterion for portfolio selection. Further topics addressed in this issue include wavelet analysis for portfolio selection - especially in cases where the short- and long-term correlations between assets vary - and currency option pricing when the underlying exchange rate is non-mean-reverting.
The technique of linear programming approximation is the standard used in the determination of portfolios that minimize conditional value-at-risk. In the first paper of this issue, "Suboptimality in portfolio conditional value-at-risk optimization", Edgars Jakobsons conducts a numerical study to assess the suboptimality of this approach. He finds that, in contrast to light-tailed distributions and when a small number of assets are used, a significantly large Monte Carlo sample is needed for acceptable accuracy when the number of assets is moderately large and return distributions are heavy tailed.
In the issue's second paper, "Outperforming benchmarks with their derivatives: theory and empirical evidence", Alejandro Balbás, Beatriz Balbás and Raquel Balbás use conditional value-at-risk as a risk measure to assess the risk-adjusted performance of certain strategies. They then show how one can construct portfolios including common options with the property that their expected returns increase while their risk decreases (good deals). In an empirical assessment of these strategies, they illustrate their superior performance relative to benchmarks based on their underlying indexes and commodities.
"Wavelet decomposition and applied portfolio management" by Theo Berger is the third paper in the issue, and it considers the application of wavelet analysis to portfolio selection. This technique is particularly well suited to the determination of correlations between assets along different time-scales. The author shows in particular that in market downturns, minimizing long-run volatility is inefficient while minimizing short-term volatility beats standard benchmarks.
In the issue's fourth and final paper, "Pricing options on trend-stationary currencies: applications to the Chinese yuan", Michael Mebane considers currency option pricing and hedging when the exchange rate drifts, instead of the standard assumption of mean reversion. His model is particularly relevant within the realm of sovereign currency control and skewed markets. Through an empirical validation, he shows that it outperforms the standard Garman-Kohlhagen and Heston models.
University of Florida
This paper considers the portfolio optimization problem, with conditional value-at-risk as the objective.
This paper looks for optimal explicit constructions and empirical tests in regards to pricing and hedging derivatives with coherent risk measures.
In order to separate short-term noise from long-term trends, this paper decomposes financial return series into their time and frequency domains.
This paper derives a closed-form version of a model with a trend-stationary, stochastic volatility exchange rate, using both a linear and quadratic trend.