Reflecting market globalization, foreign exchange trading is the largest and most liquid market in the world. This issue of The Journal of Risk contains a paper that contributes strategies to hedge foreign exchange risk over short-term horizons, such as trading sessions. Other topics included in this issue are empirical assessments of value-at-risk (VaR) models and market and credit risk integration, and a numerical method to determine marginal risk in asset portfolios.
The first paper in this issue, "Evaluating the performance of the skewed distributions to forecast value-at-risk in the global financial crisis", is by Pilar Abad, Sonia Benito Muela, Carmen López-Martín and Miguel Angel Sánchez Granero. The authors use data from major world market indexes and compare the out-of-sample performance of a number of common fat-tailed and skewed distributions, based on loss functions reflecting regulator and firm objectives. Their study illustrates the trade-offs between statistical fitting of asset returns and VaR forecast accuracy. It also contrasts the VaR modeling preferences of firms and those of their regulators, with the former favoring models that result in lower regulatory capital.
In our second paper, "Stochastic receding horizon control for short-term risk management in foreign exchange", Farzad Noorian, Barry Flower and Philip H.W. Leong propose a dynamic technique that adapts models to new data in order to hedge foreign currency risk over the short term. These strategies are focused more on spot positions than on derivatives, which are typically used for longer horizons. Through a study based on real-world data, the authors show that their approach can significantly reduce hedging costs relative to standard benchmarks.
"Finite difference methods for estimating marginal risk contributions in asset management", by Michael Olschewsky, Stefan Lüdemann and Thorsten Poddig, is the issue's third paper. In contrast to standard approaches, for which explicit expressions to estimate marginal risk contributions come at the cost of stringent assumptions, the authors instead rely on a general risk model based on kernel density estimation for return distributions and on finite differences to estimate marginal risk contributions. Through use of a numerical study, the authors show that statistical estimation errors dominate numerical differentiation errors, thus implying that finite differences can be assessed on very coarse discretizations.
The fourth and final paper in this issue, "Comparing risk measures when aggregating market risk and credit risk using different copulas" by Jakob Maciag, Frederik Hesse, Rolf Boeve and Andreas Pfingsten, considers the aggregation of market risk and credit risk in a manner that fits the top-down approach promulgated in Basel III. Through use of a simulation study they find consistent ranking of riskiness by different risk measures and copulas. However, they also find that tail-based measures are much more sensitive to input parameters than deviation-based measures.
University of Florida
Evaluating the performance of the skewed distributions to forecast value-at-risk in the global financial crisis
This paper models the tail behavior of daily returns and forecasting VaR in order to evaluate the performance of several skewed and symmetric distributions.
The authors of this paper formalize a methodology to manage short-term FX risk.
The authors of this paper simulate realistic total bank return distributions by means of a top-down copula approach for different parameter settings.
This paper studies the use of finite difference methods for estimating risk contributions.