Journal of Risk

Farid AitSahlia
Warrington College of Business, University of Florida

The papers in this issue of The Journal of Risk address the generalization of value-at-risk (VaR) to multiple dimensions and the application of machine learning techniques to risk identification and categorization. They also provide empirical studies that assess the impact of forecast maturities and sampling frequencies on foreign exchange rates and option pricing.

Our first paper, “Risk measures: a generalization from the univariate to the matrix-variate” by M. Andrea Arias-Serna, Francisco J. Caro-Lopera and Jean-Michel Loubes, exploits the flexibility afforded by the parameters of the beta distribution to generalize VaR expressions to multiple dimensions. The authors use an example to illustrate the extension of Gaussian hypergeometric functions from one-dimension settings to zonal polynomials in a multiple-dimension case.

“A general framework for the identification and categorization of risks: an application to the context of financial markets” is the second paper in this issue. In it, Micha Bender and Sven Panz introduce a general framework for the identification of risk types in specific contexts using unsupervised learning techniques; they offer an illustration using financial markets. The authors show how their approach can help identify risk categories at various levels of granularity in a systematic and consistent fashion that is more efficient than current manual and biased methods.

“Modeling realized volatility with implied volatility for the EUR/GBP exchange rate”, our third paper, sees Anna Rokicka and Janusz Kudła use heterogeneous autoregression models of realized volatility in an empirical study to illustrate the impact of option maturities on the forecast accuracy of EUR/GBP exchange rates. They show in particular that weekly and daily frequencies work best when returns are asymmetric but that monthly horizons perform better in the presence of measurement error.

The issue’s fourth and final paper, “Option pricing using high-frequency futures prices”, is by Stavros Degiannakis, Christos Floros, Thomas Poufinas, George Filis and Konstantinos Gkillas. They use futures-based estimates of variance in an ad hoc Black–Scholes model. In an empirical study involving Standard & Poor’s 500 data, the authors illustrate how both longer time to maturity and data sampling frequency adversely affect pricing accuracy, in contrast to option moneyness, with at-the-money options performing best.