Journal of Risk

Farid AitSahlia    
Warrington College of Business, University of Florida

Return barrier options, semiparametric and dynamic estimation of common risk measures, and risk parity strategies are addressed in this issue of The Journal of Risk.

In the first paper in the issue, “Return to the barrier: option pricing and calibration in foreign exchange markets”, Justin Lars Kirkby, Claudio Aglieri Rinella, Jean- Philippe Aguilar and Nathaniel Rupprecht examine calibration aspects for a new type of option, called the return barrier option (RBO), that is directly exposed to jump risk in its underlying asset. Analyzing Lévy models, stochastic volatility models with and without jumps, and stochastic local volatility models, among others, Kirkby et al find that in the context of foreign exchange, RBOs are sensitive to model assumptions. In addition, the authors observe that relying exclusively on either Lévy or jump-diffusion processes is insufficient to price these options properly. Instead, combining stochastic volatility with jumps results in more accurate pricing.

In “Semiparametric GARCH models for value-at-risk and expected shortfall: an object-driven procedure”, the issue’s second paper, Yuanhua Feng and Christian Peitz propose the use of object-driven smoothing techniques on semiparametric generalized autoregressive conditional heteroscedasticity (GARCH) models for standard risk measures such as value-at-risk (VaR) and expected shortfall (ES). Feng and Peitz’s approach is based on kernel regression with different bandwidths, and parametric models are therefore included as special cases. Through numerical illustrations, the authors show that their algorithm can identify when a semiparametric GARCH model is preferable to a parametric form, and vice versa.

Our third paper, “A dynamic method-of-moments copula model approach for market risk estimates”, is by Wolfgang Aussenegg and Christian Cech, who apply a method-of-moments approach to estimate VaR and ES in a dynamic setting. They account for heteroscedasticity by using exponential GARCH models on volatility-adjusted returns. Their empirical results show that the method-of-moments technique reduces the copula estimation time for VaR and ES while preserving accuracy for large portfolios that include a variety of asset classes.

Closing out the issue, in “Risk parity strategies with risk factors”, Yu-Jui Chen, Chen-Yen Hsieh, Huei-Wen Teng, Ming-Che Hu and Alex YiHou Huang use Standard & Poor’s 500 data and a sample selection based on risk factors such as size, value and momentum to empirically evaluate the performance of risk parity strategies against traditional value-weighted and equally weighted portfolios. Chen et al find that both naive and dynamic risk parity strategies generate the highest risk-adjusted returns, especially for sample selections based on a single factor (most notably, firm size). In particular, they show that for investments with a long horizon, risk parity constitutes an effective and profitable alternative risk-adjusted strategy.