Journal of Risk

Farid AitSahlia    
Warrington College of Business, University of Florida

This issue of The Journal of Risk addresses the forecasting of a common volatility index, the correlation between collateralized debt obligations (CDOs) and a systematic factor, the risk spillover between Chinese financial institutions, and the estimation of covariance matrixes based on data captured in different currencies.

In the first paper in the issue, “Forecasting the Volatility Index with a realized measure, volatility components and dynamic jumps”, Xinyu Wu, Yuyao Wang and Bo Zhang integrate realized volatility, two volatility components and the dynamic jump intensity into a single exponential generalized autoregressive conditional heteroscedasticity model with dynamic jumps (the REGARCH-2C-Jump model). Due to its nonaffine nature, REGARCH-2C-Jump accounts for the realized measure bias caused by nontrading hours and microstructure noise and avoids the imposition of positiveness conditions for model estimation, in contrast to the standard affine model. Through an empirical validation, Wu et al show that their model is superior to affine GARCH models used in forecasting the Chicago Board Options Exchange Volatility Index (VIX).

The issue’s second paper, “Cumulative accuracy profile curves for correlating collateralized debt obligations to systematic factors” by David Lozinski and Chris Stavnitzky, uses Lorenz, or cumulative accuracy profile (CAP), curves to calibrate the correlation between CDO defaults and a general systematic factor affecting bonds in a diversified portfolio. In this fixed-income context, a Lorenz curve relates the cumulative proportion of CDO defaults to the losses incurred in the general bond portfolio. Based on Monte Carlo simulation involving a variety of related industries and a wide range of bond ratings, Lozinski and Stavnitzky show that CDO defaults are highly correlated with a general systematic risk factor, and their findings provide additional context for the role of CDOs in the global financial crisis.

In “Unveiling multiscale dynamics: exploring financial risk spillover and influencing factors among Chinese financial institutions”, our third paper, Ce Guo, Qiwei Xie, Jingyu Li and Dandan Zhang propose an approach to study dynamic patterns of financial risk spillover between Chinese financial institutions across different time scales and periods. Guo et al make use of the maximal overlap discrete wavelet transform (MODWT) to decompose time series of transaction prices according to their long-, medium- or short—term impact. They combine the MODWT with a vine copula to construct a multiscale network. Through a standard Delta-conditional value-at-risk (ΔCoVaR) model and a spatial econometric model, the authors are able to assess systemic risk and identify factors that influence risk spillover.

The short final paper in the issue, “Converting a covariance matrix from local currencies to a common currency” by Gianluca Fusai, Domenico Mignacca and Khalifa Al-Thani, addresses a problem faced by institutions that operate globally and need to estimate covariances based on returns captured in different currencies. By exploiting a simple algebraic property to estimate the covariance matrix between log returns expressed in different currencies, the authors’ approach avoids the need to first convert all the data into a single currency and then carry out covariance estimation.