Journal of Risk

Contributions offered in this issue of The Journal of Risk include one on systemic risk estimation; one on realized volatility forecasting that accounts for financial stress; one on counterparty default risk minimization; and one on accurate performance evaluation of a portfolio relative to a benchmark.

In our first paper, “Nonparametric estimation of systemic risk via conditional value-at-risk”, Ahmed Belhad, Davide Lauria and A. Alexandre Trindade introduce concomitant- and saddle-point-based estimators of conditional value-at-risk to assess systemic risk. They show that these estimators are consistent under mild regularity conditions and that the saddle-point approximation is very helpful for asymptotic normality. Finite-sample insights are further gained from simulations, and an application is illustrated on leading market indexes from the United States, Europe and Asia.

In the issue’s second paper, “Forecasting the realized volatility of stock markets with financial stress”, Chuan Guo and Yiyun Feng incorporate a recently developed financial stress index into the heterogeneous autoregressive model, which is known for capturing long memory, in order to forecast realized volatility. Guo and Feng validate their approach empirically by showing that it generates improved forecast accuracy for long-term horizons.

In “Counterparty risk allocation”, the third paper in the issue, Rainer Baule addresses the problem of minimizing the risk of an exposure to counterparties. Baule shows that if there are two counterparties, only sparse solutions (eg, allocation to only one counterparty) are optimal for spectral risk measures such as expected shortfall and discrete loss distributions. Further, the risk-minimizing strategy does not depend monotonically on the confidence level associated with the expected shortfall.

Ruihong Jiang, David Saunders and Chengguo Weng round out this issue with their paper “The statistics of capture ratios”, in which they focus on relative portfolio performance evaluation in a manner that accounts for whether the benchmark is “up” or “down” through the so-called capture ratios. Whether returns are identically and independently distributed or serially correlated, Jiang et al derive asymptotic results for the distributions of the nontrivial estimators for these ratios. These results are then used to show, via a simulation analysis, that large samples are required for reliable statistical inference.

You need to sign in to use this feature. If you don’t have a account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here: