This issue of the Journal of Risk marks the fifth year of publication of the journal. The journal is making good progress in its mission to disseminate academic and practitioner research into financial risk management. It has benefited from an increasing flow of good quality research in the topics of financial market risk, credit risk, and operational risk.
Over the last five years, the journal has received 261 submissions and has published 85 papers. The average acceptance rate has been 35% over the last five years and 27% over the last two years. Given the specialized nature of the journal and its youth, these are encouraging statistics.
The papers in this issue reflect the breadth of recent risk management topics. They include a paper on portfolio credit risk, three papers on market risk that emphasize the multivariate nature of modern risk models, and a paper on option pricing.
In “Evaluating credit risk models using loss density forecasts,” H. Frerichs and G. Löffler address the verification of credit risk models. This is a crucial issue in the context of the Basel Committee revision of the 1988 Capital Accord. To the great disappointment of risk managers, the Basel Committee has still not allowed the use of internal portfolio credit risk models in the Basel II revision. Perhaps this is because many of the parameters are difficult to estimate and verify. This paper provides a more positive view, however. The authors show that tests for basic model parameters have acceptable statistical power.
The second paper, by J-D. Fermanian and O. Scaillet, “Nonparametric estimation of copulas for time series,” provides a nonparametric method of estimating copulas, that is, functions building joint distributions from their univariate marginal distributions. This topic is important given the observation that correlations are not necessarily constant, as assumed with multivariate normal distributions. Non-linear dependencies may cause extreme losses in periods of high volatility. The paper derives the asymptotic properties of kernel-based estimators of copulas and presents some examples.
In “Hedge funds revisited: distributional characteristics, dependence structure and diversification,” H. Geman and C. Kharoubi provide another application of copula functions. The paper analyzes the distributional characteristics of hedge fund indices and confirms that hedge fund returns are far from normally distributed. It then investigates their diversification properties relative to standard equity indices. The paper models non-linear correlations using copula functions. Most hedge funds show increased dependence with stock markets for large down moves. This implies that diversification benefits from hedge funds could disappear in the event of a financial crisis.
Next, the paper by S. Turkay, E. Epperlein, and N. Christofides, “Correlation stress testing for value-at-risk,” discusses stress tests of the correlation matrix of risk factors. When value-at-risk (VAR) is based on a covariance matrix method, stressing the correlations enables the risk manager to assess the sensitivity of VAR to changes, or errors, in correlations. In practice, this is a difficult problem due to the need to keep the matrix positive definite. This paper provides a method to perturb subsets of the correlation matrix while maintaining positive definiteness.
Finally, the paper by J. R. Sobehart and R. Farengo, “A dynamical model of market under- and overreaction,” provides an extension of the conventional Black–Scholes model starting from a more complex stochastic process that incorporates overreactions. This leads to the derivation of a distribution for returns, hedging strategies, and the pricing of options that can accommodate volatility smiles.
The mission of the Journal of Risk is to further our understanding of risk management. Contributions to the journal are welcome from academics, practitioners, and regulators in the field. With this in mind, authors are encouraged to submit full-length papers.