Journal of Risk

Farid AitSahlia    
Warrington College of Business, University of Florida

This issue of The Journal of Risk covers portfolio diversification through risk parity, the application of neural networks to stochastic volatility model calibration, a comparison of credit risk with the market jump hazard rate in times of turmoil, and the impact that the communication tone of central bankers has on investor sentiment.

In the issue’s first paper, “Approximate risk parity with return adjustment and bounds for risk diversification”, Viraat Singh and Ali Hirsa incorporate returns into risk diversification strategies to exploit the presence of profitable assets that would otherwise be overlooked by strategies that focused only on risk minimization. They use real-world data to show that their approach can be implemented with software packages that are commercially available and that can solve the related convex programs efficiently.

In “The power of neural networks in stochastic volatility modeling”, the second paper in this issue, Caspar Sch¨on and Martin Walther address the calibration challenges associated with stochastic volatility models for options nearing their maturities. In particular, the authors show that neural networks can capture non-Markovian features and, based on data covering periods of high market volatility, accurately and efficiently approximate implied volatilities for rough stochastic volatility models.

Our third paper is “A tale of two tail risks” by Xin Huang, who conducts an empirical analysis to compare banking credit risk and the financial market jump hazard rate. Using data covering both the 2007–9 global financial crisis and the 2020–21 Covid-19 pandemic, Huang finds that banking credit tail risk negatively affects the jump hazard rate, capturing the countercyclical nature of volatility, which leaves less room for jumps to occur in bad times and more room in good times.

Closing out the issue is “The impact of divergence in communication tone on investors’ willingness to invest in eurozone small- to medium-sized enterprises”. Dimitris Anastasiou, Stelios Giannoulakis, Christos Kallandranis and Styliani-Iris Krokida use proprietary survey data from the euro area and machine learning to gauge investors’ “willingness to invest” in small and medium-sized enterprises (SMEs) based on the tone of central banks’ communications. They show that a divergence in tone between central bankers negatively affects investors’ willingness to invest in SMEs, thereby demonstrating the importance of communication in effecting monetary policy.