Journal of Computational Finance

Time flies, dear Reader. In 2013, I agreed to act as the editor-in-chief of The Journal of Computational Finance for five years. That seems like only yesterday, but here we are, five years later, and I believe it is now time to pass on the editorship.

“Kriging metamodels and experimental design for Bermudan option pricing” by Mike Ludkovski is the issue’s second paper. In it, the regression Monte Carlo framework, as proposed by Longstaff and Schwartz in 2001, is enhanced with stochastic Kriging to better fit the continuation values. Kriging is a well-known statistical technique, and in this context it offers a quantification of the quality of the regression approximation within optimal stopping problems. Moreover, a framework of improved experimental design based on space-filling and adaptivity is explained. The paper comes with several numerical experiments, confirming the robust and efficient solution technique.

Sergio De Diego, Eva Ferreira and Eulalia Nualart have contributed to the issue with “Importance sampling applied to Greeks for jump–diffusion models with stochastic volatility”. Their paper, the third in this issue, develops a variance reduction technique in the context of the so-called Robbins–Monro algorithm for option pricing under jump–diffusion with stochastic volatility. The main technique behind importance sampling is changing the drift term to reduce the variance of a sample. The idea is to determine an optimal drift in the Girsanov transform, under which the variance of the estimate is minimal. Numerical results for the Black–Scholes and Heston models with jumps as well as the Barndorff–Nielsen–Shephard model confirm the efficiency improvement.

Our fourth paper, “Importance sampling for jump–diffusions via cross-entropy”, is by Rebecca Rieke, Weifeng Sun and Hui Wang. This is another paper on the importance sampling variance reduction technique being applied to a class of jump–diffusion processes. An efficient and easy-to-implement importance sampling scheme is proposed, based on the method of cross-entropy, which is combined with an expectation–maximization algorithm. The sampling distributions are selected from the family of exponentially tilted distributions and their mixtures. A variety of path-dependent options are simulated with this technique to illustrate its fine performance.

Finally, I would like to thank the very friendly people in the editorial office in London, who have helped me with The Journal of Computational Finance whenever necessary. At present, it is Ciara Smith and Sarah Campbell at Infopro Digital, but previously I had Dawn Hunter, Carolyn Moclair and Jade Mitchell helping me out. Nada Mitrovic also helped me with the journal at the CWI – Center for Mathematics & Computer Science in Amsterdam. This is warmly acknowledged!

Let me also thank all the associate editors for their invaluable help with and support of the journal over the years. And, last but certainly not least, I thank the authors and readers. Your continued interest in this fascinating journal is hugely appreciated.

I wish you very enjoyable reading of this June 2018 issue of The Journal of Computational Finance.

Cornelis W. Oosterlee