Podcast: SocGen quants on exotics calibration, machine learning and autocallable pricing

Deep learning techniques are being explored by the quants to speed up exotics pricing

Podcast 30 July 2018
Monika Ghose

Pierre Henry-Labordère, a member of the global markets quantitative analytics team at Societe Generale Corporate & Investment Banking, and Hamza Guennoun, a senior quantitative researcher within the same team, dialled in from Paris to talk about their new paper, Equity modelling with local stochastic volatility and stochastic discrete dividends.

The quants explained the mechanics of the physics technique they use in the paper, called the particle method, to calibrate exotic options with both equity and dividend underlyings, such as knockout dividend swaps, in the presence of stochastic dividends.

The quants said pricing of these products have been challenging until now because dividends would have to be modelled stochastically and this would lead to a slightly inaccurate calibration.

Fixing this problem by calibrating the model in a more accurate way can make it easier to risk-manage products with stochastic dividends, said SG CIB’s Guennoun.

“Exotics on both equities and dividends are not very liquid, but this is partly due to the fact that traders don’t have the tools to risk-manage the products,” Guennoun said. “This model can increase the liquidity of such products. It will allow the trader to handle at the same time, in a flexible way, the joint density of the equity underlying and also dividends.”

The quants also spoke about their future research projects, which over the next couple of years will focus mainly on the application of deep learning techniques to the calibration of multi-dimensional local stochastic volatility models.

“We are also trying to speed up the pricing of exotics. This is actually a crucial issue for the business. This will allow clients to have access to the price very quickly and to play with the payoff parameters and with several underlyings. To do that one idea would be to learn the pricing formula using neural networks,” said Guennoun.

One example of this would be autocallables, which are typically written as the worst of three stock underlyings. Having a faster pricing method would, therefore, allow clients to tweak the parameters and observe the prices very quickly.   


0:00 Introduction

1:20 Knockout dividend swaps

2:33 Applications of knockout dividend swaps

4:03 Issues with the calibration of models with stochastic dividends

5:26 The particle method

7:01 The particle method in exotics pricing

8:56 Discrete dividends

11:07 Future projects

12:45 Autocallable pricing

To hear the full interview, listen in the player above, or download. Future podcasts in our Quantcast series will be uploaded to Risk.net. You can also visit the main page here to access all tracks or go to the iTunes store to listen and subscribe.

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact info@risk.net or view our subscription options here: http://subscriptions.risk.net/subscribe

You are currently unable to copy this content. Please contact info@risk.net to find out more.

You need to sign in to use this feature. If you don’t have a Risk.net 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