Perspectives from Academia and Industry: Synergies in Financial Research 

Edition: 'On the value of information in data driven markets' 

 

In the past years the shift towards data driven analytics in the financial industry has brought an array of unexplored possibilities to quantitative researchers both in academia and industry. In order to create a platform of dialogue between London based universities and the financial Industry, the math finance group from King's College London initiates the series Perspectives from Academia and Industry: Synergies in Financial Research.

Regular workshops on current relevant topics in Mathematical Finance will be organised, bringing together international leading experts from Academia and Industry. 

 

All those interested in attending these events are welcome.

 

On November 21st we will be hosting the first workshop: 

'On the value of information in data driven markets'

devoted to topics related to the measurement, processing and the value (and inherently the limitations) of data driven modelling with a focus on FX trades and Fixed Income.

 

We are particularly delighted to give weight to the importance of the topic by welcoming speakers with a formative impact on these directions. We also provide a platform to junior quants and researchers and allocate plenty of time for discussions to facilitate a dialogue.

  

Workshop Programme:

 

Date and Time:  Wednesday 21 November 15:00-18:00 

Location: K3.11, Kings College London, King’s Building.

 

3:00          Welcome

3:05-3:15  Flash-talk by Laurids Nielsen (Deutsche Bank) Machine Learning for High Frequency FX Predictions

3:15-3:50  Jan Obloj (Oxford University) From Optimal Transport to Robust Pricing and Hedging via Neural Networks

3:50-4:25  Jessica James (Commerzbank) Yield Curves, Cross-currency Basis and the Nature of Arbitrage in the Post-Crisis World

4:25-4:50  Drinks reception and time for discussions

4:50-5:25  Damiano Brigo (Imperial College London) Option Pricing Range under Statistically Indistinguishable Models: A New Look at Historical and Implied Volatilities

5:25-6:00  Roel Oomen (Deutsche Bank) Electronic FX trading – where Game Theory meets Data Science

***********************************************************

 

Programme and Abstracts:

 

Damiano Brigo (Imperial College London)

Title: Option Pricing Range under Statistically Indistinguishable Models: A New Look at Historical and Implied Volatilities

Abstract: Given a Black and Scholes geometric Brownian motion with a fixed volatility, we construct alternative stock price processes with the same laws (and returns’ distributions) in a given discrete-time grid. We then illustrate how option prices based on such statistically indistinguishable alternative stock price models span the whole arbitrage free range, independently of the mesh size of the initially chosen discrete time grid. We explain how this framework allows to reconcile historical and implied volatilities. In the final part of the talk, if time permits, we will partly explain the above result by showing that option pricing only relies on pathwise properties of the underlying, as in Bender et al (2008) based on semimartingales quadratic variation and subsequent generalizations, including Armstrong et al (2018) who make use of Davie’s type rough differential equations, and of “rough brackets”. This entails the definition of a Gubinelli derivative, which in the classical Black and Scholes’ framework is the Gamma sensitivity of the option.

About the Speaker: Damiano Brigo is Chair and Head of Mathematical Finance at Imperial College London.  Formerly Managing Director at Fitch Ratings and Head of Credit Models in Banca IMI, Damiano is in the board of a number of financial institutions. Building on his joint industry and academic experience, Damiano teaches executive education to regulators, banks and hedge funds. Damiano published 100+ works in Mathematical Finance, Probability and Statistics, and field reference books in interest rate and credit modeling (H-index 38, 7000+ citations). Damiano has been the most cited Risk Magazine author in 1998-2017.  Damiano holds a PhD in differential geometric stochastic filtering.

 

Jessical James (Commerzbank)

Title: Yield Curves, Cross-currency Basis and the Nature of Arbitrage in the Post-Crisis World

Abstract: The markets changed after 2008.  A new appreciation of credit risk and a raft of regulations have led to unexpected consequences and a multiplication of valuation curves.  FX hedging has become vastly more complex and some deals which look like riskless profit turn out to be nothing of the sort.  But in this maze there are occasional gems and opportunities.  We delve deep into the topic of cross currency basis swaps to show the audience how the world has changed.

About the Speaker: Jessica James is the Senior Quantitative Researcher in the Rates Research team at Commerzbank., where she covers foreign exchange and fixed income.  She joined Commerzbank from Citigroup where she was Global Head of the Quantitative Investor Solutions Group. Previously, she lectured in physics at Trinity College, Oxford. Significant publications include ‘FX Option Performance’,  ‘Handbook of Foreign Exchange’ (Wiley), 'Interest Rate Modelling' (Wiley), and 'Currency Management' (Risk books). She is on the Board of the Journal of Quantitative Finance, a Fellow of the Institute of Physics, and is a Visiting Professor at UCL and Cass Business School.  

 

Laurids Nielsen (Deutsche Bank)

Title: Machine Learning for High Frequency FX Predictions

Abstract: This talk discusses the applicability of machine learning methods for forecasting price movements in high frequency foreign exchange markets. With OLS as a benchmark, the performances of several machine learning models are investigated. This is done for both a short and a longer prediction horizon and for multiple currency pairs. It is shown that machine learning methods are capable of predicting the short horizon significantly better than the linear models but tend to struggle when the prediction horizon is longer and data more noisy.

About the Speaker: Laurids holds a BSc in Mathematical Economics from Aarhus University and an MSc in Mathematics and Finance from Imperial College. Currently, he is working as an eFX quant at Deutsche Bank. 

 

Jan Obloj (University of Oxford)

Title: From Optimal Transport to Robust Pricing and Hedging via Neural Networks

Abstract: I introduce the robust approach to pricing and hedging which does not start with an priori probability measure but is instead data driven. The framework is designed to interpolate between model-independent and model-specific settings and to allow to address and quantify model risk. I explain briefly how classical fundamental notions and theorems in quantitative finance extend to the robust setting. I then focus on a simple two-dimensional study case of pricing and hedging a spread option, introducing suitable numerical methods and presenting concrete examples. I use vanilla option prices, together with agent-prescribed bounds on key market characteristics, to drive the interval of no-arbitrage prices and the associated hedging strategies. The setting can be seen as a constrained variant of the classical optimal transportation problem and comes with a natural pricing-hedging duality. I discuss numerical methods based on discretization and LP implementation but subsequently focus on a deep NN optimization. At the end I will outline some of the higher-dimensional challenges for such methods as well as way to coherently combine option price data with past time series data in one estimation procedure.

About the Speaker: Jan Obloj is a Professor of Mathematics and a member of the Mathematical and Computational Finance group, at the Mathematical Institute, University of Oxford. He is also a Fellow of St John’s College and an associate member of the Oxford-Man Institute of Quantitative Finance. He is an expert on embedding problems and time-change techniques for stochastic processes and has been at the forefront of developments in robust approach to pricing and hedging. He has given numerous invited talks, including the Bruti-Liberati lecture at the QMF in Sydney and a plenary lecture at the SIAM Financial Mathematics Conference. His research is supported by a European Research Council grant. Jan holds a PhD degree from Unviersite Paris 6 and University of Warsaw and, before moving to Oxford in 2008, he was a Marie Curie Postdoctoral Fellow at Imperial College London. 

 

Roel Oomen (Deutsche Bank)

Title: Yield Curves, Cross-currency Basis and the Nature of Arbitrage in the Post-Crisis World

Abstract: In this talk, I will discuss recent developments in electronic FX trading and show how data science applied to dense financial data can be used to make practical decisions around execution optimisation. I will go over a  number of case studies taken from a live trading environment.

About the Speaker:  Roel is the head of FIC quantitative trading at Deutsche Bank. He started his industry career as a quant in cash equity algo trading in 2006, and subsequently held various roles in electronic FX spot trading, including co-head of the business. Roel holds a PhD in econometrics, is a senior research fellow at the London School of Economics, and has published widely on the econometric analysis of high frequency data and FX trading.

 

Organisers: Roxana Dumitrescu, Blanka Horvath, Teemu Pennanen