26th MathFinance Conference

Abstract

We begin by discussing the origin of impact in OTC markets driven by dealer’s risk management in response to and/or in anticipation of client flow. We further address dealer’s risk due to informational content of their own quotes and demonstrate how to preserve the inventory-management value of skew while limiting the information it leaks. For dealers exposed to client flow with superior information, we show that defence is not the only possible choice, the dealer can rationally accept adverse selection if the information helps manage the rest of the portfolio.

Biography

Alexander is a director at Foreign Exchange, Rates and Commodities, HSBC, specializing in the design and development of market making and execution algorithms. He moved to finance in 2007 from the field of theoretical chemical physics. He held a senior research position at the National Institute of Advanced Industrial Science and Technology, Japan. Alexander received his PhD in physics and mathematics from the Moscow Institute for Physics and Technology.

Abstract

The analysis of volatility smiles is a central component of options pricing theory. An attempt to unify the many models and frameworks into a theory applicable to any underlying is what this paper is. I draw conclusions from the rates, equities, and index markets on the general dynamics of skew, and provide statistical properties of cryptocurrency options skews. This shines light not only on the dynamics unique to crypto options, but also on the implications for exotic option pricing and for the pricing of prediction market bets, like polymarket. A principal component analysis is used in conjunction to the implied volatility regimes to better understand the supply and demand economics that lead to skew.

Biography

Mike Beckhusen is the founder of VolaLabs, a Berlin-based quantitative research firm building implied volatility surfaces. His current work focuses on systematically decomposing vol-surface P&L into skew components and trading them as market-neutral factors. Before VolaLabs, he worked in bank risk and derivatives functions at Deloitte, covering market, credit and interest-rate risk with a focus on VaR, ICAAP and structured-derivatives pricing. Prior to his career in Quantitative Finance he was a professional race car driver, giving him an unique perspective on risk management.

Abstract

We present a generative model for smooth arbitrage free non-parametric option surfaces (“SANOS”) and discuss applications to hedging derivatives.

Biography

Dr. Hans Buehler is a leading quantitative finance executive and researcher known for pioneering work in machine-learning-based derivatives hedging. He led the Equity Derivatives Quant Research at Deutsche Bank before joining J.P. Morgan in 2008, where he became Global Head of Equities Quantitative Research and later co-lead the Markets’ Analytics, Automation & Optimization “AI” effort.

At JPMorgan, he spearheaded the “deep hedging” framework, applying neural networks to trading and risk management—work that earned him Risk.net’s 2022 Quant of the Year award.

In 2022 Hans joined XTX Markets as Deputy CEO and later Co-CEO. He has served as a visiting professor at TU Munich and is now a visiting professor at the Mathematical Institute of the University of Oxford. Hans holds a PhD in Financial Mathematics from TU Berlin.

Abstract

Fx Trading evolves all the time with its drivers shifting as technology and events swing markets, but the goal remains the same for all participants:

• How do I maximise my position?
  • Regardless if a corporate looking to fund operational needs or a day-trader looking to maximise straight P’n’L
• And what does the future hold for my job?

We will look at the factors to be faced in this industry and see what that means of us.

Biography

Jeremy Carvell is a banking and FX markets specialist currently working at 360T (360 Treasury Systems AG), a leading global trading platform for foreign exchange and money market products.

Post university, Jeremy started in the City of London as a Fixed Income analyst, before following in the genetical footsteps to become a financial journalist and, later, editor at various institutions including Euromoney and Bloomberg. Eventually being re-located to Frankfurt to take on account management, he switched back into banking to join Citibank managing accounts on the CATS-OS Warrants multi-dealer platform. From this, he joined 360T as employee #19 and working with liquidity providers across the EMEA region in optimizing their electronic trading strategies, liquidity sourcing, and execution quality.

With this strong background in foreign exchange and the bond markets, Jeremy has extensive experience engaging with banks, liquidity providers, and institutional clients on topics such as liquidity fragmentation, algo execution, and trading workflow efficiency.

Outside of his professional work, he is actively engaged in industry and community initiatives, including charity fundraising, motorcycling and a keen tennis player.

Abstract

Efficiency in the complete stack is key for a business, and in the structured business this needs to be combined with flexibility.

Biography

Michael obtained his BA in Engineering from Cambridge and his PhD in Robotics from Oxford. He started as Quant at Credit Suisse Financial Products (1998-2000), then was a quant at Citi (2000-2003), then a structured IR products trader at Citi (2003-2005). In 2005 he joined Swiss Re as an IR Options trader to help build the structured IR Options desk from scratch (risk managing using Markov and BGM models). In 2008 he moved to Commerzbank, and is since 2015 Global Head of IR and FX Options Trading.

Abstract

We explore the calculation of the sensitivity with respect to the implied volatility surface for products valued with Monte Carlo simulation in the local volatility model. The theoretical foundation as well as practical implementation details will be discussed.

Biography

Jürgen works as a Quant at Leonteq AG. He started his career working on foreign exchange models. He extended his field of interest to equity, commodity, and interest rate derivatives and hybrids. He has a keen interest in market microstructure and in applying machine learning methods to quantitative finance problems.

Abstract

In the light of the upcoming macro hedge accounting framework IFRS Risk Mitigation Accounting, we discuss how to identify overhedging in an easy asset & liability framework employing linear instruments. Furthermore, we present various methodologies separating the overhedging from the non-overhedging part to create hypothetical derivatives and show a brief demo of our self-developed software RMA Solution.

Biography

Karl Hofmann leads Deloitte Germany’s chapter Financial Engineering (including all quant activities on financial instruments valuation, market risk, counterparty risk and liquidity risk) co-leads Deloitte’s international initiative on IFRS Dynamic Risk Management. He holds a PhD in Mathematics from TU München.

Abstract

In this talk, we present recent results on the construction and interpolation of arbitrage-free implied volatility surfaces.

The first part focuses on a characterization of differentiable, arbitrage-free implied volatility slices. Using Fukasawa’s normalizing volatility transforms (NVTs), we show that every such slice satisfies a first-order Abel ODE. We then provide numerical examples illustrating how this characterization can be used to construct and interpolate implied volatility smiles across strikes.

In the second part, we study the interpolation of volatility slices across expiries. We establish a connection between calendar and strike arbitrage in NVT coordinates and discuss how this relationship can be used to construct arbitrage-free implied volatility surfaces.

Biography

Vladimir Lucic is a Visiting Professor at Imperial College London and Head of Quants at Marex Solutions, London. During his 25+ career as a quant Vladimir focused on various aspects of quant modelling and volatility investment strategies. He has published in premier academic and practitioner’s journals. Vladimir is the recipient of the Risk.net Quant of the Year award for 2025.

Abstract

The research project “Risk Culture Indicators and their Role in Financial Institutions’ Resilience” builds on Fritz-Morgenthal, Hellmuth, and Packham (2016), who developed a set of nine risk culture indicators from annual reports and linked these to stress test outcomes. We develop an LLM-based pipeline that scores bank risk culture automatically from published annual reports, targeting a broad panel of European banks in the time period 2013 to 2023. The pipeline combines retrieval-augmented generation with navigation over the reports’ structure and returns, for every score, the evidence passages it rests on. We validate against the original 2013 expert assessment on two levels: does the model retrieve the passages the experts cited, and do its scores match theirs. In short, does it score like the expert, and for the expert’s reasons?  The resulting time series will feed new analyses of the link between risk culture and bank resilience under changing regulatory and economic conditions.

Biography

Natalie Packham is Professor of Mathematics and Statistics at Berlin School of Economics and Law. Natalie has several years of industry experience as a front office software engineer at an investment bank, and is frequently involved in industry-related research and consulting projects. Her research expertise includes Mathematical Finance, Financial Risk Management and Computational Finance, and her academic work has been published in Mathematical Finance, Finance & Stochastics, Quantitative Finance, Journal of Applied Probability, European Economic Review and many other academic journals. She is associate editor of “Quantitative Finance”, “Methodology and Computing in Applied Probability” and “Digital Finance” and co-director of the Germany Chapter of the Global Assocation of Risk Professionals (GARP). Natalie holds an M.Sc. in Computer Science from the University of Bonn, a Master’s degree in Banking & Finance from Frankfurt School, and a Ph.D. in Quantitative Finance from Frankfurt School.

Sami Alkhoury is Research assistant, HWR Berlin, project “Explainable AI for FinTech Risk Management” (since 07/2023).
Research focus: Explainable AI, LLM-based risk culture assessment, ML-based real estate valuation.
M.Sc. International Finance, HWR Berlin (2023)

Abstract

Rolf Poulsen says: I will talk about some things about implied volatility surfaces I don’t think I knew at last year’s MathFinance Conference — i.e. about a year ago. Ideally, things will go as 1-2-3, a-b-c:

(1) Good clean data: (a) A John & Yoko data set; (b) Express[ed] volatility, Dr. Jäckel vs. Dr. Schadner; (c) Bottlenecks, Achilles heels, and mistletoes

(2) When bad things happen to good data: (a) Counting is difficult; (b) Discounting is difficult; (c) A day late and a dollar short

(3) Models or not: (a) Lucky Feller;  (b) The big Heston trap is the little things; (c) Implied volatility surfaces are geometric objects

Biography

Over the last 25 years Rolf Poulsen has worked (researching and teaching) on a variety of topics in quantitative finance; for instance hedging of exotic options, exchange rate models and markets, and optimal mortgage choice. He will talk about these subjects to all who are interested – and some who aren’t.

Abstract

There are several reasons why the FX options business is diversifying. In many areas, it is becoming simpler or more standardized, meaning that primarily (a few) large players offer (banks) or execute (client) highly complex products.

Biography

Stefan has been working as a sales specialist for interest rate and foreign exchange derivatives for 25 years. The range of clients he and his team manages is very diverse within their area of responsibility, spanning from MiFID retail clients to multinational corporations. As a result, he encounters a wide variety of questions and challenges in the field of interest rate and foreign exchange management on a daily basis, each requiring different solutions and approaches. In his opinion, the trend in FX options for many clients and banks is moving more toward simpler structures.

Biography

Thorsten Schmidt is Professor for Mathematical Stochastics at University Freiburg and Senior Financial Engineer at MathFinance. From 2017-2019 he was fellow of the Freiburg institute of Advanced Studies (FRIAS). Prior to this he was professor for Mathematical Finance at Chemnitz University of Technology since 2008, held a replacement Professorship from Technical University Munich in 2008 and was Associate Professor at University of Leipzig from 2004 on. Besides many stays at leading universities, he was guest professor at ETH Zurich and at Université d’Evry.

His Ph.D. he obtained from University in Giessen in 2003 on credit risk. He is Editor of Statistics & Risk Modeling,

Associate Editor for Mathematical Finance, International Journal of Theoretical and Applied Finance and was Associate Editor for Journal of Banking and Finance and Statistical and Probability Letters. He is an elected member of the International Statistical Institute and is a member of the Board of the Bachelier Finance Society.

With numerous articles in the areas of Mathematical Finance and Probability in internationally leading journals and frequent presentations on conferences around the world, he is a well-known scientist in the area of affine models, interest rates, credit risk, incomplete information, risk management, filtering, and insurance mathematics. He has a strong background in statistics and information technology and teaches probability, mathematical finance and machine learning at the university of Freiburg.

Abstract

Perfect for early-career quantitative investors and analysts, traders, and market data professionals, The Quantamental Revolution is an essential read for portfolio managers interested in improving their investment processes. The engaging anecdotal vignettes coupled with academic rigor provide the reader with an authentic front row seat to the evolution of Quantamental investing on Wall Street over the past three decades.

In The Quantamental Revolution: Factor Investing in the Age of Machine Learning, veteran quantitative investor and strategist, Milind Sharma, delivers a comprehensive discussion of factor investing, risk premia, smart betas, multi-factor models and the deployment of ML ensembles towards monetizing alpha in the hedge fund world.

Biography

Milind Sharma’s ~30 years of market experience span running proprietary trading desks at Deutsche Bank (now Saba) & RBC (GAT entity) as well as hedge funds (QuantZ) & mutual funds (MLIM). His funds have won many awards over the years including those from Morningstar, Lipper, WSJ etc. He was a founding member of Quant Strategies at Merrill Lynch AM (now BlackRock) and Manager of the Risk Analytics and Research Group at Ernst & Young. His QuantZ/ QMIT entities have been a pioneer in applying ML to factor investing. QuantZ is the world’s 1^st fully synthetic hedge fund accessible via QIS swaps.

His publications have appeared in the Journal of Investment Management, Risk, Wiley, Elsevier, World Scientific etc in addition to a new Wiley book – “The Quantamental Revolution”. His educational background includes Oxford, Carnegie Mellon, Vassar, UWC & Wharton. He has guest-lectured and (or) run Capstone courses at Columbia, Carnegie Mellon, NYU, UCLA, UCSD etc.

He is founder & President of New York’s premier quant society <QWAFAxNEW>.

Abstract

The Climate Risk Brain is an institutional memory for climate-risk management and reporting. It combines agentic AI, a temporal knowledge graph, immutable evidence and governed risk models to connect counterparties, assets, hazards, scenarios, decisions and disclosures. The knowledge graph provides the semantic middleware needed to make AI outputs current, traceable and auditable rather than merely plausible. Its guiding principle is Data Integrity before Agentic Intelligence: agents collect evidence, detect inconsistencies and orchestrate workflows, while authoritative calculations remain in controlled analytical services. This is essential because, e.g., classification errors can propagate into exposure measures, capital requirements and climate-related disclosures. The result is a continuous, evidence-based climate-risk process in which the brain remembers, agents act, models calculate and humans govern.

Biography

Martin Simon is a Professor of Data Science at the Frankfurt University of Applied Sciences and serves as a Docent in Computational Finance at the Lappeenranta-Lahti University of Technology (LUT) in Finland. He holds a PhD in Numerical Mathematics from the University of Mainz and brings many years of industry experience to his academic roles. He is one of the founders of the Sustainable Finance Research Lab at the Frankfurt University of Applied Sciences, where he plays a pivotal role in advancing research on sustainable finance and climate risk. With a strong interdisciplinary background, his research interests span across data science, computational finance, and the application of advanced numerical methods in financial modeling.

Simon is involved in several international collaborations and research projects, particularly focusing on the integration of machine learning techniques into financial risk assessment and quantitative finance. In addition to his academic roles, he actively engages with the financial industry, providing expert insights on the challenges of implementing cutting-edge technologies like AI and machine learning in financial risk management.

His work is recognized both in academia and industry, where he actively bridges the gap between theoretical research and practical application. He is a regular speaker at international conferences, where he discusses topics like climate risk modeling and advances in computational finance.

Abstract

This talk presents a smart contract-based prototype for trading and settlement of derivatives on Ethereum. Users open collateralized long and short positions on-chain, while a market maker provides liquidity and hedges net exposure off-chain. The prototype is deliberately product-agnostic and supports a range of implementation choices, from direct price feeds and index construction to model-based pricing for derivatives. It also allows for configurable fee structures, exposure limits and access restrictions.

Biography

Jan Szyszko vel Chorazy, CFA is a Senior Consultant at WEPEX GmbH, a consulting firm focused on financial markets, where he works on technical, risk, and blockchain/digital assets topics for capital markets clients — drawing on a background in derivative pricing and risk model validation across a variety of asset classes.

Abstract

Local volatility (LV) and stochastic local volatility (SLV) models can be calibrated in such a way, that they reprice path-independent options and therefore fit plain vanilla options. However, they will produce different prices for path-dependent options. In foreign exchange (FX), one approach is to introduce a mixing factor to generate a new model that proportionally mixes LV and SLV according to this factor. The objective is to approximate the observed prices of path-dependent claims while ensuring that path- independent options are still fitted exactly.

To derive the parameters of an SLV model one usually calibrates a stochastic volatility model (SV) to plain vanilla options to some extent and then adds a local volatility component (the L in SLV) to achieve an (almost) exact fit to the vanilla price surface.

In this talk we will demonstrate how calibrating SV using certain methods can result in significantly different good fits to the vanilla price surface. In particular, quality may vary with the tenor. Therefore, the local volatility component in the corresponding SLV model may vary in severity, as it needs to compensate for the SV component’s lack to fit.

This will result in significantly different SLV model prices for path-dependent options, calling the concept of the mixing factor into question.

We will propose some methods to improve the SV calibration and in particular want to emphasize how important already a good SV calibration is to maintain a meaningful concept of mixing SLV and LV.

Biography

Andreas works as a financial engineer at MathFinance AG. He specialises in implementing numerical methods and stochastic models for the valuation of derivatives. and is particularly interested in collaborating with various teams to design and develop quantitative methods and libraries.

Biography

Uwe Wystup is managing director of MathFinance AG, Professor of Financial Option Price Modeling and Foreign Exchange Derivatives at University of Antwerp and Honorary Professor of Quantitative Finance at Frankfurt School of Finance & Management, certified expert for interest rate and currency risk management at Frankfurt’s Chamber of Commerce, honorary judge at Frankfurt Commercial Court and the State Financial Court. Before, he has actively worked in FX derivatives trading as Financial Engineer, Global Structured Risk Manager and Advisor since 1992, including Citibank, UBS, Sal. Oppenheim and Commerzbank. He is one of the few hybrids in the world working in the intersection of the derivates market and academic research. Uwe earned his PhD in mathematical finance from Carnegie Mellon University, is currently He acts internationally as expert witness and mediator in financial disputes. In his consulting work he has advised many leading banks and companies in the world including DBS, OCBC, Deka, GPS, Moneycorp, ICY Software, Murex, SuperDerivatives, Refinitiv, Pictet, La Caixa, DWS, Union Investment, NordLB, LBBW, AXA, Natixis, Reserve Bank of India and many others.

His first book Foreign Exchange Risk was published in 2002, quickly became the market standard and has also been translated into Mandarin. His second book FX and Structured Products appeared in 2006, followed by a second edition in 2017. Many of his papers appeared in scientific journals. See https://www.mathfinance.com/publications/ for publications.