Forschung und Projekte

• random matrix generation and unified matrix decomposition
• equivariant topology, Mackey functors and induction theory
• topological data analysis and deep learning models
• impact of the interest rate benchmark reform on valuation models
• principals of esg markets
• credit Markov chains
• collateral allocation in multi loan structures
• market risk and systemic risk of crypto currencies
• model risk and epistemology

• Forum Finanzmathematik (internal)
• Forum Data Science (internal)
• DISO (Darmstadt Institute of Statistics and Operation Research)

• auxilium, project boilerplate for an automated testing and deployment
• businessdate, library for generating business dates and adding periods
• dcf, library for interest rate calculations and discounted cashflows
• shortrate, stochastic risk factor model library
• timewave, time evolution simulation engine

New Products and New Markets

• supporting implementation of exotic interest rate derivatives with risk model integration
• implied volatility smile risk methodology
• quanto clique Fx option with contingent exposure
• options on cross currency swaps in a convexity framework
• inflation products: year-to-year and zero-coupon swaps and floors
• autocallable CMS spread options
• perpetual bond pricing
• SABR volatility and implied rate correlation
• Bermudan swaptions

Pricing and Risk Library

• The framework offers a web service (rest api), product coverage of most popular interest rate and fx instruments as well as a simulation framework for derivative exposure calculation. Moreover, the latest loss loan provision models are available.

Model Risk Management and Risk Model Validation

• Establishing Principles for Model Inventories, Model Life Cycle Management and Model Roles and Responsibilities with Unified Model Validation Reporting
• defining model risk and model uncertainty metrics
• conducting various validation including validation of credit risk portfolio unexpected loss model and multiple market risk value-at-risk models

EMIR

• implementation of initial margining methodology for central clearing
• fire drill methodology for LCH compliance (central counter party clearing). On semiannual basis LCH clearing member must demonstrate their capability to participate in a simulated auction of defaulted member portfolio. The new fire drill process included various automated validation checks in order to ensure best practice valuations.

Hedge Accounting

• Economically hedging such interest rate risk with swaps is easy but the difference between derivative fair-value accounting and loan at-cost accounting makes it challenging. In order to minimise the gap between both accountings IAS 39 offers hedge accounting. Profit and loss of hedged risk position can be offset with the fair value profit and loss once a hedge is demonstrating tested effectiveness for each value date.
• Also micro hedge accounting, i.e. one-to-one hedge relation between derivative and bond (see below), have become a nontrivial task due to basis risk, hedging a contingent nonlinear portfolio of residential mortgages with cancelation tights is even more complex. I developed dynamic hedging strategies and accompanying robust valuation methodologies for hedge effectiveness testing.

IFRS 9

• Development and agile implementation of stochastic default models by credit Markov chains with macro economic forecast by real business cycle theories and firm value factor models and conditional recovery and resolution models with stochastic collateral value evolution.

IFRS 13

• Starting from a comprehensive list of infecting clause types, we developed for each type a booking structure and valuation methodology. Some them are straight forward, like caps, floors or convexity adjustments, others, mainly idiosyncratic clauses, require complete new modelling ideas.
• fair value adjustments (CVA/DVA/FVA)

Negative Interest Rates

• Admissible models for plain vanilla option were know for a long time, e.g. Bachelier or shifted Black76, others had to be improved (e.g. arbitrage free SABR) or enjoy a renaissance like the HullWhite short rate model.
• Beside implementing those models in risk engines further challenges came along by hidden zero strike floors such as in floating loans or collateral agreements. We implemented models for negative rates, redesigned all calibration workflows for option pricing, exposure simulation and validated the impact of CSA floors. Moreover, I contributed to opinion forming discussions how to account zero strike floors.

OIS Discounting

• introduction of a multi curve interst rate framework and impield volatility calibration

Operational Risk

• Extensive data analysis and clustering with respect to time dependence provided aggregated data sets which meets the core assumption from extreme value theory. In addition we introduced infective disease modelling to provide better frequencies predictions of future events.
• For value-at-risk calculations the single loss approximation approach is extended using copulas to a multi entity loss distribution model.