Log in

Login to your account

Username *
Password *
Remember Me

MathFinance Conference 2017

20th and 21st April 2017, Frankfurt

MathFinance Conference has been successfully running since 2000 and has become one of the top quant events of the year. The conference is specifically designed for practitioners in the areas of trading, quantitative and derivatives research, risk and asset management, insurance, as well as academics.

As always, we expect around 100 delegates both from the academia and the industry. This ensures a unique networking opportunity which should not be missed. A blend of world renowned speakers ensure that a variety of topics and issues of immediate importance are covered.

This event is a must for everyone in the quantitative financial industry.

Details of the event will follow soon.

Please click here for registration (single / group)

Please check out the confirmed speakers.

We want to thank our sponsors d-fine, Deloitte and fintegral.

d fine logo

 

 fintegral logo neu

deloitte schwarz

MathFinance Conference 2016

21st and 22nd of March 2016 / Frankfurt am Main, Germany

 

 

conf2016

Alexander Kabanov stressed that a life insurance contract is the most difficult complex derivative. We learned that discounting uses the so-called Ultimate Forward Rate (UFR), a relatively high interest rate that assumes a scenario that interest rates to go higher again over the next decades.

Bruno Dupire showed us how to trade estimates of historical volatility. Hot news: The usual estimates based on high and low are not tradable. Surprisingly, it is not because the historic maximum and minimum are not stopping times but because they do not depend quadratically on the final value. He then introduced a new high and low based estimate that is tradable and unbiased.

Michèle Vanmaele presented locally risk-minimizing (LRM) strategies. In a jump diffusion model, one can either use a replication of a European style derivative contract that minimizes the variance of the hedging error at maturity. The alternative is to replicate the payoff exactly, but inject cash on the way. The latter is the LRM strategy. Compared to Merton this requires more frequent rehedging towards the end. Michèle uses Föllmer-Schweizer decomposition.

Julie Othman reduced the dimensions of the auxiliary variables required to value a volatility swap to one, which allows a fast computation in a PDE approach, also for Greeks.

Andreas Pfadler showed us how to efficiently calculate CVA sensitivities using Adjoint Algorithmic Differentiation (AAD). This can be applied to a portfolio of interest rate swaps. I am happy to learn that teaching American Monte Carlo is still highly useful. AAD improves up the calculation of sensitivities, but obviously requires the user to have access to the valuation source code. It doesn’t work with a black box.

Christoph Becker examined what drives the average correlation of stock returns. For the average he worked with equi-correlation and block-correlation matrices. He incorporated the Fama-French portfolios (High minus Low, Big minus Small, Winner minus Losers) and showed that with two driving factors equity correlations can be statistically explained in various market regimes with an R² of 90%. Christoph’s approach can immediately be applied to stressed Value-at-Risk calculations, as in fact required by banking supervision. It means that in stressed markets volatilities and correlations should be stressed systematically along with general market uncertainty.

Jürgen Hakala this time considered negative interest rates and the question how to realistically exploit cash-arbitrage. First of all, market rates don’t always follow the central bank rate. Secondly, cash as arbitrage has a number of practical obstacles. While a 1m³ safe could contain 6000 million CHF in 1000 CHF bills, there is only a limited amount of cash. Jürgen points out that we could soon experience an exchange rate between cash money and electronic money by letting cash expire gradually. Furthermore, he asked how low can interest rates go and how to choose a potential lower boundary of negative interest rates. Central banks are currently testing the ground to lowering rates further into negative territory. This can be achieved by removing cash or gradually expiring cash. For derivatives markets, a change of a model to negative rates needs to be approved by the supervisors, which can take 18 months. The unanswered key question is obviously what to do in between.

Manuel Wittke and Patrick Büchel report that new regulatory requirements jack up capital requirements. Sensitivities begin to matter, even in the Fundamental Review of the Trading Book (FRTB). They presented a Cheyette framework with stochastic volatility: the Tolle-Schwartz model.

Donging Qu simplified the interest rate smile by transforming a forward-model to a spot model via Girsanov change of measure, which allows for the first time to derive a local volatility model for Interest Rates. He then derived a local volatility formula in the normal Bachelier model. Once the interest rate smile is transformed to a model in the spot all the advanced techniques for local volatility modelling can be recycled from foreign exchange and equity derivatives.

Mats Kjaer presented multi-currency FVA (Funding Value Adjustment). He stressed that cross currency basis spreads are not Brownian motion-like, but more like electricity, as they spike in crises. He gets a systematic result taking all the trimmings into account a practitioner faces.

Adil Reghaï noted that quant investment strategies require a good quantitative risk management strategy. The question of which distribution financial assets follow has been answered: they are q-Gaussian. He uses a Hidden Markov Model (HMM) to implement an Advanced Risk Perception Indicator (ARPI); based on this there are investment funds that shift their assets based on ARPI; the indicator had already been public before the 2008 financial crisis.

Yuri Greenfield presented a riddle how to optimally design a portfolio of corporate loans, assuming the loans are ordered by quality. In case the goal is to minimize portfolio variance finding the optimal selection of loans is easy. However, if the goal is to find the best quantile - even in the simplified zero-correlation case – then it requires results from combinatorical probability theory yet to be discovered. Yuri points out that Markovitz portfolio optimization didn’t really help him, because the default of a loan is a digital event.

Attilio Meucci on the other hand is again working on portfolio optimization assuming the portfolio has tradable instruments such as stocks. He added further extensions to view and means by allowing views on mean, variances and correlation matrices. These included views also allow a term structure. He derived a systematic formula in the frame-work of a Multi-variate Ornstein-Uhlenbeck process. It has all the bells and whistles. The only point left to argue about is the normal return assumption, both in the prior and the posterior distribution, especially, after Adil’s presentation.

We want to thank you our sponsors d-fine, Delotte and fintegral.

We hope to welcome you next year for another round of an exciting MathFinance Conference 2017!

Go to top