The Krakow Quant Hub seminar was initiated in 2015. The list below summarised the talks that have been given on our seminar so far.
Please note that the views and opinions expressed during the seminar are the speakers' own and do not necessarily reflect the views and opinions of their current or past employers (the company/university name given next to the speaker is just for the reference purposes).
| Time | Speaker | Title & Abstract |
|---|---|---|
20.10.2020
|
Dariusz Gątarek (IBS PAN) |
Swap Rate a la Stock: Bermudan Swaptions Made Easy This talk will show how Markovian projection together with some clever parameter freezing can be used to reduce a full-edged local volatility interest rate model - such as Cheyette (1992) - to a "minimal" form in which the swap rate evolves essentially like a dividend-paying stock. Using a number of numerical examples, this talk will compare such a minimal "poor man's" model to a full-edged Cheyette local volatility model and the market benchmark Hull-White one-factor model. Numerical tests demonstrate that the "poor man's" model is in fact sufficient to price Bermudan interest rate swaptions. The main practical implication of this finding is that - once local volatility, dividend and short rate parameters are properly stripped from the volatility surface and interest rate curve - one can readily use the widely popular equity derivatives software for pricing exotic interest rate options such as Bermudans. |
14.01.2020 
|
Marcin Pitera (Jagiellonian University) |
Backtesting Expected Shortfall In this talk we will discuss a new approach to backtesting of Expected Shortfall that was recently proposed in article "Backtesting Expected Shortfall: a simple recipe" (Journal of Risk, Vol 22, No 1, pp 17-42). The test in based on a very simple and intuitive approach: instead of comparing projected risks with realised P&Ls, as done in VaR framework, we compare the aggregated projected risks with aggregated realised P&Ls in order to construct appropriate test statistic. Apart from discussing its statistical and mathematical properties, we show how to implement the backtest in R from scratch and run a couple of validation exercises on-the-fly. In particular, we compare the ES backtest proposal with the standard VaR breach test and discuss interconnection between these two frameworks. |
| 21.05.2019 | Ewa Marciniak (IMR, HSBC) |
Introduction to stochastic control: dividend optimization problems During the talk, the optimal dividend distribution problems for the companies with surplus-dependent premiums will be introduced. Two types of models will be concidered. The first one is suitable for the insurance companies for which it is assumed that they gain profits continuously and bear random losses as a result of claims. The second model (the so-called dual model) is considered to be more appropriate for the companies that incur continuous expenses related to running a business and gain its profits randomly. The control mechanism is based on the choice of the size and time of payments of the dividends. The goal is to maximize the expected value of the discounted dividends paid out up to the ruin time. The talk is based on the joint work with Z. Palmowski. |
| 16.04.2019 | Filippo Macaluso (State Street) |
How to Sample From a Distribution When Only the Moments are known with an application to Polynomial Processes We develop a novel (asymptotically)-exact efficient simulation-based procedure to sample from a multivariate distribution when only its moments are known and to price derivatives. To achieve this goal, we combine two strands of the statistical literature: The first one is concerned with the approximation of the density ratio of the original target to an auxiliary measure by orthonormal polynomial series in weighted L2-spaces, the second relates to simulation-based methods where the target density is not available and thus an approximation is used. Tests on densities implied from stochastic volatility models and tests on the general class of polynomial processes where the characteristic function is not available display that the proposed simulation scheme is computationally efficient and robust. The talk is based on a joint work with Antonietta Mira and Paul Schneider. |
| 26.02.2019 | Tadeusz Czernik (IMR, HSBC) | Fourier Transform technique applied to derivative pricing In this talk we present selected integral transform based approaches to option pricing. Two of them, namely Carr and Madan approach and Lewis approach, will be discussed in detail. We also show how to overcome a few drawbacks by introducing generalized functions. |
| 29.01.2019
|
Fabio Marelli (IMR, HSBC) | A trip through XVAs We introduce the three main valuation adjustment - CVA, DVA, and FVA - applied to the risk-free price of a derivative to account for counterparty risk and funding needs. Their mathematical formulation is introduced, together with practical modelling challanges and market practices used to mitigate their impact on derivative portfolios. We discuss some controversial effects of these adjustments and we look at them from an accounting and regulatory perspective. We conclude with a snapshot of the current US derivative market. |
| 27.11.2018
|
Damian Jelito (Institute of Mathematics, Jagiellonian University) | New normality test based on the conditional second moments and the 20-60-20 rule We introduce a new test of normality based on the so-called 20-60-20 rule. In a nutshell, the rule states that the conditional variances of normally distributed random variable are equal, when the conditioning is based on the 20-60-20 ratio. The new test has a clear financial interpretation and can be used to assess the impact of fat-tails on central data normality assumption. Using financial data we compared the new test with some well-known tools and obtained interesting results. We also derived asymptotic distribution of the test statistic. The talk is based on a joint work with Marcin Pitera. |
| 30.01.2018 | Igor Vexin |
Overnight Period at the Stock Market At the beginning I will familiarize the auditory with the stylized facts on the overnight return at the stock market. Then I will present my method of incorporating the overnight return into the estimation of the whole day covariance matrix. Finally, I will extend my method to the case of asynchronous markets. |
| 21.11.2017 | Rafal Muchorski (Senior specialist, Millennium Bank SA, Market Risk and Liquidity Team) |
A quasi Monte Carlo method and its applications in arbitrage-free pricing of selected exotic derivatives The author will present a novel approach for pricing European Basket options, and some other related derivatives, by applying a numerical technique based on adaptive sparse grids. Due to its semi-analitycal nature, the method allows for an efficient approximation of prices of multi-asset options for dimensions larger than 5, while preserving some flexibility of calibration to typicaly observed implied volatility surface for single and multi-asset options. The presentation will also include related aspects such as analytical error approximation, optimal choice of adaptive grids and some theoretical results for the analysis of index-type implied volatilities, associated with each Basket option. |
| 12.06.2017 | Laurent-Olivier Valigny |
2008 crisis, valuation uncertainty and risk management The presentation will highlight the main market events that firms have been exposed to since 2008 and will examine the consequences and new requirements in terms of pricing models, valuation of illiquid products, model and market risk. |
| 16.05.2017 | Pawel Morkisz (AGH University of Science and Technology, Faculty of Applied Mathematics) |
Optimal pointwise approximation of SDEs from inexact information Pointwise approximation of solutions of systems of stochastic differential equations will be presented. We assume that an approximation method can use values of the drift and diffusion coefficients which are perturbed by some deterministic noise. We also investigate lower bounds on the error of an arbitrary algorithm and establish optimality of the defined randomized Euler algorithm. Finally, we report some numerical results. |
| 04.04.2017 | dr Pawel Przybylowicz (AGH) |
On numerical approximation of solutions of stochastic differential equations with jumps. In the talk we will present introduction to numerical approximation of solutions of stochastic differential equations driven by Wiener and Poisson processes. We will discuss basic theoretical facts about SDEs with jumps (existence and properties of strong solutions) together with main techniques of obtaining approximation schemes. Some recent developments in the field will also be presented. Finally, we will report results of numerical experiments. |
| 10.01.2017 | dr Marcin Jaskowski (Erasmus University, Departament of Econometrics; HSBC - IMR) |
The puzzle of Cross-Sectional Dependence in Credit Spread Changes. In order to understand credit risk puzzle and the apparent segmentation of the stock market from the credit markets, we need to be able to assess the strength of cross-sectional dependence in credit spreads. This turns out to be a non-trivial task due to the extreme data sparsity typical for any panel of credit spreads extracted from corporate bond transactions. The problem of data sparsity led previous researchers to some erroneous conclusions which we point out. Understanding previous pitfalls is an important first step allowing us to provide a new estimate of the latent factor in credit spread changes and its characteristics. |
| 13.12.2016 | dr Marcin Pitera (Jagiellonian University) |
Optimising a risky portfolio on infinite time horizon - |
| 07.06.2015 | Grzegorz Goryl |
Correlated counterparty risk - |