Market risk under control

10 December 2014

Marcin Chlebus.jpgDecember 3rd Marcin Chlebus defended his doctoral dissertation ?Market risk measurement using Value-at-Risk ? two-step approach? at the Faculty of Economic Sciences at University of Warsaw. The thesis attempts to answer the question how the market risk can be effectively estimated using econometric models. One of the recommendations is to adapt tools and models to the state of the analyzed portfolio: different when at tranquillity, different when in time of turbulence.

Current regulations[1] do not impose any particular way of predicting the market risk, however, they require acting in accordance to a vast range of qualitative and quantitative standards. The most important requirement is calculating market risk using Value at Risk (VaR) which needs to be calculated with 99% of confidence. Short-term analysis and high norm of confidence are quite demanding requisites and as a result they may generate statistical fiction in prognosis. Thus the main challenge is to discover ways to find oneself in this space of abstraction.

Marcin Chlebus examined data of 79 companies enlisted in the Warsaw Stock Exchange and tested various prognosis models (2645) to determine the most effective ways of predicting and managing market risk. One of the main conclusions is to resign from universal solutions. Chlebus proposes that in the first step analysts should diagnose the state of the portfolio?s market (a state of tranquility or turbulence), and in the next one select the right VaR model.

What is more, Marcin Chlebus investigates adequacy of the VaR measurements in reference to a cost of model usage. The author emphasizes costs associated with the selection of the wrong way of predicting market risk, including: underestimation of the prognosis and the cost of maintaining excess capital where conservative strategy is implemented. The main difficulty is to handle and awaken risk appetite at the same time. All this while maintaining control over the symptoms of the upcoming state of turbulence and in accordance to the banking supervision standards. What is the solution? In emerging markets in a state of tranquility the model GARCH (1,1) will be suitable. In state of turbulence models referring to the Extreme Value Theory which take into account the empirical distributions of return rates will be more adequate. Thanks to the modulation of the models an increased compatibility without unnecessarily increasing costs is possible.

[1] Basel III (2010), CRD IV/CRR in EU.


Marcin Chlebus ? PhD at the Faculty of Economic Sciences who specializes in the development of models used to measure different types of risk (operational risk, market risk, credit risk). His interest is focused on advanced risk measurement methods (internal models for market risk, models consistent with the AMA approach, models of credit risk parameters PD, LGD, EAD). He has experience in this area, both from the perspective of academic and professional.

The PhD thesis has been prepared under the guidance of prof. UW dr. hab. Ryszard Kokoszczyński. Reviewers: prof. NCU dr hab. Piotr Fiszeder, prof. UW dr hab.  Marian Wisniewski.

The full text of the doctoral thesis available after contact with the author: Pwq[8hsZ-YH7DzSJx3!riE]#[=[ZHxKiNHQ3wMpKS^ru{Z.

Services * use cookies stored on users' computers. They are necessary for the proper operation of mechanisms websites Faculty of Economic Sciences.

Detailed information on cookies Operating Policy, University of Warsaw in services are available on