Bevezetés a sztochasztikus analízisbe közgazdászoknak (Introducing stochastic analysis)

A dolgozat célja, hogy rövid bevezetést adjon a folytonos idejű sztochasztikus analízisbe. A hazai pénzügyi oktatási gyakorlat nagyrészt a diszkrét idejű és gyakran diszkrét állapotterű modellekre épül. Ennek oka a folytonos időparaméterű sztochasztikus folyamatok elméletétől való érthető idegenkedés. A folytonos időparaméterű sztochasztikus analízis a modern matematika egyik csúcsteljesítménye, amely teljeskörű matematikai megértése egyrészt feltételezi, hogy az olvasó tisztában van a modern analízis szinte minden részletével; másrészt a matematikai részletek pontos megértése nem sok segítséget jelent a pénzügyi gondolatok elsajátításakor. / === / In the article we present a short, intuitive introduction to stochastic analysis. Our presentation is aimed for economist and we try to discuss only the most elementary properties of the stochastic analysis. Instead of precise proofs we present some simplified intuitive arguments. The central concept of the discussion is the quadratic variation and the Itō's lemma.

Analysis of operational risk of banks – catastrophe modelling

Nowadays financial institutions due to regulation and internal motivations care more intensively on their risks. Besides previously dominating market and credit risk new trend is to handle operational risk systematically. Operational risk is the risk of loss resulting from inadequate or failed internal processes, people and systems or from external events. First we show the basic features of operational risk and its modelling and regulatory approaches, and after we will analyse operational risk in an own developed simulation model framework. Our approach is based on the analysis of latent risk process instead of manifest risk process, which widely popular in risk literature. In our model the latent risk process is a stochastic risk process, so called Ornstein- Uhlenbeck process, which is a mean reversion process. In the model framework we define catastrophe as breach of a critical barrier by the process. We analyse the distributions of catastrophe frequency, severity and first time to hit, not only for single process, but for dual process as well. Based on our first results we could not falsify the Poisson feature of frequency, and long tail feature of severity. Distribution of “first time to hit” requires more sophisticated analysis. At the end of paper we examine advantages of simulation based forecasting, and finally we concluding with the possible, further research directions to be done in the future.