izpis_h1_title_alt

Modelska negotovost in izračun tvegane vrednosti : magistrsko delo
ID Bratanič, Tomaž (Author), ID Košir, Tomaž (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (1,59 MB)
MD5: 5934DF349493350E2B5FB617C34EF056

Abstract
Tvegana vrednost (angl. Value-at-Risk (VaR)) je standardna metrika tveganega kapitala v bančništvu in zavarovalništvu. V magistrskem delu predstavim algoritem, ki omogoča izračun numerične ocene za tvegano vrednost skupne izgube pri različnih scenarijih odvisnosti med posameznimi mejnimi tveganji. Algoritem omogoča izračun zanesljivih (ostrih) mej tveganih vrednosti tudi v portfeljih z več sto mejnimi tveganji. Predstavim pomembnejše analitične meje za skupno tvegano vrednost portfelja in povzamem ključne lastnosti mej izračunanih s pomočjo predstavljenega algoritma. Izkaže se, da dodatne informacije o pozitivni odvisnosti med pari faktorjev portfelja v splošnem občutno ne izboljšajo zgornjih mej. Nasprotno pa, ko so na voljo informacije o mejnih porazdelitvah višjega reda, vodijo do zelo izboljšanih mej. Na več primerih s praktično uporabnostjo pokažem, kako lahko s pomočjo opisanega algoritma določimo natančne meje za skupno tvegano vrednost portfelja. Te meje lahko razumemo kot modelsko negotovost, ki izhaja iz možnih scenarijev odvisnosti.

Language:Slovenian
Keywords:Kopula, Fréchetovi razredi, modelska negotovost, algoritem prerazporeditve, meje tveganih vrednosti
Work type:Master's thesis/paper
Typology:2.09 - Master's Thesis
Organization:FMF - Faculty of Mathematics and Physics
Year:2018
PID:20.500.12556/RUL-101097 This link opens in a new window
COBISS.SI-ID:18363225 This link opens in a new window
Publication date in RUL:26.04.2018
Views:1788
Downloads:550
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:English
Title:Model uncertainty and value at risk
Abstract:
Value-at-Risk (VaR) is the industry and regulatory standard for the calculation of risk capital in banking and insurance. I present a rearrangement algorithm that allows us to numerically estimate the VaR for a portfolio position as a function of different dependence scenarios on the factors of the portfolio. This algorithm allows us to compute reliable (sharp) bounds for VaR also in portfolios with hundreds of marginal risks. I describe the most important analytical bounds and provide a short discussion on their sharpness. Also, other properties of bounds given by the algorithm are presented. It turns out that additional information about positive dependence between pairs of marginal risks in general does not tighten the upper bound of possible VaR values given by the algorithm. On the other hand information about joint distribution of subsets of marginal distributions can lead to much tighter bounds. This bounds can be understood as the model uncertainty that results from the possible structures of dependencies between the different marginal risks of the portfolio.

Keywords:Copula, Fréchet classes, model uncertainty, rearrangement algorithm, Value-at-Risk

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back