Posouzení modelů odhadu tržního rizika s využitím dea přístupu

Translated title of the contribution: Examination of market risk estimation models via DEA approach modelling

Aleš Kresta, Tomáš Tichý, Mehdi Toloo

Research output: Contribution to journalArticlepeer-review

1 Citation (Scopus)

Abstract

Measuring and managing of financial risks is an essential part of the management of financial institutions. The appropriate risk management should lead to an efficient allocation of available funds. Approaches based on Value at Risk measure have been used as a means for measuring market risk since the late 20th century, although regulators newly suggest to apply more complex method of Expected Shortfall. While evaluating models for market risk estimation based on Value at Risk is relatively simple and involves so-called backtesting procedure, in the case of Expected Shortfall we cannot apply similar procedure. In this article we therefore focus on an alternative method for comprehensive evaluation of VaR models at various significance levels by means of data envelopment analysis (DEA). This approach should lead to the adoption of the model which is also suitable in terms of the Expected Shortfall criterion. Based on the illustrative results from the US stock market we conclude that NIG model and historical simulation should be preferred to normal distribution and GARCH model. We can also recommend to estimate the parameters from the period slightly shorter than two years.

Translated title of the contributionExamination of market risk estimation models via DEA approach modelling
Original languageCzech
Pages (from-to)161-178
Number of pages18
JournalPoliticka Ekonomie
Volume65
Issue number2
DOIs
Publication statusPublished - 2017
Externally publishedYes

Keywords

  • Data envelopment analysis
  • Historical simulation
  • Market risk
  • Model quality
  • NIG
  • Value at risk

ASJC Scopus subject areas

  • Sociology and Political Science
  • Economics and Econometrics

Cite this