Premium
Non parametric VaR Techniques. Myths and Realities
Author(s) -
BaroneAdesi Giovanni,
Giannopoulos Kostas
Publication year - 2001
Publication title -
economic notes
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.274
H-Index - 19
eISSN - 1468-0300
pISSN - 0391-5026
DOI - 10.1111/j.0391-5026.2001.00052.x
Subject(s) - bootstrapping (finance) , econometrics , parametric statistics , variance (accounting) , reliability (semiconductor) , computer science , covariance , value (mathematics) , value at risk , statistics , mathematics , machine learning , economics , risk management , power (physics) , physics , accounting , management , quantum mechanics
VaR (value‐at‐risk) estimates are currently based on two main techniques: the variance‐covariance approach or simulation. Statistical and computational problems affect the reliability of these techniques. We illustrate a new technique – filtered historical simulation (FHS) – designed to remedy some of the shortcomings of the simulation approach. We compare the estimates it produces with traditional bootstrapping estimates. (J.E.L.: G19).