Document Search
Add To My Bookshelf Sign in or Register Save And Annotate
Keywords:

BEKK model sign in to follow this
equities sign in to follow this
autocorrelation sign in to follow this
Dow Jones sign in to follow this
Categories:

VaR Methods sign in to follow this
--Evaluation/Comparison sign in to follow this
Half-Life:
Impact:
Discuss This Paper
Sign in to follow this page
Recent Comments
  more
Construction of Value-at-Risk Forecasts Under Different Distributional Assumptions within a BEKK Framework
Year Of Publication: 2014
Month Of Publication: May
Resource Link: Click here to open
Pages: 51
Download Count: 0
View Count: 1305
Comment Num: 0
Language: English
Source: working paper
Who Can Read: Free
Date: 7-26-2014
Publisher: Administrator
Summary
In order to construct Value-at-Risk (VaR) forecasts that take into account the serial autocorrelation and non-normality observed in financial returns, we combine a scalar BEKK (sBEKK) model with different assumptions on the returns distribution. The set of distributions we consider comprises the Normal, Student, MEP and their skewed equivalents, obtained via a transformation function. Using a sample of 10 assets from the Dow Jones Industrial Average over a period of eight years, we estimate the parameters of the sBEKK model via Maximum Likelihood and then compute a series of 700 one-step ahead forecasts of the conditional covariance matrix using a rolling fixed-window scheme. Equally-weighted portfolios are then constructed to compute the out-of-sample VaR. The accuracy of the VaR forecasts is assessed for each distribution by implementing a series of statistical backtesting procedures. The results unanimously show that inclusion of heavy-tails into the distributional specification yields more accurate VaR forecasts. By contrast, the use heavy tailed and skewed distributions does not result in significant improvements in the VaR acc
Author(s)
Braione, Manuela Sign in to follow this author
Scholtes, Nicolas K. Sign in to follow this author
This document's citation network:
Similar Documents:
Close window
Sign up in one step, no personal information required. Already a Member?



Email:
Repeat Email:
User Name:
Password:
Confirm Password:

Sign Up


Welcome to GloriaMundi!
Thanks for singning up



continue or edit your profile