Editing Value at risk (section)

== Computation methods ==

VaR can be estimated either parametrically (for example, [[variance]]-[[covariance]] VaR or [[Greeks (finance)#Delta|delta]]-[[Greeks (finance)#Gamma|gamma]] VaR) or nonparametrically (for examples, historical [[simulation]] VaR or [[Resampling (statistics)|resampled]] VaR).<ref name="Dowd" /><ref name="Unbearable" /> Nonparametric methods of VaR estimation are discussed in Markovich<ref name="Markovich">{{Citation|last=Markovich|first=N.|title=Nonparametric analysis of univariate heavy-tailed data|publisher=Wiley|year=2007}}</ref> and Novak.<ref name="Novak">{{cite book|last=Novak|first=S.Y.|title=Extreme value methods with applications to finance|publisher=Chapman & Hall/CRC Press|year=2011|isbn=978-1-4398-3574-6}}</ref> A comparison of a number of strategies for VaR prediction is given in Kuester et al.<ref>{{cite journal|last1=Kuester|first1=Keith|last2=Mittnik|first2=Stefan|last3=Paolella|first3=Marc |author2-link=Stefan Mittnik |title=Value-at-Risk Prediction: A Comparison of Alternative Strategies |journal=Journal of Financial Econometrics |year=2006 |volume=4 |pages=53–89 |doi=10.1093/jjfinec/nbj002|doi-access=free }}</ref>

A McKinsey report<ref name="McKinsey">{{cite web|title=McKinsey Working Papers on Risk, Number 32|author=McKinsey & Company|url=http://www.mckinsey.com/~/media/McKinsey/dotcom/client_service/Risk/Working%20papers/Working_Papers_on_Risk_32.ashx|format=pdf}}</ref> published in May 2012 estimated that 85% of large banks were using [[historical simulation (finance)|historical simulation]]. The other 15% used [[Monte Carlo methods in finance|Monte Carlo methods]] (often applying [[Principal_component_analysis#Quantitative_finance|a PCA decomposition]]) .