[R-SIG-Finance] pointers on using VaR.gpd with return series?

[Sylvain BARTHELEMY]

I think that there is a problem with the VaR.gpd function, as it works on
USD/EUR and not on EUR/USD values

library(PerformanceAnalytics)
eurusd <- get.hist.quote("EUR/USD", provider="oanda", start = "2006-01-01")
usdeur <- get.hist.quote("USD/EUR", provider="oanda", start = "2006-01-01")

library(VaR)
v1 <- VaR.gpd(as.vector(eurusd))
v2 <- VaR.gpd(as.vector(usdeur))


output:

> v1 <- VaR.gpd(as.vector(eurusd))
Error in optim(init, gpd.liklhd, hessian = TRUE, method = "Nelder-Mead") :
         non-finite finite-difference value [2]
In addition: There were 50 or more warnings (use warnings() to see the 
first 50)

> v2 <- VaR.gpd(as.vector(usdeur))
There were 11 warnings (use warnings() to see)

---
Sylvain Barthélémy
Research Director, TAC
www.tac-financial.com | www.sylbarth.com


-----Message d'origine-----
De : r-sig-finance-bounces at stat.math.ethz.ch
[mailto:r-sig-finance-bounces at stat.math.ethz.ch] De la part de Brian G.
Peterson
Envoyé : mardi 21 août 2007 14:38
À : R-SIG-Finance
Objet : [R-SIG-Finance] pointers on using VaR.gpd with return series?

Does anyone have any hints on utilizing VaR.gpd on return series instead 
of price series?

I have tried converting a return series to a wealth index (using 
cumprod), but this still seems to cause problems with the VaR package.

for example:

library(PerformanceAnalytics)
data(edhec)
wi<-cumprod.column(1+edhec)
library(VaR)
vt<-VaR.gpd(wi[,1])

Error in optim(init, gpd.liklhd, hessian = TRUE, method = "Nelder-Mead") :
         non-finite finite-difference value [2]
In addition: There were 50 or more warnings (use warnings() to see the 
first 50)

A message to the maintainer of the package has without response.

If I don't get any usable feedback, I'll probably move to using fit.GPD 
from QRMlib (which I have had some luck with in the past) to add a 
general Pareto method to the VaR functions in PerformanceAnalytics, as I 
feel that parametric VaR functions on a broader set of distributions and 
copulae should be more widely available.

Regards,

    - Brian

_______________________________________________
R-SIG-Finance at stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-finance
-- Subscriber-posting only. 
-- If you want to post, subscribe first.