News impact curve?
Thanks Matthieu... Just a small correction, in the latest rgarch version, the spec object goes first (before data) in the fitting and filtering routines, else should be named, i.e. garch_norm = ugarchfit(data = dmbp[,1], spec = spec_garch_norm) fit_gjr = ugarchfit(data = dmbp[,1],spec = spec_garch_gjr) Regards, Alexios
On 8/30/2010 2:58 PM, Matthieu Stigler wrote:
Excellent! Exactly what I was looking for... thanks!! Are you planing to
put your excellent rgarch package on CRAN once? Hopefully!
So a small example for people in the list in case interested in the answer:
library(rgarch)
data(dmbp)
spec_garch_norm<-ugarchspec(variance.model = list(model = "sGARCH",
garchOrder = c(1, 1)))
garch_norm<-ugarchfit(dmbp[,1],
spec_garch_norm,solver.control=list(trace=FALSE))
spec_garch_gjr<-ugarchspec(variance.model = list(model = "gjrGARCH",
garchOrder = c(1, 1)))
fit_gjr = ugarchfit(dmbp[,1],spec_garch_gjr)
ni=newsimpact(z = NULL, garch_norm)
ni2=newsimpact(z = NULL, fit_gjr)
plot(ni$zx, ni$zy, ylab=ni$yexpr, xlab=ni$xexpr, type="l", main = "News
Impact Curve")
lines(ni2$zx, ni2$zy, lty=2, col=2)
legend("topleft", leg=c("GARCH", "gjr-GARCH"), lty=c(1,2), col=1:2,
bg="white")
Thanks!
Le 30. 08. 10 15:09, alexios a ?crit :
The function 'newsimpact' in the rgarch package will return the values needed to plot the this, given a fitted object (alternatively use plot on the fitted object and choose option 12). Regards, Alexios Ghalanos On 8/30/2010 1:51 PM, Matthieu Stigler wrote:
Hi
I searched for a function in R to compute the news impact curve of a
GARCH (showing the effect on h_{t-1} of a positive/negative shock of
e_t), but did not find...
Before I do it manually... is anyone aware of such feature in R?
Thanks!
Matthieu
_______________________________________________ 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. -- Also note that this is not the r-help list where general R questions should go.