Differences between univariate and multifit + p-values in output

[krauuuus]

I've been using rugarch to fit univariate GARCH models on assets with external regressors. The spec is this one :

spec<-ugarchspec(variance.model=list(model="eGARCH",
                                     garchOrder=c(1,1), external.regressors=as.matrix(all3$dummy)), 
                 mean.model=list(armaOrder=c(0,0), include.mean=T), 
                 distribution.model="sstd")

Then, I've fitted to my data. First, only on one series (here, BTC), then on the multivariate data object

fit<-ugarchfit(spec, returns1[,1])

mspec = multispec( replicate(spec, n = 3) )
fitlist = multifit(multispec = mspec, data = returns[,1:3])

The coefficients are different for the first column. Is there a reason that I haven't seen or is it normal ? Also, in fitlist , I can't figure out where are the p-values, do you have an idea ?

[alexiosg]

I can't replicate this. Here's a minimally reproducible example:

dji <- as.xts(dji30ret)
spec <- ugarchspec(variance.model=list(model="eGARCH", garchOrder=c(1,1), external.regressors = coredata(dji[,10])),
mean.model=list(armaOrder = c(0,0), include.mean = T),
distribution.model = "sstd")
mod <- ugarchfit(spec, dji[,1], solver = "hybrid")
multis <- multispec(replicate(3, spec))
multifit <- multifit(multis, data = dji[,1:3], solver = "hybrid")
coef(multifit)[,1] == coef(mod)
multifit@fit[[1]]@fit$matcoef

[alexiosg]

Closing this.