Heteroskedasticity in a VEC model, adding robust standard errors and plotting forecasts

I'm dealing with a data set which required me to log and take first differences of it to induce stationary. A VECM as a result was prodouced.

Everything seemed fine until I checked for hetroskedasticity using arch.test command in R and found that it was present.

How do I add factor in robust standard errors into my forecasts? More specifically, what should my code look like?

• Sorry, but asking for code is off topic. I would gladly supply the code at the end of my answer if the code was sufficiently short and simple, but it isn't (well, it is not difficult either, but a bit more tedious than I would like to get involved now). – Richard Hardy Oct 23 '17 at 16:17

The arch.test tells you the residuals display autoregressive conditional heteroskedasticity. To properly account for it, you can model the conditional variance of your time series with a multivariate GARCH model, e.g. BEKK-GARCH or DCC-GARCH; the latter is available in the rmgarch package in R. There you may
• specify (dccspec) the conditional mean model as VAR (knowing that VECM has an equivalent representation as VAR) alongside the conditional variance model;
• fit the model (dccfit); and
• produce forecasts from them in an automated way (dccforecast).
Check out the rmgarch vignette (p. 3-6) and the help files.