Static n-steps ahead forecast with regressors

[martinla]

Hi everybody, I'm new in the forum, nice to meet you.

Let me first explain you what I am working on, so the questions will be more clear.

I have a model with an ARMA structure plus some factors, and I want to use it to do some forecasting.
In particular, I want to do 2, 4 and 12 steps ahead forecasts. For example, in may-00 I want to forecast jul-00, sep-00 and may-01. I want to do it for every month in the sample, using the information up to the month. So, for example, in may-00 I would have the dependent variable and the factors up to may-00.
For the evaluation of the model, I should not use the real factors, but forecast of them based on some model such as a VAR model.

So, the questions are:

1) Do you know how to run 2, 4 and 12 periods ahead forecast for each month in the sample, using information up to the month?

2) Do you know how to estimate the model with the factors, but doing the forecasting with some estimation of them?

Thanks in advance!

Martin