Forecasting with VECM and Random walk

[ali]

I have to compare out-of-sample forecasts of VECM and random walk models. How can estimate random walk model in Eviews? I need simple driftless random walk.
And second, how to implement out-of-sample forecasts using VECM? I tryed to do the model in Procs menu and then solve it, but it makes just 1 step of the forecasts...
Also, I need to estimare root mean square error from these forecasts...
Can you help plz?

[trubador]

How can estimate random walk model in Eviews?

Search the forum please.

how to implement out-of-sample forecasts using VECM?

Try "Dynamic Solution".

I need to estimare root mean square error from these forecasts

This code might be helpful.

[ali]

Still do not understand.
If I need to implement 4, 10, 16, 20 step ahead forecasts, how can I specify it in the solve menu?

[trubador]

You need to adjust the sample size accordingly. Please read the help files and/or search the forum for detailed explanations and examples.

[ali]

Does it mean, if, for example I need to implement 4 step aheap forecasts for weekly observations with the in-sample series ending 31.12.2006, I must define forecast sample as 7.1.2005 to 28.1.2005?