EViews.com

The only way I can think of is to use bootstraped standard errors by resampling the data using @resample. Things are complicated if you have a time series instead of a cross section. However, all this has to be done "by hand". It would be nice to have a "bootstrap standard errors" option implemented in future versions. In fact, some recent developments in this area (Vinod & Lopez-de-Lacalle, 2009) look most promising, as this bootstrap can be used on every kind of data including CS, TS, (unbalanced) panels etc.

Reference: Vinod, H.D., and Lopez-de-Lacalle, J. (2009), "Maximum Entropy Bootstrap for Time Series", Journal of Statistical Software, 29(5).

As to the idea of bootstrapping time-series data, my (and this is solely "my") opinion is that the literature on bootstrapping dependent data hasn't really settled down yet and that while there are lots of techniques (block bootstrap, etc. ) there is not yet a generally accepted canon...

But we will keep our eye on things. As always, reference like the one you provide are useful in shaping our thinking.

hi, i would argue that non normalty of resids signals poor specification of the model (Modern econometrics, R.L.Thomas and others)

Why on earth should error terms be normal? (Unless the errors are averages of other underlying processes)