EViews.com

Could you please help me interpreting or maybe fixing the problem of non-normal distribution of prediction errors (measured by RMSE of out-of-sample forecast errors)?

Thanks!