I’m working with monthly copper returns (1990-2016) and the purpose is to obtain variance forecasts of the next month in a rolling windows framework of 1 step. My In-sample data takes returns since 1990M02 to 2006M11, so the out-sample data covers 2006M12 to 2016M11.
The problem is that obtained forecasts differ around ten times the monthly observed variance (which I defined commonly as the mean of quadratic difference between daily returns and the average of these within the month).
I programmed it, since I’m combining the garch model (with variance regressors) with a genetic algorithm. Both this and the GARCH(1,1) model exhibit the problem mentioned above.
Does someone know why that happens? I attached both files if it helps. I would be so grateful
pd: I'm using Eviews 8.1 version
For technical questions regarding estimation of single equations, systems, VARs, Factor analysis and State Space Models in EViews. General econometric questions and advice should go in the Econometric Discussions forum.
1 post • Page 1 of 1
Who is online
Users browsing this forum: No registered users and 2 guests