Forecasting of Constant Conditional Correlation Model
Posted: Thu Oct 25, 2012 5:09 am
Hi：
I am running Eviews 6. I have some problems about forecasting after I finished the parametric estimation of Constant Conditional Correlation Model.
I refered User's Guide, and tried to do the following:
1.From system object of finishing estimation > Proc > Make Model
2.In Model object > Solve
By these two steps, I can obtain the forecasting of mean equations.
I also need the forecasting of volatility equations, but Model object didn't do that.
I think I need the forecast values of out-of-sample to solve my problem, how could I do that?
Attaching Constant Conditional Correlation Model
mean equations:
rs(t)=c(1)+es(t)
rf(t)=c(2)+ef(t)
volatility equations:
hs(t)=m(1)+a(1)*es(t-1)+b(1)*hs(t-1)
hf(t)=m(2)+a(2)*ef(t-1)+b(2)*hf(t-1)
hsf(t)=cor(rs, rf)*sqrt(hs*hf)
Any advice would be greatly appreciated!
I am running Eviews 6. I have some problems about forecasting after I finished the parametric estimation of Constant Conditional Correlation Model.
I refered User's Guide, and tried to do the following:
1.From system object of finishing estimation > Proc > Make Model
2.In Model object > Solve
By these two steps, I can obtain the forecasting of mean equations.
I also need the forecasting of volatility equations, but Model object didn't do that.
I think I need the forecast values of out-of-sample to solve my problem, how could I do that?
Attaching Constant Conditional Correlation Model
mean equations:
rs(t)=c(1)+es(t)
rf(t)=c(2)+ef(t)
volatility equations:
hs(t)=m(1)+a(1)*es(t-1)+b(1)*hs(t-1)
hf(t)=m(2)+a(2)*ef(t-1)+b(2)*hf(t-1)
hsf(t)=cor(rs, rf)*sqrt(hs*hf)
Any advice would be greatly appreciated!