Out-of-Sample statistical forecast accuracy Statistics

[ehabyamani]

Greetings,

I am running out-of sample forecast test in SAS using PROC UCM, using the following code:

PROC UCM DATA = fx.developedroll NOPRINT; BY country beg_date end_date;ID DATE INTERVAL=MONTH; 
MODEL FXret = ret;IRREGULAR;level var=0;
estimate back=1 outest=OOSRollEst1_BOret; *estimation span;
FORECAST back=1 LEAD=1 outfor=fx.OOSRollRes1_ret plot=forecasts;  run;

However, I don't know how to find/calculate the following statistics from the UCM 'statistics of fit' output:

1. Clark and West (2007) test

2. Giacomini and White (2006) test

3. and Giacomini and Rossi (2010) test

Thanks in advance

[rselukar]

I support the UCM procedure.  I am unfamiliar with these references.  I will go over them and see if you might be able to compute these test statistics using the UCM output.  Of course, this will take some time.  In the mean time if someone already knows about these matters, please respond.

[ehabyamani]

I appreciate your willingness to help. I tired several codes but I wan't ab;e to figure it out. I will really appreciate your help.

[ehabyamani]

Greetings,

I am running out-of sample forecast test in SAS using PROC UCM, using the following code:

PROC UCM DATA = fx.developedroll NOPRINT; BY country beg_date end_date;ID DATE INTERVAL=MONTH; 
MODEL FXret = ret;IRREGULAR;level var=0;
estimate back=1 outest=OOSRollEst1_BOret; *estimation span;
FORECAST back=1 LEAD=1 outfor=fx.OOSRollRes1_ret plot=forecasts;  run;

However, I don't know how to find/calculate the following statistics from the UCM 'statistics of fit' output:

1. Clark and West (2007) test

2. Giacomini and White (2006) test

3. and Giacomini and Rossi (2010) test