Eviews gives different number of iterations for same model??

[EViews Glenn]

Can you update your copy of EViews 8 so we are looking at the same thing? You're over a year out-of-date.

[strypste]

I am using the University server to access EViews and they do not allow me to update. I will contact the IT department to ask about that.. I'll keep you posted.

[EViews Glenn]

I'm having trouble getting the programs to run as there is a lot of system specific stuff there that doesn't run on my computer. Can you try to strip it down so that it's not doing the file saves and the like?

[strypste]

I commented all these bits out.. If you put all the programs + workfile in one folder, then it should run..

[EViews Glenn]

Program 2 still gives some errors. As to program 1, it was running for a couple of hours before I stopped it. Should it take that long?

[strypste]

To reduce the run time, you could set line 95 (the sample loop) to, for example, "for !f = 30 to 50". Then it will only do the estimations for sample period 30 till 50.

With respect to program two, I downloaded it on my computer and it runs ok.. What error do you get? Did you set "Maximum numbers before halting" to a high number?

[strypste]

Just to let you know that my EViews is now updated to Standard edition - May 13 2014 build.

The problem, however, is not sorted.

For example, when I run the program for France and sample 45 and I run the program for France for samlpe 43 to 45 (in a loop), then I get different results (i.e. different number is my matrix "bestmodel_fr45 and different number of iterations in the objects "best3_fr45")..

Any news on your end?

Regards
S

[EViews Glenn]

Code: Select all

0 is not a valid index for vector-series-coefficient ARLAGS in "BESTMODEL_{!F}(2,1) = ARLAGS(BESTMODEL_45(1,1))". 0 is not a valid index for vector-series-coefficient BEST_INITIAL_FR in "BESTMODEL_{!F}(3,1) = BEST_INITIAL_FR(BESTMODEL_45(1,1) ... FR_AR0__0 is not defined in "FREEZE(BEST3_FR45) FR_AR0__0". FR_AR0__0 is not defined in "DO_ FR_AR0__0.MAKERESIDS RESIDS_FR". FR_AR0__0 is not defined in "DO_ FR_AR0__0.MAKERESIDS(S) RESIDS_STAND_FR". RESIDS_STAND_FR is not defined in "FREEZE(TAB0) RESIDS_STAND_FR.STATS". RESIDS_STAND_FR is not defined in "FREEZE(TAB1) RESIDS_STAND_FR.TESTSTAT(MEAN=0, VARIANCE=1)". TAB0 is not defined in "RES_DIAG(1) = @VAL(TAB0(6,2))". TAB1 is not defined in "RES_DIAG(2) = @VAL(TAB1(11,4))". TAB0 is not defined in "RES_DIAG(3) = @VAL(TAB0(10,2))*@VAL(TAB0(10,2))". TAB1 is not defined in "RES_DIAG(4) = @VAL(TAB1(18,4))". TAB0 is not defined in "DELETE TAB0". TAB1 is not defined in "DELETE TAB1". FR_AR0__0 is not defined in "FREEZE(TAB1) FR_AR0__0.CORREL(10)". TAB1 is not defined in "RES_DIAG(5) = @VAL(TAB1(9,6))". TAB1 is not defined in "RES_DIAG(6) = @VAL(TAB1(9,7))". TAB1 is not defined in "RES_DIAG(7) = @VAL(TAB1(15,6))". TAB1 is not defined in "RES_DIAG(8) = @VAL(TAB1(15,7))". TAB1 is not defined in "DELETE TAB1". FR_AR0__0 is not defined in "FREEZE(TAB1) FR_AR0__0.CORRELSQ(10)". TAB1 is not defined in "RES_DIAG(9) = @VAL(TAB1(9,6))". TAB1 is not defined in "RES_DIAG(10) = @VAL(TAB1(9,7))". TAB1 is not defined in "RES_DIAG(11) = @VAL(TAB1(15,6))". TAB1 is not defined in "RES_DIAG(12) = @VAL(TAB1(15,7))". TAB1 is not defined in "DELETE TAB1". FR_AR0__0 is not defined in "FREEZE(TAB1) FR_AR0__0.ARCHTEST(3)". TAB1 is not defined in "RES_DIAG(13) = @VAL(TAB1(4,2))". TAB1 is not defined in "RES_DIAG(14) = @VAL(TAB1(4,5))". TAB1 is not defined in "DELETE TAB1". FR_AR0__0 is not defined in "FREEZE(TAB1) FR_AR0__0.ARCHTEST(9)". TAB1 is not defined in "RES_DIAG(15) = @VAL(TAB1(4,2))". TAB1 is not defined in "RES_DIAG(16) = @VAL(TAB1(4,5))". TAB1 is not defined in "DELETE TAB1". RESIDS_STAND_FR is not defined in "DELETE RESIDS_STAND_FR". FR_AR0__0 is not defined in "DO_ FR_AR0__0.FORECAST FCAST". FCAST is not defined in "M{!MODEL}_FCAST_ERR_{%COUNTRY}(45,1) = G_FR(@DTOO("2008M09")) - FCAST(@DTOO("2008M09"))". FCAST is not defined in "M{!MODEL}_FCAST_ERR_{%COUNTRY}(45,2) = G_FR(@DTOO("2008M09")+2) - FCAST(@DTOO("2008M09")+2)". FCAST is not defined in "M{!MODEL}_FCAST_ERR_{%COUNTRY}(45,3) = G_FR(@DTOO("2008M09")+5) - FCAST(@DTOO("2008M09")+5)". FCAST is not defined in "M{!MODEL}_FCAST_ERR_{%COUNTRY}(45,4) = G_FR(@DTOO("2008M09")+8) - FCAST(@DTOO("2008M09")+8)". FCAST is not defined in "M{!MODEL}_FCAST_ERR_{%COUNTRY}(45,5) = G_FR(@DTOO("2008M09")+11) - FCAST(@DTOO("2008M09")+11)". FCAST is not defined in "DELETE FCAST". FR_AR0__0 is not defined in "VECTOR ESTIMATES_{%COUNTRY}_M{!MODEL}_{!F} = FR_AR0__0.@COEFS". ESTIMATES_FR_M3_45 is not defined in "ESTI_PARA_{%COUNTRY}_M{!MODEL}_{!F}(1) = @COLUMNS(@TRANSPOSE(ESTIMATES_FR_M3_45))".

[strypste]

??? I don't get these errors? What did you change in the program? Did you set !f equal to 0 or to _0?

[EViews Glenn]

Just to update you. Restored programs to originals and ran. Identical results in bestmodel_45 columns 1 and 2 in the respective workfiles. More importantly, identical values for best3_fr45 in the two workfiles. Trying out a couple of hypotheses right now.

[EViews Glenn]

I've been playing with this for a few days now and consistently getting the same results for BEST3_FR45 in the two programs.