BEKK model with VAR (1)

[upani]

Hi All,

I would like to estimate BEKK VAR (1) model and save the residuals. I just tried look in to the forum and found some post relevant in my context. I am using the following

Code: Select all

OPEN DATA "C:\Users\DELL\Documents\Estimation\gold.xlsx"
DATA(FORMAT=XLSX,ORG=COLUMNS,julian=Date) 1 3484 FP1 SP1
cal(julian=date)

* Create logs from Simple
*
set lfp   = log(FP1)
set lsp   = log(SP1)


set dlfp = lfp-lfp{1}
set dlsp = lsp-lsp{1}

set rfp = log(FP1/FP1{1})*100
set rsp = log(SP1/SP1{1})*100
***var-bekk-daily-total***
*****VAR1 lfp lsp*******
set lagrsp = rsp{1}
system(model=var1)
variables lfp lagrsp
lags 1
end(system)
***BEKK***
GARCH(P=1,Q=1, model= var1,mv=bekk, dist=t, robusterrors, pmethod=simplex, piters=20, $
method=bfgs, iters=500, rvectors=rd, hmatrices=hh,$
stdresids=rstd, factorby=eigen) /sacc lacc{1}
GARCH(P=1,Q=1, model= var1,mv=bekk, dist=t, ASYMMETRIC, robusterrors, pmethod=simplex, piters=20, $
method=bfgs, iters=500, rvectors=rd, hmatrices=hh,$
stdresids=rstd, factorby=eigen) /sacc lacc{1}

I would like to know whether the code is correct. I need your opinion in this regard.

With sincere regards,
Upananda Pani

[TomDoan]

You don't have a CONSTANT in your VAR, and you probably should have one.

I don't know what the /sacc lacc{1} are supposed to be at the end of your GARCH instructions, but they are
(a) wrong (those aren't in the data set, and you can't use lag notation for the dependent variable) and
(b) unnecessary anyway since the MODEL option is providing the information about the dependent variables.

That will give you a standardized residuals into RSTD, but note that you're doing those back to back so the first set (from the symmetrical GARCH) will get overwritten by the second GARCH before you do anything with them.

[upani]

Dear Sir,

Thanks for your reply. I have modified the variable name. The code follows.

Code: Select all

set lagrsp = rsp{1}
system(model=var1)
variables lfp lagrsp
lags 1
end(system)
***BEKK***
GARCH(P=1,Q=1, model= var1,mv=bekk, dist=t, robusterrors, pmethod=simplex, piters=20, $
method=bfgs, iters=500, rvectors=rd, hmatrices=hh,$
stdresids=rstd, factorby=eigen) /rfp lagrsp
GARCH(P=1,Q=1, model= var1,mv=bekk, dist=t, ASYMMETRIC, robusterrors, pmethod=simplex, piters=20, $
method=bfgs, iters=500, rvectors=rd, hmatrices=hh,$
stdresids=rstd, factorby=eigen) /rfp lagrsp

Hope i am correct this time.

With regards,
Upananda Pani

[TomDoan]

See point (b). You don't need to list the dependent variables on the GARCH instruction. The MODEL option takes care of that.