Page 2 of 4

Re: Favar QUESTION

Posted: Sun Apr 09, 2017 9:31 pm
by hamidlalkhezri
dakila wrote:I could not understand the questions. I never mentioned I(0) and I(1) factors. The add-in replicates Bernanke Boivin Eliasz (2005).

can You help me to find number of factors in Favar model?

Re: Favar QUESTION

Posted: Mon Apr 10, 2017 2:31 pm
by dakila
After the estimation, variables named _facrot1 _facrot2 ... will be created.

Re: Favar QUESTION

Posted: Mon Apr 10, 2017 11:41 pm
by hamidlalkhezri
dakila wrote:After the estimation, variables named _facrot1 _facrot2 ... will be created.

aftar the estimation variables are create. now how i estimation FAVAR model?How to obtain the number of lags?
thank u.

Re: Favar QUESTION

Posted: Tue Apr 25, 2017 12:04 pm
by hamidlalkhezri
dakila wrote:After the estimation, variables named _facrot1 _facrot2 ... will be created.

hi. can you help me to esimation favar model?
After estimating favar model this message appears:"Missing Value found in covariance/correlation matrix"

Re: Favar QUESTION

Posted: Tue Apr 25, 2017 3:02 pm
by dakila
Could you post the file?

Re: Favar QUESTION

Posted: Sat Jul 08, 2017 12:11 pm
by ali_Economist
Hi, I am trying to run the model, if i use the dialog box it gives "Syntax error" if use the commandline it says tcode not define, i tried best to give the tcode by following the program codes but it still keep on giving error message. if you please help. filed attached.

Re: Favar QUESTION

Posted: Thu Jul 13, 2017 1:34 am
by ali_Economist
Hi, i am trying to run the baysian favar, it is giving an error 500, if you please help.

Regards
Ali

Re: Favar QUESTION

Posted: Thu Jul 13, 2017 2:05 am
by ali_Economist
got it sorted, please ignore my both questions about FAVAR and BFAVAR. thanks.

Re: Favar QUESTION

Posted: Thu Jul 13, 2017 3:20 am
by ouerk
Hi
my "xdata" is a specific short term interest rate wich contain negatif values after 2010. this variable is stationnary after first difference transformation and i kwew that de tcode don't accept this transformation. I try to pass by Dlog but i can't (the value are negatif ,so i ca'nt transform them to log )

suggestions ??

thank you in advance !

Salima

Re: Favar QUESTION

Posted: Thu Jul 13, 2017 5:41 pm
by dakila
If I were you I will never transform interest rate. If you really want the first difference transformation, I will try to include it.

Re: Favar QUESTION

Posted: Fri Jul 14, 2017 4:42 am
by ouerk
Thank you for your answer !you are right, i will not transform the interest rate for the study !
One last question , i would like to produce 'conditional forecasts' . I knew how to do it with simple Var but not with Favar .

it is possible to produce conditional forecast with FAVAR ?

Thank you !!

SALIMA

Re: Favar QUESTION

Posted: Sat Jul 15, 2017 1:13 am
by dakila
Yes. it is possible to estimate conditional forecast with the FAVAR add-in. You will need to install the Confcast add-in.

after the estimation you should use the following command (for BBE example):
First you will need to estimate the factor loading equation for selected variable (for example series108):

Code: Select all

eq01.ls series108 _facrot1 _facrot2 _facrot3 ffr


then you should forecast the factors on conditional path of ffr :

Code: Select all

favar01.confcast "ffr ffr" "2001m12 2002m6" "0.25 0.50"


finally you should forecast the selected variable (series108):

Code: Select all

eq01.forecast


Keep in the mind all variables are standardized (zero mean and unit variance).
Therefore you should transform your variable back to the non-standardized one.

Re: Favar QUESTION

Posted: Sat Jul 15, 2017 11:53 am
by ouerk
Thank you for your precious help !

For the study, i must compare the impact of "expansionary monetary policy shock' to show if the unconventional monetary policy and conventional monetary policy have the same impact .

The unconventional monetary policy consist to 'decrease 'policy rate since the crisis = exclusively 'expansionary moneraty policy" so i can't comment the impulse reponses of the selected variables to 'recessionary monetary policy shock' (+0.25 bps) like Bernanke.

is it possible to obtain impulse reponses of selected variable to 'expansionary monetary shock" (-0.25 bps or - 1 bps ) ??

Thank you in advance for your help !!

Salima

Re: Favar QUESTION

Posted: Sun Jul 16, 2017 2:18 pm
by dakila
Yes, it is possible. First you need to save the IRF. For instance:

Code: Select all

favar(factor=3,horizon=48,rep=1000,ci=0.9,save=1) 13 xdata xslow xir tcode yx_name @ ffr


After the estimation you will need to calculate the unit IRF:

Code: Select all

matrix irf_scl=irfxmat201/irfxmat201(1,1)


then scale it by parameter you want for example -0.25

Code: Select all

irf_scl = irf_scl*(-0.25/ffr_std)
irf_scl.line(m)


where ffr_std is standard deviaton of FFR. Keep in the mind the all variables are standardized so you need to standardize the shock.

Re: Favar QUESTION

Posted: Wed Jul 19, 2017 4:04 am
by ouerk
thank you for your help, Dakila !!

i obtain impulse response matrix (for shock egal to -0,25) but i don't have confidence intervals , the axis 0 and the name of the selected variables (see workfile). what should i do to get these items (like figure 01 in your exple) ?

for the conditional forecast, i should transform for exple series2 to the non standardized on
-before i estimate eq01 ?
- before i forecast the factors ?
- before i forecast series2 ?

salima