Drawing the IRF for 25 combinations in a loop form

[Khati Zolfaghari]

Hello, I try to estimate and draw a chart of IRF for 25 combinations based on this model :

yi,t+k=𝛼ik+𝛾tk+𝛽k∆Ri,t+∑𝜗jk∆Ri,t−j+∑𝜃jkyi,t−j+𝜀tk

that

∑𝜗jk∆Ri,t−j

is from 1 to 5

and also

∑𝜃jkyi,t−j

is from 1 to 5

∆Ri,t

is the shock

I do not know how to solve it and draw the chart for 25 combinations in a loop form ( because it could be lag one versus lag one, lag one versus two ......, lag tow versus lag one, lag two versus lag two ....and so on)

I hope I receive your guidance.


Many thanks in advance.

regards,

[Khati Zolfaghari]

please help me.
regards,

[Justin Niakamal]

If I understand your question, here's a starting point. You can try to rework this to suit your needs.

clear all

tempfile temp
save `temp', emptyok

forvalues x = 1/5 {
    webuse lutkepohl2, clear
    tsset
    
    var dln_inv dln_inc dln_consump, lags(1/`x')
    irf create order`x', step(10) set(myirf`x', replace)
    use myirf`x'.irf, clear
    
    append using `temp'
    save `temp', replace
    erase myirf`x'.irf
}

use `temp', clear

* as an example
keep if response == "dln_inc" & impulse == "dln_consump"
encode irfname, gen(lag_order)

xtset lag_order step
xtline oirf, overlay

[Khati Zolfaghari]

@Justin Niakamal thank you so much for your reply.
my question is that I am dealing with the panel data from 1996-2019.

yi,t+k = 𝛼i+𝛾t+ 𝛽∆Ri,t +∑𝜗j ∆Ri,t−j + ∑𝜃j yi,t−j + 𝜀t

𝛽∆Ri,t is the shock
∑𝜗j ∆Ri,t−jis from 1 to 5

∑𝜃j yi,t−jis from 1 to 5

I want to draw the chart for 25 combinations in a loop form ( because it could be lag one versus lag one, lag one versus two ......, lag tow versus lag one, lag two versus lag two ....and so on).

your example really helps me, but the problem is that here "lags(1/`x')" are for all independent variables. I think for example of mine and for the "IRF" I should do another? I could not get different charts.

I did in this way , but it is not correct,

clear
foreach v in "ltotalfertility"  {
foreach i in "d_dumrecession"    {
forvalues x = 1/5               {    
foreach vart in "0" "1" "2" "3" "4" "5" {
cd "C:\Users\KHATEREH\Desktop\MyProject/`v'"
use "C:\Users\KHATEREH\Desktop\fertility_xtscc.dta"   
xi: statsby _b _se e(r2_a), clear:  xtscc f(`vart').`v' d_dumrecession l(1/`x')`i' l(1/`x')`v' i.year, fe
irf create order`x'.irf , clear
foreach var in d_dumrecession {
    gen t_`var' = _b_`var'/_se_`var'
    gen time = "obs`vart'"
save `v'`vart'.dta, replace
    }
    } 
    }
    }    
    }

I do not know how to solve it and draw it.


many thanks,
regards ,

[Khati Zolfaghari]

I hope I receive your assistance soon . I need your help.

Regards,

[Justin Niakamal]

Since you have panel data (I overlooked that in your notation) you should look into pvar (from the Stata Journal). I've not used xtscc (SSC), but it looks like it doesn't support IRFs as indicated by the post estimation commands. So one option is to use pvar -- throw your dummy variables exog(varlist) and save the data from pvarirf

[Khati Zolfaghari]

@Justin Niakamal Thank you so much for your reply. I am sharing my codes and, finally, it gives me at the end 25 charts of combination.

clear
foreach   v in "totalfertility"         {
forvalues x = 1/5                       {
forvalues y = 1/5                       {                
foreach vart in "0" "1" "2" "3" "4" "5" { 
cd "C:\Users\KHATEREH\Desktop\MyProject/`v'"
use "C:\Users\KHATEREH\Desktop\charts A\xtscc-correct-dataset.dta"   
xi: statsby _b _se e(r2_a), clear: xtscc d`vart'`v' l(0/`y')dum_recession l(1/`x')d0`v' i.year,fe
foreach var in dum_recession {
    gen t_`var' = _b_`var'/_se_`var'
    gen time = "obs`vart'" 
save `y'`x'`v'`vart'.dta, replace
    }
    }
    }    
    }
    }

and then :

clear
foreach v in  "totalfertility"  {
forvalues x = 1/5               {
forvalues y = 1/5               {        
use "C:\Users\KHATEREH\Desktop\MyProject/`v'/`y'`x'`v'0.dta"
foreach var in  "1" "2" "3" "4" "5" {
append using  "C:\Users\KHATEREH\Desktop\MyProject/`v'/`y'`x'`v'`var'.dta"
*drop _merge
}
gen date_string = substr(time, 4, 2)
destring date_string, gen(year)

keep  year  _b_dum_recession _se_dum_recession t_dum_recession 
order  year  _b_dum_recession _se_dum_recession t_dum_recession 
gen label="all"
bysort label (year) : gen first = _n == _N-5
**expand only the first obs
expand 2 if first==1
**** generate the new time variable
bysort label (year) :gen time2 = _n-2
***recode the variables
replace _b_dum_recession=0 if time2==-1
replace _se_dum_recession=0 if time2==-1  
replace t_dum_recession=0 if time2==-1  
bysort label (year) : gen low = _b_dum_recession-(_se_dum_recession*1.645)
bysort label (year) : gen high = _b_dum_recession+(_se_dum_recession*1.645)
gen a=0 
*a=0 is for charts  
save "C:\Users\KHATEREH\Desktop\MyProject/`y'`x'`v'all.dta", replace
}
}
}

I am sharing this, maybe of intrest.

Thanks a lot.