pweights not allowed , please help me

[Khati Zolfaghari]

hello, I am using the xtscc and I want to estimate the ATE following the jorda`s paper. I have a panel data from 1996-2019 but when I estimate the model I get the error that "pweight is not allowed" , I try to find the logic behind this because when I do through the simple regression I can use the "pweight" but in the "xtscc" no.

gen a=forward_dumrecession

gen invwt=a/probit_hat + (1-a)/(1-probit_hat)

capture drop a invwt 
foreach v in  ltotalfertility          {
foreach vart in "0" "1" "2" "3" "4" "5"{
    * SAME OUTCOME REG IN BOTH T&C THIS TIME, REST ALL THE SAME (eq18)
   xi: xtscc d`vart'`v'  l(0/2)forward_dumrecession l(1/1)d0`v'  i.year [pweight=invwt],fe  
    *set sample
    gen samp=e(sample) 
    *actual
    predict mu0 if samp==1 & forward_dumrecession==0 
    *actual
    predict mu1 if samp==1 & forward_dumrecession==1 
    * ghost
    replace mu0 = mu1 - _b[forward_dumrecession] if samp==1 & forward_dumrecession==1 
    * ghost
    replace mu1 = mu0 + _b[forward_dumrecession] if samp==1 & forward_dumrecession==0 
    *from Lunt et al
    generate mdiff1=(-(a-probit_hat)*mu1/probit_hat)-((a-probit_hat)*mu0/(1-probit_hat))
    generate iptw=(2*a-1)*d`vart'`v'*invwt
    generate dr1=iptw+mdiff1
    *constant for convenience in next reg to get mean
    qui gen ATE_IPWRA=1 
    qui reg dr1 ATE_IPWRA , nocons 
    eststo DR1_d`vart'`v'
    sum dr1
    local dr1m = r(mean)
    gen Isq  = (dr1-`dr1m')^2
    sum Isq
    estadd scalar RobustSE = sqrt(r(mean)/r(N))
    estadd scalar pvalue = normal(`dr1m'/sqrt(r(mean)/r(N)))    
    capture drop iptw Isq mdiff1 dr1 mu1 mu0 samp ATE_IPWRA
    capture scalar drop dr1m
    }
    }

I hope I receive your assistance.

regards,

[Felix Bittmann]

I do not know the command xtscc but a quick workaround for similar problems is to use aweights instead of pweights. Test with caution.

[Khati Zolfaghari]

@Felix Bittmann thank you so much for your reply. xtscc is used for "Driscoll and Kraay (Rev. Ec. Stat. 1998) standard errors", and here i need the inverse probebility for this I used "pweight" , but when I recieved an error.
aweights which I think here is not for inverse probability weight as I see. it is inverse of variance if I say correct.

regards,