Dear All,
I am trying to estimate synthetic control difference-in-difference case studies. The unit of observation is state-quarter, with 34 states and 26 quarters. 15 of the states are treated, with varied treatment periods spread between the 5th-24rth quarters. The remaining 19 states belong to the donor group for creating the synthetic controls. So eventually, I will have 15 case studies, one for each state.
I have tried the following code which has given me missing results:
I am not really sure why synth is unable to estimate the RMSPE. The problem may simply be with my loops....
I will be very grateful for any help I can get.
Sincerely,
Sumedha.
I am trying to estimate synthetic control difference-in-difference case studies. The unit of observation is state-quarter, with 34 states and 26 quarters. 15 of the states are treated, with varied treatment periods spread between the 5th-24rth quarters. The remaining 19 states belong to the donor group for creating the synthetic controls. So eventually, I will have 15 case studies, one for each state.
I have tried the following code which has given me missing results:
Code:
local outcome narc /*otherreason misuse suicide reasonunknown nomineff modeff majeff de
> ath effmiss */
. local controlvars age015 age1624 age2534 age3544 age4554 age5564 age65plus agemiss male pop
.
. collapse (mean) `outcome' `controlvars' qtrtreat , by (trunit qtr)
.
. tsset trunit qtr
panel variable: trunit (strongly balanced)
time variable: qtr, 1 to 26
delta: 1 unit
.
. foreach y in `outcome' {
2. forvalues lname = 1/15 {
3. preserve
4. keep if (trunit==`lname'trunit>15)
5. egen maxp_`lname'=max(qtrtreat)
6. local p=maxp_`lname'
7. display `p'
8. local p1 = `p' - 1
9. display `p1'
10. /*qui*/ synth `y' `controlvars' `y'(1(1)`p1') , trunit(`lname') trperiod(`p') fig ke
> ep(synthetic_casestudy_`y'_`lname', replace)
11. restore
12. }
13. }
(364 observations deleted)
5
4
Synthetic Control Method for Comparative Case Studies
First Step: Data Setup
Data Setup successful
Treated Unit: 1
Control Units: 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
Dependent Variable: narc
MSPE minimized for periods: 1 2 3 4
Results obtained for periods: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
24 25 26
Predictors: age015 age1624 age2534 age3544 age4554 age5564 age65plus
agemiss male pop narc(1(1)4)
Unless period is specified
predictors are averaged over: 1 2 3 4
Second Step: Run Optimization
Optimization done
Third Step: Obtain Results
Loss: Root Mean Squared Prediction Error
RMSPE .
Unit Weights:
Co_No Unit_Weight
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Predictor Balance:
Treated Synthetic
age015 .1887048 .
age1624 .2452776 .
age2534 .183628 .
age3544 .1065218 .
age4554 .0979623 .
age5564 .0652333 .
age65plus .0674047 .
agemiss .0452675 .
male .4161599 .
pop 11544.82 .
narc(1(1)4) 1 .
(364 observations deleted)
I will be very grateful for any help I can get.
Sincerely,
Sumedha.
Comment