Hello Statalist,

My name is Bam. I'm a master's degree student and a new user of Statalist.

I find some critical problems for using 'Synth' command in Stata.

The original research used the Synthetic Method to find the effects of using tobacco policy in California in 1989. They used "Synth" to compare California to other states in the US (See Abadie, Diamond, and Hainmueller (2010) or 'help synth' in stata for the original case). They used California to be the treated unit (receive policy) and other states of being a synthetic unit (do not accept the policy). The data from 1970 to 2000.

The form of data is state | year | tobacco consumption | xx | ...
California 1970 xx x
. . . .
California 2000 xx x
Ohio 1970 xx x
. . . .
Ohio 2000 xx x
. . . .
. . . .
and use code:

. tsset state year
. synth cigsale beer(1984(1)1988) lnincome retprice age15to24 cigsale(1988) cigsale(1980) cigsale(1975), trunit(3) trperiod(1989)

Then, they got the difference between the tobacco consumption of California and other states.

In my case, I adjust the concept to my research to use Synth to find the difference of worker's wage after using minimum wage policy (policy was used from 2013 to 2016) in only 1 country (policy was used in hold country) by comparing the year before policy (2007-2012) to the year after policy (2013-2016).

I want to set the year before using policy to be the synthetic unit and set the year after using policy to be the treated unit.

my data is from 2007 to 2016 for example,

year | wage | xxx | xxxx
2007 5000 xxx xxx
2007 5100 xxx xxx
2007 3000 xxx xxx
. . . .
. . . .
2008 6000 xxx xxx
2008 6000 xxx xxx
. . . .
. . . .
2016 8000 xxx xxx
2016 7000 xxx xxx

(The sample in my data was a random and different amount in each year)

The problem is I can't use the Synth command because I can't set the time variable (tsset).
It shows code:

. repeated time values in sample r(451);

Also, if I set the unit like unit | year | wage | xxx | xxxx
1 2007 5000 xxx xxx
1 2008 5800 xxx xxx
. . . .
1 2012 4800 xxx xxx
2 2013 9000 xxx xxx
2 2014 5800 xxx xxx
. . . .
2 2016 3200 xxx xxx
by unit=1 is the synthetic control unit and unit=2 is the treated unit.

use code:

. tsset unit year

I still can't use the synth command because the data is too small. The Stata code:

. synth wage index wage(2012) wage(2011) wage(2010) wage(2009) wage(2008) wage(2007), trunit(2) trperiod(2013)
. no observations r(2000);

I don't know how to adjust my data to fit with the synth condition.

I hope any Statalist know and help me solve this problem or tell me the mistake.


Thank you very much for your help.