synth conformability error

[flavia]

Hello,
I'm trying to run a synthetic control method (Abadie, Diamond, Hainmueller, 2009) for a balanced panel of 21 countries. I'm trying to figure out whether a switch from a free floating exchange rate regime to a hard peg (i.e. Finland's accession to the EMU in 1997) has had any impact on the gini coefficient. Although I have no missing values for the Independent or dependent variable, and neither for my values of predictor, Stata always spits out "conformability error". Has anybody encountered this problem before? If yes, how have you dealt with it?
Thank you in advance.
Flavia

[Stephen Jenkins]

Flavia: to boost your chances of getting a response or, indeed, a helpful response, please respect Forum etiquette, as set out in the FAQ. For example, provide full bibliographic citations perhaps also including URL and/or DOI (you didn't). You should also show us exactly what you typed and exactly what Stata returned. Post all this within CODE delimiters. They're accessible via the advanced editor (A button above where you compose your message, and then # button).
Also, please use your real name (firstname lastname) in the Forum. It's simple to re-register: hit the Contact Us button on bottom right of screen, and make your request. Thank you.

[Kilian Heilmann]

Dear Flavia,

I ran into the same error and fixed it by making some minor changes in the synth.ado. The conformability error seems to be caused by a non-critical command that can be captured. Here is what I did to make it work:

1) Go into the ado folder where Stata saves the packages and find the synth.ado file (should be in the folder ado/plus/s)
2) Open the synth.ado and replace line 3 "program synth , eclass" with "program synth2 , eclass"
3) go to line 672 and replace "mat rowname `wsolout' = `colabels'" with "capture mat rowname `wsolout' = `colabels'"
4) save the file as synth2.ado

Now you can run the synthetic control group with the same options using the command "synth2" instead of the original "synth".

Best,
Kilian

[Jason Ward]

Hi all,

I'm running into this same problem. Using the synth package with Stata 14 IC on a Mac (OS 10.10.5). I followed the synth faq and made sure to update all and get the latest synth package from the authors' site. Trying to use a line of code like this:

(going to skip the code delimiters since this code isn't using sysuse data or anything and I don't think the code is my problem, very simple versions of it are giving me trouble. I think it is related to the issue described by Killian above).

synth `x' total_undergraduates `x'(2003) `x'(2005) `x'(2007) `x'(2009), trunit(221759) trperiod(2011) nested fig keep(resout_univname_`x')

Getting "conformability error r(503)". So I am trying to go find and edit /save the synth.ado as described above but I cannot find it. The ado folder in my Stata folder only has a "base" folder in it and when I search for "synth.ado" I come up with nothing. I have hidden folders shown so I don't believe it is an issue where I just don't see a "plus" or "updates" folder.

This is not just a synth problem. I can't find any of my installed .ado files. A search turns up all the basic Stata .ado files but nothing I've ssc installed.

Any help would be greatly appreciated. Thanks much!
Jason Ward

[Jason Ward]

I found the .ados and tried the hack described by Killian above, but it does not help with this problem. The search for answers continues.

[Joel Levin]

Hi all,

I'm running into this same problem. Using the synth package with Stata 14 IC on a Mac (OS 10.10.5). I followed the synth faq and made sure to update all and get the latest synth package from the authors' site. Trying to use a line of code like this:

(going to skip the code delimiters since this code isn't using sysuse data or anything and I don't think the code is my problem, very simple versions of it are giving me trouble. I think it is related to the issue described by Killian above).

synth `x' total_undergraduates `x'(2003) `x'(2005) `x'(2007) `x'(2009), trunit(221759) trperiod(2011) nested fig keep(resout_univname_`x')

Getting "conformability error r(503)". So I am trying to go find and edit /save the synth.ado as described above but I cannot find it. The ado folder in my Stata folder only has a "base" folder in it and when I search for "synth.ado" I come up with nothing. I have hidden folders shown so I don't believe it is an issue where I just don't see a "plus" or "updates" folder.

This is not just a synth problem. I can't find any of my installed .ado files. A search turns up all the basic Stata .ado files but nothing I've ssc installed.

Any help would be greatly appreciated. Thanks much!
Jason Ward

You can find the default path that your SSC-installed .ado files are in by typing 'sysdir' in the command prompt. The default folder for SSC files is the 'plus' folder.

[Liu Qiang]

Dear Flavia,

I ran into the same error and fixed it by making some minor changes in the synth.ado. The conformability error seems to be caused by a non-critical command that can be captured. Here is what I did to make it work:

1) Go into the ado folder where Stata saves the packages and find the synth.ado file (should be in the folder ado/plus/s)
2) Open the synth.ado and replace line 3 "program synth , eclass" with "program synth2 , eclass"
3) go to line 672 and replace "mat rowname `wsolout' = `colabels'" with "capture mat rowname `wsolout' = `colabels'"
4) save the file as synth2.ado

Now you can run the synthetic control group with the same options using the command "synth2" instead of the original "synth".

Best,
Kilian

you are correct. Thank you for your reminders.