Synth Runner - How to generate one sided p-values for negative effects

[Iven Valpassos]

Hi everybody,
I´ve been working using the package synth_runner (https://github.com/bquistorff/synth_runner) in order to perform inference in synthetic control method analisys. After creating a synthetic control, the package calculates p-values measuring the number of control units that would have suffered effects higher than the effects in the treated unit, if they were the treated units (this is the placebo test) and dividing it by the total number of control units. By default, effects are considered in module, that is, positive and negative effects are considered in the statistic, no matter if the effect you see in your output variable is positive or negative.
The package has a built in solution that generates one-sided p-values when the effect is positive. However, in my case, the effect is negative (the actual trajectory of my output variable, per capita gdp, is under the trajectory of the synthetic inflation) and there is no built in solution to produce one sided p-values. Anyone knows how to generate the one sided p-values in this case? Is there an automated solution? Should I do it manually?

Thank you

[Phil Bromiley]

You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex.

On user-written packages, getting a result on this listserve depends a lot on whether any of the active folks use that package. I don't do synth and no one else chose to respond.

It is possible to open the ado file and modify it if you really want to. However, it is likely to be easier to use the output and calculate what you want using local macros.