Note: An earlier version of GTAPHSc.zip was replaced with a corrected update on 7/13/2021.



GTAPHSc.zip archive includes policy experiments for the GTAP-HS model and database (based on the GTAP 10p2 Data Base). Four policy experiments are included:

1. gtapv7     numeraire shock

2. vfretal    implementation of the imp retaliatory tariffs on U.S. vegs, fruits and nuts exports by all U.S. trading partners

3. uschtars   implementation of ongoing trade frictions between U.S. and China (up to the 2nd round of U.S.-China trade frictions)

4. alltars    implementation of all ongoing trade frictions between U.S. and its trading partners (up to the 2nd round of U.S.-China trade frictions)



"default_cepii_v2.prm" parameter file includes HS6/disaggregated commodity trade elasticities sourced from Fontagne et al (2019).

"default_gtap.prm" parameter file includes uniform trade elasticities for the disaggregated commodities (based on the standard GTAP Data Base elasticity values).



Solution time per one policy experiment is around 20-30 minutes depending on computing power and policy experiment type.

In the manuscript, for the policy experiments with GTAP-HS model with CEPII elasticities, mean results of the SSA runs are reported. These mean estimates might slightly differ from the one time solution run.



TWARGTAP.zip archive includes four policy experiments (as listed above) for the standard GTAPv7 model. 



Both experiment files can be run in a RunGTAP software (https://www.gtap.agecon.purdue.edu/products/rungtap/default.asp).

Open RunGTAP, select File/Version Archive/Load ZIP and choose corresponding archive. All parameters, closures and shocks are saved with experiments.



To replicate the Systematic Sensitivity Analysis (SSA) results, if desired, users should first run an experiment from the "GTAPHSc.zip" archive. 

Then go to the "ssa_shocks" folder of the "GTAPHSc" archive and copy elasticity variations from the "SSA_variation.txt" file (all content of the file).

In the RunGTAP go to Tools->Sensitivity->w.r.t. Parameters. Copied content of the "SSA_variation.txt" file should be paste to the "Set up Parameter sensitivity calculation" window.

Replication of the SSA runs for each policy scenario could take around 20-40 hours depending on computing power, policy experiment type and specification of the multiprocessing feature for SSA runs.

To derive mean and error bars reported in the paper, procedures described in the manuscript should be followed.