Sensitivity of CGE simulation results to competing SAM updates

Recent research shows a renewed interest in the properties of non-survey updates of input-output tables and social accounting matrices (SAM). Along with the scaling RAS method, new procedures related to entropy minimization and other metrics have been examined. Whether they will replace or complement the RAS method is an open question. The performance of the updating procedures has been tested with proximity measures to a reference input-output table or SAM, but we look at the performance of three updating mechanisms by comparing the simulation results from adopting alternative databases for the calibration of a reference CGE model.