The aim of this article is to evaluate sensitivity of policy impact to the use of different methods of regionalization based on I-O approach. Policy impact is that coming from the application of the SAPARD pre-accession instrument for the period 2000-06 in three rural regions.