Shouyong Shi(1998) presents a general equilibrium model which shows a persistent monetary propagation mechanism. There the high persistence is obtained by a combination of search frictions in the goods and labor markets and the presence of final goods inventories. The present paper addresses the question of robustness of these results, especially, how sensitive are Shi's results to parameter changes and to different model specifications. Calibration of the parameters to intervals is used to perform a global sensitivity analysis. The calibration exercise reveals that the model is quite robust to changes in parameters. Comparing different model versions - including a CIA model which appears as a special case when buyers and sellers match always - we can disentangle and quantify the contributions of the various frictions in accounting for the persistent propagation. Search-frictions in the goods market and inventory holdings are necessary for persistent propagation of monetary shocks. Labor market frictions are not crucial but prolong the output responses and reduce their magnitude.