In this study, we determine the reliability and exogeneity of four popular monetary policy shock measures, namely the narrative series of Romer and Romer (2004), the high-frequency series of Barakchian and Crowe (2013), the high-frequency series of Gertler and Karadi (2015), and the hybrid series of Miranda-Agrippino and Ricco (2018b). To this end, we employ the Proxy-SVAR model and different empirical diagnostic tools to determine the shock measures’ information content. We find that the measure of Miranda-Agrippino and Ricco (2018b), combining the insights from the narrative approach and high-frequency identification, outperforms the other three series.