This paper investigates empirically the usefulness of monetary aggregates as information variables in the conduct of monetary policy. For this purpose, some recent advances on the topic were used, which include the analysis of both real-time and revised final data, and the application of Bayesian model averaging to allow for model uncertainty regarding the lag length and number of cointegrating relationships. In this paper, money is considered as an information variable for Wt (e.g. output or prices) if the following two criteria are satisfied: (i) Mt is strongly exogenous, and (ii) Mt Granger-causes Wt. Strong exogeneity is relevant because it validates conditional forecasting of Wt using monetary aggregates as conditioning variables. The results show no strong evidence supporting the usefulness of monetary aggregates as information variables for prices or output. However, this does not preclude their potential use as information variables for other macroeconomic targets such as financial stability. It is worth mentioning that the results do not imply that monetary policy in Peru is not useful.