The objective of this paper is to consider methodology for modelling time series data of monetary aggregates such as monetary base and broad money. A brief review is made with regard to the likelihood‐based cointegration analysis of I(2) (integrated of order 2) data and I(2)‐to‐I(1) transformations. The paper then investigates procedures for econometric modelling of monetary aggregates, which are in general deemed to be I(2) variables analogous to price indices. It is shown that I(2)‐to‐I(1) transformations centering on a money multiplier play an important role in the modelling procedures. Finally, the study presents an empirical illustration of the proposed methodology using monetary aggregate data from Japan.