Analysing divisia aggregates for the euro area

In this paper, different Divisia monetary aggregates for the euro area are constructed over the period from 1980 to 2000. Theoretically, one main difference of these aggregates is their reaction to exchange-rate variations. Empirically, the aggregates are compared with respect to three issues. First, the demand for the Divisia aggregates is evaluated using the cointegrated VEC model and single-equation techniques, where stable demand functions are estimated. Second, the information content of these aggregates as regards future output is investigated. Evidence is presented that one of the Divisia aggregates has most information content from a forward-looking perspective. Third, using the P-star framework, the importance of money for future price movements is examined. Adapting an in-sample analysis, Divisia aggregates are important for HICP development and to some extent for GDP deflator movement. The out-ofsample forecasting exercise presents, on the one hand, evidence that simple-sum M3 includes more information for the HICP, whereas one of the Divisia aggregates helps to predict the future GDP deflator. On the other hand conspicuous control errors exist. In sum, the paper supports the view that money should have an important role in the conduct of monetary policy in the euro area.