Modelling the demand for M3 in the Euro area

Modelling monetary transmission is central to understanding the role of monetary policy in the Euro area, and money demand is commonly seen as a link in that transmission mechanism. Since the beginning of the 1990s, many studies have suggested that the demand for Euro area broad money is stable over the long run because the estimation of an area-wide demand for money function provides an appropriate solution to a number of potential causes of misspecification of the single-country relations (such as spillover effects and currency substitution), and enjoys the positive consequences of a statistical averaging effect. On the other side, it must be stressed that previous benefits can be achieved at the risk of introducing parameter heterogeneity into the area-wide relationship. In order to shed some light on the issue, this study is first devoted to an analysis of the main econometric features of the money M3 demand at Euro area and single country levels, then it compares the two sets of results in a common framework that, differently from all previous studies, explicitly takes account of the potential nonstationarity of the variables of interest in both estimation and testing phases. The comparison shows that the area-wide money demand is more smooth and less subject to shocks than the single-country ones. Finally, a number of poolability tests run over subgroups highlight that low precision associated with the estimates of the parameters of the national models makes it impossible to exclude that their long-run specifications do in fact coincide.