'All the money in europe?' An investigation of the economic properties of EC-wide extended monetary aggregates

This paper explores the economic properties of several EC-wide monetary aggregates which are extended to include different measures of cross-border holdings (CBHs). The EC-wide aggregate obtained by simply adding national definitions of broad money is used as a benchmark, since econometric analysis confirms previous findings that it can play a useful role in the coordination of monetary policy at the European level. Cointegration analysis and error-correction modelling in fact show that the demand for it is stable and predictable, with a statistical performance which compares most favourably with national equations. And Granger-causality tests, conducted in both bivariate and multivariate settings for a wide range of specifications, indicate that it has predictive value for both nominal and real Community income. Aggregates including CBHs perform differently according to the definition. "Very broad" measures (which include EC-residents' CBHs kept outside the EC or non-ECresidents' CBHs kept within the EC but denominated in non-EC currencies) are poorly linked with EC-wide income, since they do not pass cointegration tests and do not Granger-cause Community income. In contrast, extended definitions "focused on the EC" (which hinge on the inclusion of CBHs denominated in EC-currencies andlor kept within the EC) are shown to possess the economic properties necessary for money measures to be useful for monetary policy and economic analysis. Although such extended aggregates do not yet outperform the traditional measure of EC-wide broad money, the results suggest that they may shortly become an increasingly important tool for monetary analysis at the EC level.