Money's predictive role in output: evidence from recent data

We measure and assess the predictive performance of money in explaining output for the US, the UK, the Euro Area and Poland. Our approach, which uses more recent data and a larger set of areas, focuses on whether money 'Granger causes' output while taking into account the statistical properties of the data, which has implications for Granger causality tests. In addition to simple sum monetary aggregates, we use Divisia monetary aggregates, which are theoretically shown to be the actual measure of monetary services. Our results confirm that movements in money are still relevant in explaining movements in output, particularly when Divisia monetary aggregate is used. While the models using 'narrow' money, both simple sum and Divisia, provide mixed evidence of support, what appears to be robust across countries, is the predictive power of 'broad Divisia' in explaining output. Furthermore, our research finds no evidence that the statistical significance of money decreases as more variables are added to the model, or models with level variables produce more significant results than models with growth variables. In fact, rather than the level of money, it is the growth rate of money and the deviation of the growth rate of money from its time trend, which 'Granger causes' output.