The Effect of Temporal Aggregation on the Test of Wagner's Law

This paper applies the Granger-Sims causality framework to historical data from the United States, West Germany, and Great Britain to develop estimates of the causal ordering between government expenditures and national income. Based on these estimates, it is clear that when one tests the Wagnerian and Keynesian predilections with annual and quarterly data, different results are obtained because of the use of values of variables calculated for different units of time. These results suggest that studies attempting to test Wagner's law and the direction of causality implied by Keynesian macroeconomic models should recognize the temporal aggregation problem.