We analyze the relationship between website visits, magazine demand and the demand for advertising pages using Granger non-causality tests on the ba- sis of an extensive and externally audited quarterly data set for the German magazine market spanning the period I/1998 to II/2004. We use traditional panel data estimators and an estimator suitable for heterogeneity across mag- azines. We find very robust evidence for positive effects from website visits to circulation. There is no evidence of causality running in the opposite direction. Our findings are contrary to the widespread belief that the Internet will cannibalize print media markets.
Granger causality heterogeneous panel data models Mean Group Estimation website visits magazine circulation