On innovative activity over the business cycle: a note

In a recent article, Geroski and Walters (Economic Journal, 105, 1995, 916-928) suggest that changes in demand Granger cause changes in innovative activity but not vice versa. This note shows that inappropriate modelling methods can lead to such conclusions which exclude the possibility of two-way causality between production and innovation. It suggests that applied economists must test data validity before any meaningful econometric analyses are conducted.