The Causality between R&D and Productivity in Manufacturing: An international disaggregate panel data study

This study analyses the causality between productivity and domestic and foreign R&D on the basis of panel data with respect to 22 manufacturing sectors in 14 OECD countries during the period 1972-94. A unit root analysis shows that the relation between the log of total factor productivity (TFP) and the logs of domestic and foreign R&D is cointegrated. Causality tests are performed on corresponding dynamic VAR and error correction augmented VAR models, estimated both on the total panel and on 22 sub-panels, sector by sector. Their results show that, although there are feedbacks, both on average and in a clear majority of sectors the causation runs mainly from the R&D variables to TFP rather than the other way round. This causation is, moreover, shown to be, in the first place, long-run in nature. It is concluded that, when considered in conjunction, these results suggest an essentially supply-type rather than demand-pull kind of interpretation of the cointegrating relation between TFP and R&D.