Unobservable industry characteristics and the innovation-concentration-advertising-maze: Evidence from an econometric study using panel data for manufacturing industries in the FRG, 1979–1986

Abstract

In discussing the nexus between innovations and market structure it is often argued that industry characteristics (called ‘opportunities’) might play an important role as determinants of innovation, and that simultaneity rather than one-way causality prevails. We consider a three-equation model for innovation, advertising, and concentration. Based on pooled cross-section time-series data for 26 German manufacturing industries we estimate single equation models with and without fixed industry and/or time effects (to control for unobservable industry or time effects, respectively) and simultaneous equation systems including fixed effects, and controlling for extreme cases (‘outliers’) or not. Furthermore, we use two different measures for innovations, i.e., the percentage of shipments due to new products, and the percentage of firms which classified themselves as innovators. Our results can be summarized as follows: (1) The firm size has no significant effect on innovation. One can, therefore, not conclude from this data set that large firms are more innovative than small ones; (2) unobservable industry effects do matter; (3) the treatment of outliers does matter; (4) simultaneity does matter (5) the way innovations are measured does matter; (6) different stories could be told based on the results of the systems of interdependent equations estimated.