In and out of Pasteur’s quadrant: revisiting trends in corporate science

Based on a very large sample of publications and patents, we examine trends in US corporate science from 1980 to 2014 and revisit the hypothesis of its decline, as formulated in the recent literature. Drawing from Stokes’ (1997) ‘quadrants’ taxonomy, we produce two distinct indicators for corporate publications’ appliedness (consideration of use) and basicness (quest for fundamental understanding), under no assumption of a trade-off between the two. We find that corporate publications have progressively become more applied and less basic than academic ones, after controlling for fields and journals and assuming any measurement error to be time invariant. These results are mostly driven by established large firms in specific sectors, which appear to be moving away from Pasteur’s quadrant in the direction of Edison’s quadrant. Publications resulting from university-industry collaborations follow a similar trend, but in a much less pronounced way.