A long-standing conjecture in the innovation literature is that demise of research labs has contributed to declining R&D productivity. We test that conjecture using data from industrial lab directories from 1921 to 1998. We find the number, size and prevalence and scale of labs have all increased rather than declined over the 20th century. Accordingly, lab demise cannot have contributed to declining R&D productivity. We therefore considered the opposite possibility: that the rise in labs contributed to the decline in R&D productivity. There we find that labs decrease firms’ R&D productivity. Firms with labs incur an estimated lab penalty of 4.7 to 11% lower productivity than firms without labs. Using both aggregate data and comparative case studies, we find this “lab penalty” seems to stem from labs’ parochialism.
