(Mis)measuring Innovation Around the World

Using information from 30 patent offices around the world, we investigate the randomness of missing R&D (Research and Development) and non-patenting firms. In several countries, missing R&D firms routinely file for more patents than their positive R&D counterparts, while in other countries missing R&D firms exhibit only limited patent activity. Similarly, we find substantial cross-country variation in the proportion of positive R&D firms that do not file patent applications. To address these missing data problems, studies in innovation often truncate the data, strictly relying on firms with reported R&D and/or positive patent activity.

We show how excluding firms without observable innovation can induce a severe selection problem, providing the most biased estimates among the various remedies for missing innovation data. Empirical analysis demonstrates how excluding firms without R&D or patents leads to distorted inferences in tests on innovation capacity and corporate growth. Overall, our investigation indicates that corporate disclosures of innovation systematically differ around the world and that the exclusion of non-disclosing R&D and/or non-patenting firms creates biased results. 

The Brown Bag Seminars are sponsored by ERIM.
www.eur.nl/financegroup