Looking for Innovation Beyond the Patent System: Evidence from Research Disclosures

We study the content, novelty, and value of defensive publications relative to patents. We use a large language model (LLM) to apply the cooperative patent classification (CPC) system to a set of defensive publications (from 1962 to 2022) from the journal Research Disclosure, thus mapping such research disclosures and patents into a common space and allowing for a direct evaluation of textual similarities between these two types of R&D outputs. We find that while in some technologies, patents and research disclosures follow similar aggregate trends, some exhibit diverging developments over time. We also document shifts in the position of research disclosures in the patenting space that are indicative of changes in the technological landscape not captured in patents. We further show that substantial numbers of research disclosures are published before their closest patents are filed, and many contain terminology before it is first used in patents. Last, we find that in several technology areas, research disclosures have evolved from being an outlet for niche results to a vehicle to publicize technological developments of high practical relevance and value. Our results imply that when we draw conclusions about the nature of technological progress or the direction of innovation based solely on patent data, we obtain an incomplete picture.

Ganglmair, Bernhard and Alexander Kann (2025), Looking for Innovation Beyond the Patent System: Evidence from Research Disclosures, Discussion Paper Series – CRC TR 224, Bonn und Mannheim