Do Patents Enable Disclosure? Evidence from the Invention Secrecy Act

Do patents facilitate knowledge disclosure as indicated by the relationship between invention secrecy and follow-on inventions?

Understanding the relationship between patents and knowledge disclosure is crucial for patent policy and innovation strategy. Past evidence suggested unclear impacts of patents on follow-on inventions, highlighting the need to assess how secrecy might hinder knowledge diffusion and subsequent innovations. The findings of this research could inform policymakers on the potential social costs of enforcing invention secrecy.

We analyzed 2,121 patents subject to secrecy orders from 1982 to 2000, employing a two-level matching approach to estimate regression models comparing treated patents with matched control patents. Our methodology focused on citation patterns and text-based measures of invention similarity to quantify the impact of secrecy orders on follow-on inventions.

We find that a secrecy order reduced the number of follow-on inventions by approximately 40% during its enforcement period, with a further 45% decline observed even five years after the order was lifted. Our analysis indicates that patents cited by secret patents received an average of 1.35 fewer citations than controls, suggesting significant long-term effects on innovation. Additionally, using text similarity measures, we confirm the negative relationship between secrecy and subsequent inventions.

• 2,121 patents that were subjected to secrecy orders from the USPTO Patent Examination Research Dataset (PatEx)
• Control group of 1,244 patents matched on filing year and technology class

• The matched control patents are representative of the treatment patents in terms of innovation potential.
• The secrecy orders enforced lead to a direct reduction in follow-on inventions as measured by patent citations and similarity scores.