Treaty Design and the Globalization of Innovation: IP Protection, Regulatory Space, and Technology-Specific BIT Effects

• In Event: AI Governance, Regulation, and Public Institutions

Abstract

The globalization of innovation in the form of the cross-border diffusion of technologies through adoption, sourcing, and collaborative invention has grown over recent decades yet remains highly sensitive to institutional frictions. Bilateral investment treaties (BITs) have been shown to materially increase cross-border innovation diffusion and innovation-related foreign investment. It does so by strengthening cross-border property rights and reducing contracting risk for R&D-related activities including technology-focused joint ventures, mergers and acquisitions, venture capital, and licensing. However, existing empirical work treats BITs as a homogeneous binary shock, leaving room for leveraging substantial and well-documented variation in clause-level treaty design. In particular, no study has examined how specific textual features like the strength of intellectual property (IP) coverage and the scope of right-to-regulate provisions  shape the magnitude and composition of BIT-induced innovation diffusion. This paper addresses this gap directly.Using a country-pair–technology-class–year panel of patent data from PATSTAT Global covering over 100 patent offices from 1991 to 2023, I measure innovation diffusion at five stages following Bian, Meier, and Xu (2025): adoption, citation-based sourcing, transfer-based sourcing, co-invention, and co-application. BIT timing data are drawn from the UNCTAD, and treaty design is coded from treaty texts along two dimensions: IP coverage strength (explicit IP enumeration, TRIPS-plus standards, breadth of asset definitions) and right-to-regulate scope (operative right-to-regulate clauses and public-interest carve-outs). Technology-class imitation risk is measured using the obsolescence index of Ma (2023), which captures citation-decay speed across 3-digit IPC technology classes.The empirical strategy exploits staggered BIT signings in a difference-in-differences framework with country-pair–technology-class, source-country–year, host-country–year, and technology-class–year fixed effects, isolating within-pair, within-technology variation in innovation diffusion around BIT events. Treaty design enters as an interaction with BIT status, and imitation risk as a triple interaction, testing whether IP-strong BITs produce differentially larger effects in slow-obsolescence, high-imitation-risk technology classes. Event-study specifications using the Callaway and Sant'Anna (2021) estimator verify parallel pre-trends and document dynamic responses across treaty design categories. I expect that BIT effects on cross-border innovation diffusion will be significantly amplified in treaties with strong IP provisions, while right-to-regulate clauses will dampen this effect. I also expect design-driven differences to be concentrated in high-imitation-risk, slow-obsolescence technologies such as pharmaceuticals and specialty chemicals, where the marginal value of robust IP protection is greatest.This paper makes three contributions. First, it provides the first treaty-design-sensitive causal estimates of BIT effects on innovation diffusion, moving beyond the BIT dummy to exploit clause-level textual variation. Second, it demonstrates that the institutional environment for cross-border IP shapes the geography and depth of global knowledge flows. Third, it informs ongoing policy debates around investment treaty reform and the innovation consequences of right-to-regulate language, suggesting that treaty design choices carry significant and previously unmeasured consequences for global innovation diffusion.

Author

• Sadman Shafiq, University of Texas, Dallas
  Presenting Author