Text-as-Data Analysis of Preferential Trade Agreements: Mapping the PTA Landscape

Abstract

Preferential trade agreements (PTAs) form an intricate web that connects countries across the globe. In this article, we introduce a PTA text corpus and research tools for its fine-grained, automated analysis. Recent computational advances allow for efficient and effective content analysis by treating text as data. We digitize PTA texts and use textual similarity tools to assess PTA design patterns on the global, national, and chapter level. Our descriptive analysis reveals, inter alia, that PTAs are more heterogeneous as a group than, for instance, bilateral investment agreements, but that they converge in regional or inter-regional clusters of similarly worded agreements. Following our descriptive account, we provide three concrete, interdisciplinary examples of how text-as-data analysis can advance the study of trade economics, politics, and law. In trade economics, similarity measures can provide more detailed representations of PTA design differences. These allow researchers to capture more meaningful variation when studying the economic impact of PTAs. In trade politics, scholars can use treaty similarity to trace design diffusion more accurately and test competing explanations for treaty design choices. Finally, in trade law, similarity measures offer new insights into the processes of normative convergence between legal regimes such as trade and investment law.

Keywords: preferential trade agreements, free trade agreements, text-as-data, similarity, computational analysis of law, trade law, WTO