Regulatory Risk as a Financial Factor: An LLM-Derived Index of Cross-Border Data Restrictions

We introduce the Data Localization and Restriction Index (DLRI), a longitudinal measure of cross-border data restrictiveness constructed from legal texts via a large language model (LLM)-as-judge pipeline. From 218 legal documents across 417 country–year observations (2010--2023), we extract nine regulatory dimensions and aggregate them with principal component analysis into a normalized 0--1 index. Analysis of the DLRI–FDI relationship indicates heterogeneity across the levels of development: data regulation does not uniformly deter investment; instead, it anchors capital in developed countries while offering a null effect in developing economies. Beyond data policy, our approach demonstrates how generative AI can convert unstructured regulation into structured, market-relevant signals for political economy and financial risk modeling.