• We develop the FPDI to measure forest-related depth in 778 trade agreements across 213 countries (1947–2021). • Governance capacity is the most consistent correlate of FPDI across estimators, subperiods, and subsamples. • Tree-cover loss and agricultural expansion correlate with lower FPDI; forest rents and wood imports with higher scores. • Cross-country governance differences explain more FPDI variation than within-country changes over time. • Rankings are stable across 11 weighting schemes; FPDI captures treaty architecture beyond clause frequency. Trade agreements have become a vehicle for environmental commitments, but existing measures of environmental provisions in preferential trade agreements do not distinguish forest-specific content from general environmental architecture. This paper introduces the Forest Provisions Depth Index (FPDI), a composite indicator measuring the depth of forest-related commitments in preferential trade agreements across three dimensions: explicit forest content, implementation and compliance capacity, and normative anchoring. The index covers 778 agreements involving 213 countries from 1947 to 2021, uses geometric aggregation with a targeted discount when forest-specific clauses are absent, and is stock-based, accumulating provisions across a country’s in-force agreement portfolio. Country rankings are stable across 11 alternative weighting schemes, and the composite shows moderate correlation with a cumulative forest-clause count, indicating it captures treaty architecture beyond clause frequency. Panel estimates on 150 countries over 2002 to 2021 associate governance capacity, forest rents, and wood product imports with higher FPDI scores, while tree-cover loss and agricultural land expansion associate with lower scores. Persistent cross-country governance differences account for more FPDI variation than within-country changes, and agricultural expansion operates through the within-country margin. The FPDI provides a benchmark for cross-country comparison of treaty design and for research on trade-forest linkages.
Martins et al. (2026) studied this question.