Reconciling forest product statistics for enhanced sector modeling: A recursive dynamic goal programming approach

International forest product statistics contain measurement errors, missing values, and trade flow discrepancies. Our study presents a recursive dynamic goal programming approach to identify and reconcile these inconsistencies over the period from 2000 to 2022. Unlike traditional static reconciliation methods, this framework includes year-over-year consistency, optimizing deviations between reported and estimated values while adhering to technological constraints and trade balances. Results indicate substantial regional heterogeneity: Europe and North America exhibit high reported–reconciled agreement, while Oceania and Africa require larger adjustments. Export statistics consistently require the greatest corrections, with import–export reconciliation gaps often reaching 20–30 percentage points. Agreement rates vary widely across products, with residues (waste paper, chips and particles) requiring substantially more adjustment than other categories. The reconciled dataset reveals diverging trends in wood processing efficiency, characterized by gains in Europe and Asia contrasted against declining efficiency in South American sawnwood production. The reconciled price series captures major market dynamics, including the 2021–2022 pandemic-induced surge. Validation confirms elimination of all material balance violations present in raw data. This methodology improves resource potential evaluation and provides a robust, internally consistent foundation for future forest sector analysis.