Unveiling the long history of bilateral food trade

The globalization process has profoundly transformed the food system over time, leading to a significant increase in food and agricultural trade. The multidimensional impact of this shift is substantial, underscoring the need for a deeper understanding of the evolution of the International Food Trade Network (IFTN). This study addresses a gap in data availability by reconstructing missing bilateral trade matrices from 1961 to 1985 for five major crops (wheat, maize, soybeans, rice and barley). The Backward Stochastic Reconstruction – a new procedure based on the RAS algorithm – is introduced to recover missing data, focusing on five essential food items. To validate our approach, we apply the procedure to the period 1986-2016, where data is available, treating the reconstructed values as though they were actual data. Key topological indicators calculated from both this Forward Stochastic Reconstruction and actual data are compared to assess the accuracy of the method. Our findings show strong consistency between the estimated and actual data, demonstrating the robustness of the proposed procedure. Moreover, the analysis of the historical evolution of the trade highlights significant transformations in the network structure. This study provides a generalizable methodology for reconstructing historical trade flows with a minimal amount of information and generates an extended IFTN database for the five items considered, potentially applicable to other foods to study food system dynamics.