Sound Multisector Dynamic modeling hinges significantly upon the quality of data. The Food Balance Sheet (FBS) constructed by the Food and Agriculture Organization (FAO) has been widely used as a primary source of data in research related to global agriculture development, food security, nutrient, and dietary health. It provides a comprehensive overview of a country's supply (e.g., production, import, and stock variation) and utilization (e.g., food, feed, export, processing, seed, loss, etc.) patterns and macronutrient (fat, protein, and carbohydrate) for over 100 aggregated food commodities on a primary commodity equivalent (PCE) basis. However, the practical implementation of the PCE approach is considerably challenging, as it requires maintaining data balance simultaneously across different dimensions, including (1) supply and demand, (2) space (trade balance), (3) time (storage carryover), (4) along the processing chain (reasonable extraction rate). 

Here, we are motivated to improve the PCE method to provide a transparent, reproducible, consistent, and more flexible approach to aggregating agricultural commodities across different dimensions. In particular, we constructed a nested commodity mapping to understand the relationship across all 500+ commodities (with detailed supply-utilization accounts) and employed a data-based approach to calculate the extraction rate (a ratio between the production of secondary products and the processed use of the primary product). This approach maintains data balance across different dimensions, such as supply and demand, space, time, and along the processing chain. This more detailed commodity information also improves the portrayal of the food system. For example, the data we compiled provide additional information such as (1) macronutrient and dietary energy supply at the detailed commodity level, (2) a broader commodity coverage, including the balance of non-food commodities (e.g., oilseed cake, fiber crops, rubber, etc.), and (3) additional category detail, such as opening and closing stock data.

One important application of the PCA approach is to provide more flexible and consistent sector aggregation in global economic modeling. The new data has been incorporated into the Global Change Analysis Model (GCAM). This data development is an ongoing process that would benefit from the continued efforts of a broader community with regard to updating, maintaining and improving the processing. The data processing is made in an extension of an open-source R package (gcamfaostat), to ensure traceability, transparency, and reproducibility and to document the detailed mapping, parameter, and processing assumptions. The updated package further bridges the gap between FAOSTAT data and global economic modeling. Our initiative also seeks to enhance the quality and accessibility of data for the global agroeconomic modeling community, with the aim of fostering more robust and harmonized outcomes in a collaborative, efficient, and open-source framework. Our study provides important insights into better understanding the food supply chains, developing a new and more rigorous methodological basis for future modeling and analytical applications.