Application of environmentally extended input-output data to estimate greenhouse gas emissions attributable to packaged foods and beverages in Australia

Input-output tables for national economies are widely available and may provide an established mechanism for estimating environmental impacts of industry sectors. We used environmentally extended input-output (EEIO) analysis to estimate greenhouse gas emissions (GHGe) for Australian packaged foods and beverages and compared the findings to those derived from life cycle analysis (LCA). GHGe intensities in purchaser prices were used alongside corresponding median prices per kilogram from NielsenIQ Homescan, a nationally representative database with food prices. Applying GHGe estimates to a 2019 Australian packaged food database, we reported median and interquartile range (IQR) GHGe per kilogram, comparing them with LCA-derived estimates. EEIO-derived intensities were estimated for 23,550 products, and the median overall GHGe based on EEIO data was 6.87 kg CO2eq / kg (IQR 4.20 to 10.5). LCA-derived estimates were comparatively lower, showing a median overall GHGe of 2.42 kg CO2eq / kg (IQR 1.41 to 5.00) using Poore and Nemecek data and 2.35 kg CO2eq / kg (IQR 1.24 to 4.53) using FoodSwitch. Despite differences in magnitude, EEIO-derived data discriminated between the highest emitting categories in alignment with the LCA-based approaches. Future work is required to understand the discrepancies and make the EEIO method a compelling alternative to LCA-based approaches. • Input-output tables can be used to estimate environmental impacts of industry sectors. • Emissions from production of packaged foods calculated using input-output approach. • Emissions from packaged foods also estimated using prior life-cycle analysis data. • Input-output approach found higher emissions; applicability to packaged foods limited. • Hybrid approaches for environmental impact assessment may be ideal for packaged foods.