Assessing food consumption discrepancy between food balance sheet and survey dietary data in Europe

National-scale estimates of food consumption are essential for assessing a country’s agricultural situation and informing policy on food security, nutrition, trade and land-use planning. However, different data sources often produce divergent estimates. This study quantifies the discrepancy between two widely used sources of national food consumption data in Europe: one based on individual dietary surveys (EFSA database) and the other derived from FAOSTAT's Food Balance Sheets (FBS). We compare per-adult consumption across four staple food products: wheat, rice, maize, and potatoes. We express discrepancies as the ratio between EFSA and FBS values. Results show systematic differences: wheat displays the smallest discrepancy (mean=0.77; median 0.73), with rice showing similar values (mean=0.73; median=0.67), while maize (mean=0.48; median=0.36) and potatoes (mean=0.40; median=0.38) show more pronounced differences. Overall, FBS-based consumption tends to exceed survey-based estimates. Country-level means vary markedly: Sweden showed the lowest discrepancy (mean=0.89; median=0.83) and Romania the highest (mean=0.35; median=0.32). Variability of the results at the national level was significant, suggesting that methodological, cultural and supply-chain factors affect the magnitude and direction of discrepancies. Finally, this study assessed the potential sources of uncertainty and reasons that could explain the discrepancy of food consumption estimates between FBS and dietary surveys, and provides recommendations for combining data sources to obtain more robust national consumption estimates.