Food demand elasticities are crucial parameters in the calibration of simulation models used to assess the impacts of political reforms or to analyse long-term projections, notably in agricultural sectors. Numerous estimates of these parameters are now available in the economic literature. The main objectives of this work are twofold: we seek first to identify general patterns characterizing the demand elasticities of food products and second to identify the main sources of heterogeneity between the elasticity estimates available in the literature. To achieve these objectives, we conduct a broad literature review of food demand elasticity estimates and perform a meta-regression analysis. Our results reveal the important impacts of income levels on income and price elasticities both at the country (gross domestic product-GDP) and household levels: the higher the income is, the lower the level of elasticities. Food demand responses to changes in income and prices appear to follow different patterns depending on the global regions involved apart from any income level consideration. From a methodological viewpoint, the functional forms used to represent food demand are found to significantly affect elasticity estimates. This result sheds light on the importance of the specification of demand functions, and particularly of their flexibility, in simulation models.