Global drivers of food system (un)sustainability: A multi-country correlation analysis.

At present, our ability to comprehend the dynamics of food systems and the consequences of their rapid 'transformations' is limited. In this paper, we propose to address this gap by exploring the interactions between the sustainability of food systems and a set of key drivers at the global scale. For this we compile a metric of 12 key drivers of food system from a globally-representative set of low, middle, and high-income countries and analyze the relationships between these drivers and a composite index that integrates the four key dimensions of food system sustainability, namely: food security & nutrition, environment, social, and economic dimensions. The two metrics highlight the important data gap that characterizes national systems' statistics-in particular in relation to transformation, transport, retail and distribution. Spearman correlations and Principal Component Analysis are then used to explore associations between levels of sustainability and drivers. With the exception of one economic driver (trade flows in merchandise and services), the majority of the statistically significant correlations found between food system sustainability and drivers appear to be negative. The fact that most of these negative drivers are closely related to the global demographic transition that is currently affecting the world population highlights the magnitude of the challenges ahead. This analysis is the first one that provides quantitative evidence at the global scale about correlations between the four dimensions of sustainability of our food systems and specific drivers.

Global map and indicators of food system sustainability.

This paper presents the first global map of food systems sustainability based on a rigorous protocol. The choice of the metric dimensions, as well as the individual indicators included in the metric, were initially identified from a thorough review of the existing literature. A rigorous inclusion/exclusion protocol was then used to refine the list and shorten it to a sub-set of 27 indicators. An aggregate sustainability score was then computed based on those 27 indicators organized into four dimensions: environment, social, food security & nutrition and economic. The paper shows how the availability of data (or lack therefore) results in an unavoidable trade-off between number of indicators and number of countries, and highlights how optimization can be used to present the most robust metric possible given the existence of this trade-offs in the data space. The process results in the computation of a global sustainability map covering 97 countries and 20 indicators. The sustainability scores obtained for each country are made available over the entire range of indicators.