Joanna Upton is a Postdoctoral Associate and Jennifer Denno Cissé is a PhD candidate at Cornell’s Dyson School.
The international community came to agreement on the definition of food security during the 1996 World Food Summit, defining food security as existing when “…all people at all times have physical, social, and economic access to sufficient, safe and nutritious food that meets their dietary needs and food preferences for an active and healthy life.” Well before and since this consensus was reached, researchers and practitioners have been striving to improve measurement of food security. A definition is crucial for the guidance of measurement, and measurement in turn is a key “driver of diagnosis and response.”
In a new working paper with Chris Barrett, we identify four axioms that an ideal measure of food security should address if it is to encompass the internationally agreed upon definition. We then assess the state of today’s food security measures and propose a new approach to food security measurement that meets all four of these axioms.
The four food security measurement axioms we propose are:
- the scale axiom, which allows us to understand the situation of “all people,” e., nations as well as sub-groups and individuals,
- the time axiom, which measures the stability of these conditions at “all times,” i.e., between seasons and over years,
- the access axiom, which captures “physical, social, and economic access,” and finally
- the outcomes axiom, which measures whether individuals are able to lead “an active and healthy life.”
For both illustrative purposes and further insight, we review existing food security metrics, using these axioms to clarify the gaps and highlight potential trade-offs between measurement approaches. Perhaps not surprisingly, no current food security measure captures all four axioms.
In common conversation about food security and global rates of under-nutrition, we often refer more narrowly to measures of caloric availability at national scales—all but ignoring the deeper facets of access, use, and human outcomes. Household-level measures often dig deeper, but even then they usually focus on one of these components, such as food access. Most measures fail to address the definitional components of “all people” and “at all times;” often attempts to do so involve trade-offs. For example, measuring something with high frequency is the best way to meet the time axiom, but given limited budgets, the cost of measuring meaningful health outcomes frequently, especially at the individual level, creates challenges for meeting either the outcomes or scale axioms.
Drawing on the recent literature on the theory of development resilience, we take a development resilience approach to food security measurement. This probabilistic method of measuring well-being dynamics was developed by Jenn Cissé and Chris Barrett and recently explained in an ETRM blog post. Taking a health and/or dietary diversity indicator as the outcome, and conditioning on key indicators of access, this approach allows us to look not just at the outcome in a static way but at the probability of falling above a certain normative threshold over time, given access and other conditions.
We observe in this approach an unprecedented axiomatic “fit” to the internationally agreed definition of food security.
- The scale axiom can be satisfied by examining both individual food security and applying the Foster-Greer-Thorbecke (FGT) aggregation method to look at the food security of any sub-group or of the total population sampled.
- The resilience approach satisfies the time axiom in taking dynamics into account and enabling forward-looking forecasts, not just description of past observations.
- By conditioning on core access variables, like income levels and/or assets, the approach also allows one to examine the importance of heterogeneous access as a key-conditioning factor of food security, which can be essential in terms of understanding mechanisms.
- Finally, the choice of outcome variable helps us to satisfy the outcomes axiom.
As highlighted in discussions around the paper as it was presented at the International Conference of Agricultural Economists in Milan, the outcome of focus can be context-dependent. For example, where child mortality is a serious risk we might focus on a wasting measure such as weight-for-age or the mid-upper arm circumference (MUAC), whereas a dietary diversity measure might be a better outcome for assessing the stability of access to a diversity of nutritious foods.
We demonstrate this method by applying it to data on pastoralists in northern Kenya, collected by a consortium led by ILRI with funding from DFID, USAID, and the World Bank. Using the resilience approach with the Household Dietary Diversity Score (HDDS) and MUAC as outcomes provides a few insights into the data and setting. For example, while households with a higher dependency ratio (ratio of dependents to income-earners) do not have worse outcomes in a given time period, they do experience lower probabilities of maintaining positive outcomes over time, so are on average less food secure—which we fail to observe with static analysis. This dynamic approach has additional advantages for both targeting and programming in food assistance programs.
This new measure remains very much a work in progress. As with all measures of food security, it suffers from challenges in acquiring sufficient, high quality data. However, given high frequency, longitudinal, data collection in the world’s most vulnerable regions, this approach provides a new opportunity to develop robust, axiomatic measures of food security so as to improve food security programming through the use of improved measures for diagnosis, inference, prediction and targeting.
This work was made possible by the generous support of the USDA’s Economic Research Service and the International Livestock Research Institute.