Global Forum on Food Security and Nutrition (FSN Forum)

Comments on the HLPE Report version 0: Data collection and analysis tools for food security and nutrition

Athur Mabiso
Senior Technical Specialist (Economist)
Research and Impact Assessment Division, IFAD
January 2022

Comments

As discussed in Chapter 4 of the report, there is need to explore the application of big data approaches and artificial intelligence (AI) to complement existing data tools. While AI has promise to enhance information systems that are relevant for food policy action to address issues of food insecurity and malnutrition, it is not always clear where the priorities lie. In this regard, the report could help shape the global conversation on some of the priority areas where data collection and analysis tools might provide avenues for broadly improving food security and nutrition.

The combination of survey methodologies as well as alternate data gathering approaches, including crowd-sourced data, should definitely all be on the table as highlighted in the report. However, there are varied limitations depending on the approaches, that will need to be taken into account. For instance, while big data approaches are quite useful for predictive analytics, there are limitations to using these kinds of data to assess impacts of interventions/investments designed to improve food security and nutrition. General guidelines on how these different approaches can and should be used will be important to avoid misuse.

This will also imply investments in human capacities as well as technologies, especially in developing country contexts.  To be able to leverage big data analytics for food security and nutrition significant human capacity is required and it is not yet clear to what extent governments will need to invest in their current workforce versus the future generation (particularly youth and children in secondary schools). Undoubtedly, a link between the education investments, curricular changes and issues of food security and nutrition will need to be made. The report may want to include a discussion on this topic – how to leverage investments in education and technology to address food security and nutrition challenges in developing countries.

With regard to traditional data collection approaches, which largely include household and community surveys as well as censuses, the quality of data obtained from these approaches still needs to be improved and just because new data approaches are emerging, we should not lose sight of the critical data obtained from traditional means. This will mean continuing to invest in and work closely with national statistical agencies and ministries of agriculture, health, nutrition, and gender (across disciplines) to generate statistics that are relevant for all actors to make a difference in food security and nutrition.

Following the COVID-19 pandemic, the world witnessed challenges in collecting up-to-date data on food security and nutrition. However, one tool that evidently provided vital information was the use of telephone surveys. For example, the World Bank launched its high-frequency telephone surveys, which allowed collection of timely data useful for policymaking. Going forward, there will need to be a careful assessment on how best to leverage this approach of collecting data, including determining best practices to ensure quality of data collected through this approach. At the same time, the approach of using telephones to conduct surveys has significant limitations that ought to be recognized and taken into account. For instance, the individuals who are likely to have access to a telephones will often be better off compared to those who live in the remotest rural areas and where cell phone network coverage is weak or does not exist. Moreover, there is evidence that women and some of the elderly people in poor communities might not have access to a telephone. As such, statistics on food security and nutrition generated through telephone surveys, particularly in developing country context, may have significant biases that could lead to erroneous policy decisions and actions. This needs careful consideration and should be emphasized in the report.

Thus, as part of improving the capabilities of nations to collect data using digital tools such as cell phones and tablets, there will be a need to invest in the necessary network infrastructure, enabling access for all, in addition to working on digital literacy. More traditional approaches may continue to be relevant for quite some time, including last-mile connectors (e.g. extension agents, mobile money merchants, etc.) who interface with many individuals who are not digitally connected or literate. Many such “data agents” work within government ministries at community levels and may prove to be a crucial part of the system for data collection for food security and nutrition.

The report aptly highlights the risks associated with the new data/digital technologies used for data collection and in particular data analysis. One of the issues at hand is the lack of a global data governance framework. The report may benefit from citing the work that is being undertaken by the UN High-Level Committee on Programmes (HLCP) where a global data governance framework will be looked into as part of its workstream under pillar 2, new global public goods: https://unsceb.org/session-report-369.

While this work is broader than food security and nutrition, it is relevant an perhaps there is a need to highlight how a global data governance framework might be put in place specifically to address issues of food security and nutrition.

Regarding recent initiatives (section 5.5 of the report), it may be worth including for review, the work of TetraTech supported by the Bill and Melinda Gates Foundation where several development partners are providing technical input, including IFAD: Enabling Crop Analytics At Scale.

A separate initiative also worth looking at is the Development Data Partnership, which includes UNDP, IMF, IDB, World Bank and OECD together with several private sector companies such as Google, Meta (Facebook) and Esri.