Data collection and analysis tools for food security and nutrition

 The architecture of global data governance 
is comprised of an interlinked set of laws, 
conventions, protocols, and standards at the 
international, regional, national, and local levels. 
Gaps in this architecture have resulted in a lack 
of clarity that is undermining confidence in and 
adoption of new technologies and limiting the 
tools available to address harmful uses of data 
(CSIS, 2019, p. 1). 
As such, the governance of data intended 
to inform FSN policy action today must be 
addressed from a global perspective. Given 
the relevance of food security and nutrition for 
development, and the pervasiveness of food 
insecurity and malnutrition throughout the 
world, there are legitimate reasons to treat FSN 
data as a global public good, as has been long 
advocated for in research (Knottnerus, 2016) and 
as is now being proposed for the health sector 
(WHO, 2021).
The 2021 World Development Report (World 
Bank, 2021) devotes an entire chapter to the 
discussion of institutions for data governance. 
It is beyond the scope of this report to duplicate 
the informative, comprehensive treatment of 
data governance issues set forth in the World 
Development Report that apply to FSN data. 
Nevertheless, this chapter discusses some of 
the salient aspects that should be considered in 
designing effective governance mechanisms for 
FSN data.
ISSUES OF RELEVANCE FOR 
DATA GOVERNANCE
In this section, we discuss two key issues that 
continue to permeate discussions on data 
governance: questions around the concept of data 
ownership, and how to protect the right to privacy 
when dealing with personal data. These issues pertain 
to all types of data, but become particularly relevant 
in the context of FSN data when viewed from the 
perspective of the conceptual framework introduced 
in Chapter 1, which stresses the importance of the 
dimension of agency for food security and nutrition.
THE DEBATE ON THE NATURE OF DATA 
AND THE ROLE OF DATA MARKETS
Decades ago, the Nobel laureate Joseph Stiglitz 
(1999) presented the argument that information 
should be treated as a public good. According to 
the traditional economic definition, public goods 
(Reiss, 2021) are goods and services that are 
not 
excludable (meaning that once the good or service 
is available, fruition by anyone cannot be prevented, 
unless by enforcement mechanisms), and 
not rival 
in nature (meaning that “consumption” by one user 
does not reduce the availability or usefulness of the 
good or service for anyone else). The public nature 
of goods or services is one of the conditions leading 
to market failures, that is a suboptimal outcome if 
transactions or decisions are left to market forces 
alone (Bator, 1958; Stiglitz, 1989). In the case of 
public goods, in fact, efficiency arguments suggest 
that an unfettered, market-based mechanism 
would lead to their insufficient supply. Moreover, 
even if private agents engage in the production of a 
public good or service, the actual cost of making it 
available would be increased by the need to put in 
place special mechanisms to limit access to those 
who pay for it and avoid “free-riders”.
19
19 Even if in some cases – as, for example, with public health, 
education, and transportation – the government (or any other 
institution created to represent and protect the collective interest) 
might want to act as the private owner of the good and regulate 
access by requiring the payment of a fee, this is only justified when 
there is a concrete risk of overcrowding. This is very different from 
making those goods and services private. Although it is technically 
doable, privatization of essentially public goods and services is not 
necessarily desirable (Anderson, 1995).

---

[
85
5 
INSTITUTIONS AND GOVERNANCE FOR FSN DATA COLLECTION, ANALYSIS, AND USE
We strongly support the arguments made by 
Stiglitz, and extend it here to data, even in cases 
where one might want to distinguish between 
data 
and 
information (though, see Chapter 1). The main 
argument to support extending the notion of public 
good to data is that, especially now in the era of the 
internet and digitalisation, data have become the 
ultimate example of nonrivalry: millions of people 
may have access to the same data repeatedly, even 
simultaneously, without affecting the availability of the 
data to others. Moreover, now that virtually all data 
are available in digital form and stored in databases 
that can be accessed via the internet, the marginal 
cost needed to add one additional user is zero. This 
means that anything short of full, open access to data 
that has already been generated by someone and is 
stored in digital form, must be justified by arguments 
other economic efficiency (Badiee 
et al., 2021).
Market-like mechanisms through which business 
and research institutions obtain useful data 
have existed even before the digital era, but the 
argument here is that such markets should be 
recognized for what they are, namely, markets 
for 
data collection services, not for the data 
themselves. Indeed, there are good reasons why 
the development of competitive, efficient markets 
for data collection services should be promoted, 
fully exploiting the recent advances in information 
and communication technology that have made 
data collection much easier than before. It is 
the data collection service that possesses the 
characteristics of exclusivity, which supports the 
usefulness of a private transaction between a 
seller and a buyer, who is the consumer of the 
service. Treating the data itself as the object of 
the exchange presents numerous problems, 
beginning with the fact that – especially when 
data are produced and stored in digital form – 
exclusion is difficult (in addition to being morally 
questionable). Typically, treating data as the 
object of the exchange has been made possible by 
creating legal frameworks that extend provisions 
(created long ago and in very different contexts),
20
20 See 
https://en.wikipedia.org/wiki/Copyright
. Not surprisingly, the 
debate regarding whether copyright (as originally intended to protect 
the rights of the authors of literary and artistic productions) applies 
to digital resources, including what we have defined as data, is still 
very much open.
such as copyright, to various types of digitally 
stored data. Enforcement is then carried out via 
the introduction of firewalls and other technical 
barriers that limit or prevent access to the 
repository that contains the data, thus effectively 
limiting the possibility and extent of data re-
utilization.
In addition, exclusive reliance on private 
arrangements for data generation has long 
been considered inadequate. Traditionally, 
NSOs or similar agencies have been created in 
most countries to generate the data needed by 
governments to inform policies. Designed as 
autonomous public institutions, independent 
even of current executives, let alone of possible 
private interests — NSOs are still typically 
tasked with the responsibility of compiling and 
maintaining national accounts and generating 
other official statistics which are useful to 
guide policy. In the early operation of the NSOs, 
although relevant data was also generated by 
academic institutions and by private firms, the 
bulk of the data used to guide policymaking 
remained 
official and public.
The situation, however, is changing dramatically 
with the advent of the digital revolution and big 
data. Today, an incredibly large and increasing 
amount of new data and information, potentially 
relevant for policymaking, is generated outside 
the domains of official data and statistics, and 
therefore of NSOs. Many useful datasets covering 
agriculture and FSN are now available and can 
be openly and easily queried via the internet,
21
thanks to alternative arrangements promoting 
open access, such as CopyLeft,
22
Creative 
Commons
23
and Open Source Initiative.
24
These 
open-access datasets seem to be much better 
suited than copyright and fees-based licensing, 
to recognize and deal with the extant ethical 
problems related to data sharing. Furthermore, 
a global open-science movement is actively 
supporting the transition towards full, open 
access to scientific publications (Siew, 2017), and 
the principle by which data should be “as open 
as possible, as closed as necessary” (European 
Commission, 2016, p. 4) is what inspires 
accessibility among the Findable, Accessible, 
Interoperable and Reusable (FAIR) principles 

---

86 
]
DATA COLLECTION AND ANALYSIS TOOLS FOR FOOD SECURITY AND NUTRITION
(Landi 
et al., 2020). Despite this, many datasets 
are still “owned” by private entities who profit 
from an active market for data, which 
de facto 
promotes the view that data can be considered 
private, like any other private asset. Of special 
relevance is the fact that such datasets contain 
information that can be tremendously useful to 
inform development actions (including promoting 
FSN) and humanitarian interventions. It is in this 
new context that